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 1.0.2p and 1.0.2q [xx XXX xxxx]
12 *) Microarchitecture timing vulnerability in ECC scalar multiplication
14 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
15 shown to be vulnerable to a microarchitecture timing side channel attack.
16 An attacker with sufficient access to mount local timing attacks during
17 ECDSA signature generation could recover the private key.
19 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
20 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
25 *) Timing vulnerability in DSA signature generation
27 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
28 timing side channel attack. An attacker could use variations in the signing
29 algorithm to recover the private key.
31 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
35 *) Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
36 Module, accidentally introduced while backporting security fixes from the
37 development branch and hindering the use of ECC in FIPS mode.
40 Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
42 *) Client DoS due to large DH parameter
44 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
45 malicious server can send a very large prime value to the client. This will
46 cause the client to spend an unreasonably long period of time generating a
47 key for this prime resulting in a hang until the client has finished. This
48 could be exploited in a Denial Of Service attack.
50 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
54 *) Cache timing vulnerability in RSA Key Generation
56 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
57 a cache timing side channel attack. An attacker with sufficient access to
58 mount cache timing attacks during the RSA key generation process could
59 recover the private key.
61 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
62 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
66 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
67 parameter is no longer accepted, as it leads to a corrupt table. NULL
68 pem_str is reserved for alias entries only.
71 *) Revert blinding in ECDSA sign and instead make problematic addition
72 length-invariant. Switch even to fixed-length Montgomery multiplication.
75 *) Change generating and checking of primes so that the error rate of not
76 being prime depends on the intended use based on the size of the input.
77 For larger primes this will result in more rounds of Miller-Rabin.
78 The maximal error rate for primes with more than 1080 bits is lowered
80 [Kurt Roeckx, Annie Yousar]
82 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
85 *) Add blinding to ECDSA and DSA signatures to protect against side channel
86 attacks discovered by Keegan Ryan (NCC Group).
89 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
90 now allow empty (zero character) pass phrases.
93 *) Certificate time validation (X509_cmp_time) enforces stricter
94 compliance with RFC 5280. Fractional seconds and timezone offsets
95 are no longer allowed.
98 Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
100 *) Constructed ASN.1 types with a recursive definition could exceed the stack
102 Constructed ASN.1 types with a recursive definition (such as can be found
103 in PKCS7) could eventually exceed the stack given malicious input with
104 excessive recursion. This could result in a Denial Of Service attack. There
105 are no such structures used within SSL/TLS that come from untrusted sources
106 so this is considered safe.
108 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
113 Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
115 *) Read/write after SSL object in error state
117 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
118 mechanism. The intent was that if a fatal error occurred during a handshake
119 then OpenSSL would move into the error state and would immediately fail if
120 you attempted to continue the handshake. This works as designed for the
121 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
122 SSL_connect()), however due to a bug it does not work correctly if
123 SSL_read() or SSL_write() is called directly. In that scenario, if the
124 handshake fails then a fatal error will be returned in the initial function
125 call. If SSL_read()/SSL_write() is subsequently called by the application
126 for the same SSL object then it will succeed and the data is passed without
127 being decrypted/encrypted directly from the SSL/TLS record layer.
129 In order to exploit this issue an application bug would have to be present
130 that resulted in a call to SSL_read()/SSL_write() being issued after having
131 already received a fatal error.
133 This issue was reported to OpenSSL by David Benjamin (Google).
137 *) rsaz_1024_mul_avx2 overflow bug on x86_64
139 There is an overflow bug in the AVX2 Montgomery multiplication procedure
140 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
141 Analysis suggests that attacks against RSA and DSA as a result of this
142 defect would be very difficult to perform and are not believed likely.
143 Attacks against DH1024 are considered just feasible, because most of the
144 work necessary to deduce information about a private key may be performed
145 offline. The amount of resources required for such an attack would be
146 significant. However, for an attack on TLS to be meaningful, the server
147 would have to share the DH1024 private key among multiple clients, which is
148 no longer an option since CVE-2016-0701.
150 This only affects processors that support the AVX2 but not ADX extensions
151 like Intel Haswell (4th generation).
153 This issue was reported to OpenSSL by David Benjamin (Google). The issue
154 was originally found via the OSS-Fuzz project.
158 Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
160 *) bn_sqrx8x_internal carry bug on x86_64
162 There is a carry propagating bug in the x86_64 Montgomery squaring
163 procedure. No EC algorithms are affected. Analysis suggests that attacks
164 against RSA and DSA as a result of this defect would be very difficult to
165 perform and are not believed likely. Attacks against DH are considered just
166 feasible (although very difficult) because most of the work necessary to
167 deduce information about a private key may be performed offline. The amount
168 of resources required for such an attack would be very significant and
169 likely only accessible to a limited number of attackers. An attacker would
170 additionally need online access to an unpatched system using the target
171 private key in a scenario with persistent DH parameters and a private
172 key that is shared between multiple clients.
174 This only affects processors that support the BMI1, BMI2 and ADX extensions
175 like Intel Broadwell (5th generation) and later or AMD Ryzen.
177 This issue was reported to OpenSSL by the OSS-Fuzz project.
181 *) Malformed X.509 IPAddressFamily could cause OOB read
183 If an X.509 certificate has a malformed IPAddressFamily extension,
184 OpenSSL could do a one-byte buffer overread. The most likely result
185 would be an erroneous display of the certificate in text format.
187 This issue was reported to OpenSSL by the OSS-Fuzz project.
191 Changes between 1.0.2k and 1.0.2l [25 May 2017]
193 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
194 platform rather than 'mingw'.
197 Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
199 *) Truncated packet could crash via OOB read
201 If one side of an SSL/TLS path is running on a 32-bit host and a specific
202 cipher is being used, then a truncated packet can cause that host to
203 perform an out-of-bounds read, usually resulting in a crash.
205 This issue was reported to OpenSSL by Robert Święcki of Google.
209 *) BN_mod_exp may produce incorrect results on x86_64
211 There is a carry propagating bug in the x86_64 Montgomery squaring
212 procedure. No EC algorithms are affected. Analysis suggests that attacks
213 against RSA and DSA as a result of this defect would be very difficult to
214 perform and are not believed likely. Attacks against DH are considered just
215 feasible (although very difficult) because most of the work necessary to
216 deduce information about a private key may be performed offline. The amount
217 of resources required for such an attack would be very significant and
218 likely only accessible to a limited number of attackers. An attacker would
219 additionally need online access to an unpatched system using the target
220 private key in a scenario with persistent DH parameters and a private
221 key that is shared between multiple clients. For example this can occur by
222 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
223 similar to CVE-2015-3193 but must be treated as a separate problem.
225 This issue was reported to OpenSSL by the OSS-Fuzz project.
229 *) Montgomery multiplication may produce incorrect results
231 There is a carry propagating bug in the Broadwell-specific Montgomery
232 multiplication procedure that handles input lengths divisible by, but
233 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
234 and DH private keys are impossible. This is because the subroutine in
235 question is not used in operations with the private key itself and an input
236 of the attacker's direct choice. Otherwise the bug can manifest itself as
237 transient authentication and key negotiation failures or reproducible
238 erroneous outcome of public-key operations with specially crafted input.
239 Among EC algorithms only Brainpool P-512 curves are affected and one
240 presumably can attack ECDH key negotiation. Impact was not analyzed in
241 detail, because pre-requisites for attack are considered unlikely. Namely
242 multiple clients have to choose the curve in question and the server has to
243 share the private key among them, neither of which is default behaviour.
244 Even then only clients that chose the curve will be affected.
246 This issue was publicly reported as transient failures and was not
247 initially recognized as a security issue. Thanks to Richard Morgan for
248 providing reproducible case.
252 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
253 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
254 prevent issues where no progress is being made and the peer continually
255 sends unrecognised record types, using up resources processing them.
258 Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
260 *) Missing CRL sanity check
262 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
263 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
264 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
266 This issue only affects the OpenSSL 1.0.2i
270 Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
272 *) OCSP Status Request extension unbounded memory growth
274 A malicious client can send an excessively large OCSP Status Request
275 extension. If that client continually requests renegotiation, sending a
276 large OCSP Status Request extension each time, then there will be unbounded
277 memory growth on the server. This will eventually lead to a Denial Of
278 Service attack through memory exhaustion. Servers with a default
279 configuration are vulnerable even if they do not support OCSP. Builds using
280 the "no-ocsp" build time option are not affected.
282 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
286 *) In order to mitigate the SWEET32 attack, the DES ciphers were moved from
289 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
294 *) OOB write in MDC2_Update()
296 An overflow can occur in MDC2_Update() either if called directly or
297 through the EVP_DigestUpdate() function using MDC2. If an attacker
298 is able to supply very large amounts of input data after a previous
299 call to EVP_EncryptUpdate() with a partial block then a length check
300 can overflow resulting in a heap corruption.
302 The amount of data needed is comparable to SIZE_MAX which is impractical
305 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
309 *) Malformed SHA512 ticket DoS
311 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
312 DoS attack where a malformed ticket will result in an OOB read which will
315 The use of SHA512 in TLS session tickets is comparatively rare as it requires
316 a custom server callback and ticket lookup mechanism.
318 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
322 *) OOB write in BN_bn2dec()
324 The function BN_bn2dec() does not check the return value of BN_div_word().
325 This can cause an OOB write if an application uses this function with an
326 overly large BIGNUM. This could be a problem if an overly large certificate
327 or CRL is printed out from an untrusted source. TLS is not affected because
328 record limits will reject an oversized certificate before it is parsed.
330 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
334 *) OOB read in TS_OBJ_print_bio()
336 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
337 the total length the OID text representation would use and not the amount
338 of data written. This will result in OOB reads when large OIDs are
341 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
345 *) Pointer arithmetic undefined behaviour
347 Avoid some undefined pointer arithmetic
349 A common idiom in the codebase is to check limits in the following manner:
352 Where "p" points to some malloc'd data of SIZE bytes and
355 "len" here could be from some externally supplied data (e.g. from a TLS
358 The rules of C pointer arithmetic are such that "p + len" is only well
359 defined where len <= SIZE. Therefore the above idiom is actually
362 For example this could cause problems if some malloc implementation
363 provides an address for "p" such that "p + len" actually overflows for
364 values of len that are too big and therefore p + len < limit.
366 This issue was reported to OpenSSL by Guido Vranken
370 *) Constant time flag not preserved in DSA signing
372 Operations in the DSA signing algorithm should run in constant time in
373 order to avoid side channel attacks. A flaw in the OpenSSL DSA
374 implementation means that a non-constant time codepath is followed for
375 certain operations. This has been demonstrated through a cache-timing
376 attack to be sufficient for an attacker to recover the private DSA key.
378 This issue was reported by César Pereida (Aalto University), Billy Brumley
379 (Tampere University of Technology), and Yuval Yarom (The University of
384 *) DTLS buffered message DoS
386 In a DTLS connection where handshake messages are delivered out-of-order
387 those messages that OpenSSL is not yet ready to process will be buffered
388 for later use. Under certain circumstances, a flaw in the logic means that
389 those messages do not get removed from the buffer even though the handshake
390 has been completed. An attacker could force up to approx. 15 messages to
391 remain in the buffer when they are no longer required. These messages will
392 be cleared when the DTLS connection is closed. The default maximum size for
393 a message is 100k. Therefore the attacker could force an additional 1500k
394 to be consumed per connection. By opening many simulataneous connections an
395 attacker could cause a DoS attack through memory exhaustion.
397 This issue was reported to OpenSSL by Quan Luo.
401 *) DTLS replay protection DoS
403 A flaw in the DTLS replay attack protection mechanism means that records
404 that arrive for future epochs update the replay protection "window" before
405 the MAC for the record has been validated. This could be exploited by an
406 attacker by sending a record for the next epoch (which does not have to
407 decrypt or have a valid MAC), with a very large sequence number. This means
408 that all subsequent legitimate packets are dropped causing a denial of
409 service for a specific DTLS connection.
411 This issue was reported to OpenSSL by the OCAP audit team.
415 *) Certificate message OOB reads
417 In OpenSSL 1.0.2 and earlier some missing message length checks can result
418 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
419 theoretical DoS risk but this has not been observed in practice on common
422 The messages affected are client certificate, client certificate request
423 and server certificate. As a result the attack can only be performed
424 against a client or a server which enables client authentication.
426 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
430 Changes between 1.0.2g and 1.0.2h [3 May 2016]
432 *) Prevent padding oracle in AES-NI CBC MAC check
434 A MITM attacker can use a padding oracle attack to decrypt traffic
435 when the connection uses an AES CBC cipher and the server support
438 This issue was introduced as part of the fix for Lucky 13 padding
439 attack (CVE-2013-0169). The padding check was rewritten to be in
440 constant time by making sure that always the same bytes are read and
441 compared against either the MAC or padding bytes. But it no longer
442 checked that there was enough data to have both the MAC and padding
445 This issue was reported by Juraj Somorovsky using TLS-Attacker.
449 *) Fix EVP_EncodeUpdate overflow
451 An overflow can occur in the EVP_EncodeUpdate() function which is used for
452 Base64 encoding of binary data. If an attacker is able to supply very large
453 amounts of input data then a length check can overflow resulting in a heap
456 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
457 the PEM_write_bio* family of functions. These are mainly used within the
458 OpenSSL command line applications, so any application which processes data
459 from an untrusted source and outputs it as a PEM file should be considered
460 vulnerable to this issue. User applications that call these APIs directly
461 with large amounts of untrusted data may also be vulnerable.
463 This issue was reported by Guido Vranken.
467 *) Fix EVP_EncryptUpdate overflow
469 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
470 is able to supply very large amounts of input data after a previous call to
471 EVP_EncryptUpdate() with a partial block then a length check can overflow
472 resulting in a heap corruption. Following an analysis of all OpenSSL
473 internal usage of the EVP_EncryptUpdate() function all usage is one of two
474 forms. The first form is where the EVP_EncryptUpdate() call is known to be
475 the first called function after an EVP_EncryptInit(), and therefore that
476 specific call must be safe. The second form is where the length passed to
477 EVP_EncryptUpdate() can be seen from the code to be some small value and
478 therefore there is no possibility of an overflow. Since all instances are
479 one of these two forms, it is believed that there can be no overflows in
480 internal code due to this problem. It should be noted that
481 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
482 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
483 of these calls have also been analysed too and it is believed there are no
484 instances in internal usage where an overflow could occur.
486 This issue was reported by Guido Vranken.
490 *) Prevent ASN.1 BIO excessive memory allocation
492 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
493 a short invalid encoding can casuse allocation of large amounts of memory
494 potentially consuming excessive resources or exhausting memory.
496 Any application parsing untrusted data through d2i BIO functions is
497 affected. The memory based functions such as d2i_X509() are *not* affected.
498 Since the memory based functions are used by the TLS library, TLS
499 applications are not affected.
501 This issue was reported by Brian Carpenter.
507 ASN1 Strings that are over 1024 bytes can cause an overread in applications
508 using the X509_NAME_oneline() function on EBCDIC systems. This could result
509 in arbitrary stack data being returned in the buffer.
511 This issue was reported by Guido Vranken.
515 *) Modify behavior of ALPN to invoke callback after SNI/servername
516 callback, such that updates to the SSL_CTX affect ALPN.
519 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
523 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
524 methods are enabled and ssl2 is disabled the methods return NULL.
527 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
529 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
530 Builds that are not configured with "enable-weak-ssl-ciphers" will not
531 provide any "EXPORT" or "LOW" strength ciphers.
534 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
535 is by default disabled at build-time. Builds that are not configured with
536 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
537 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
538 will need to explicitly call either of:
540 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
542 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
544 as appropriate. Even if either of those is used, or the application
545 explicitly uses the version-specific SSLv2_method() or its client and
546 server variants, SSLv2 ciphers vulnerable to exhaustive search key
547 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
548 ciphers, and SSLv2 56-bit DES are no longer available.
552 *) Fix a double-free in DSA code
554 A double free bug was discovered when OpenSSL parses malformed DSA private
555 keys and could lead to a DoS attack or memory corruption for applications
556 that receive DSA private keys from untrusted sources. This scenario is
559 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
564 *) Disable SRP fake user seed to address a server memory leak.
566 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
568 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
569 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
570 was changed to ignore the "fake user" SRP seed, even if the seed
573 Users should use SRP_VBASE_get1_by_user instead. Note that in
574 SRP_VBASE_get1_by_user, caller must free the returned value. Note
575 also that even though configuring the SRP seed attempts to hide
576 invalid usernames by continuing the handshake with fake
577 credentials, this behaviour is not constant time and no strong
578 guarantees are made that the handshake is indistinguishable from
579 that of a valid user.
583 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
585 In the BN_hex2bn function the number of hex digits is calculated using an
586 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
587 large values of |i| this can result in |bn_expand| not allocating any
588 memory because |i * 4| is negative. This can leave the internal BIGNUM data
589 field as NULL leading to a subsequent NULL ptr deref. For very large values
590 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
591 In this case memory is allocated to the internal BIGNUM data field, but it
592 is insufficiently sized leading to heap corruption. A similar issue exists
593 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
594 is ever called by user applications with very large untrusted hex/dec data.
595 This is anticipated to be a rare occurrence.
597 All OpenSSL internal usage of these functions use data that is not expected
598 to be untrusted, e.g. config file data or application command line
599 arguments. If user developed applications generate config file data based
600 on untrusted data then it is possible that this could also lead to security
601 consequences. This is also anticipated to be rare.
603 This issue was reported to OpenSSL by Guido Vranken.
607 *) Fix memory issues in BIO_*printf functions
609 The internal |fmtstr| function used in processing a "%s" format string in
610 the BIO_*printf functions could overflow while calculating the length of a
611 string and cause an OOB read when printing very long strings.
613 Additionally the internal |doapr_outch| function can attempt to write to an
614 OOB memory location (at an offset from the NULL pointer) in the event of a
615 memory allocation failure. In 1.0.2 and below this could be caused where
616 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
617 could be in processing a very long "%s" format string. Memory leaks can
620 The first issue may mask the second issue dependent on compiler behaviour.
621 These problems could enable attacks where large amounts of untrusted data
622 is passed to the BIO_*printf functions. If applications use these functions
623 in this way then they could be vulnerable. OpenSSL itself uses these
624 functions when printing out human-readable dumps of ASN.1 data. Therefore
625 applications that print this data could be vulnerable if the data is from
626 untrusted sources. OpenSSL command line applications could also be
627 vulnerable where they print out ASN.1 data, or if untrusted data is passed
628 as command line arguments.
630 Libssl is not considered directly vulnerable. Additionally certificates etc
631 received via remote connections via libssl are also unlikely to be able to
632 trigger these issues because of message size limits enforced within libssl.
634 This issue was reported to OpenSSL Guido Vranken.
638 *) Side channel attack on modular exponentiation
640 A side-channel attack was found which makes use of cache-bank conflicts on
641 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
642 of RSA keys. The ability to exploit this issue is limited as it relies on
643 an attacker who has control of code in a thread running on the same
644 hyper-threaded core as the victim thread which is performing decryptions.
646 This issue was reported to OpenSSL by Yuval Yarom, The University of
647 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
648 Nadia Heninger, University of Pennsylvania with more information at
649 http://cachebleed.info.
653 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
654 if no keysize is specified with default_bits. This fixes an
655 omission in an earlier change that changed all RSA/DSA key generation
656 apps to use 2048 bits by default.
659 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
661 *) DH small subgroups
663 Historically OpenSSL only ever generated DH parameters based on "safe"
664 primes. More recently (in version 1.0.2) support was provided for
665 generating X9.42 style parameter files such as those required for RFC 5114
666 support. The primes used in such files may not be "safe". Where an
667 application is using DH configured with parameters based on primes that are
668 not "safe" then an attacker could use this fact to find a peer's private
669 DH exponent. This attack requires that the attacker complete multiple
670 handshakes in which the peer uses the same private DH exponent. For example
671 this could be used to discover a TLS server's private DH exponent if it's
672 reusing the private DH exponent or it's using a static DH ciphersuite.
674 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
675 TLS. It is not on by default. If the option is not set then the server
676 reuses the same private DH exponent for the life of the server process and
677 would be vulnerable to this attack. It is believed that many popular
678 applications do set this option and would therefore not be at risk.
680 The fix for this issue adds an additional check where a "q" parameter is
681 available (as is the case in X9.42 based parameters). This detects the
682 only known attack, and is the only possible defense for static DH
683 ciphersuites. This could have some performance impact.
685 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
686 default and cannot be disabled. This could have some performance impact.
688 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
692 *) SSLv2 doesn't block disabled ciphers
694 A malicious client can negotiate SSLv2 ciphers that have been disabled on
695 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
696 been disabled, provided that the SSLv2 protocol was not also disabled via
699 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
700 and Sebastian Schinzel.
704 *) Reject DH handshakes with parameters shorter than 1024 bits.
707 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
709 *) BN_mod_exp may produce incorrect results on x86_64
711 There is a carry propagating bug in the x86_64 Montgomery squaring
712 procedure. No EC algorithms are affected. Analysis suggests that attacks
713 against RSA and DSA as a result of this defect would be very difficult to
714 perform and are not believed likely. Attacks against DH are considered just
715 feasible (although very difficult) because most of the work necessary to
716 deduce information about a private key may be performed offline. The amount
717 of resources required for such an attack would be very significant and
718 likely only accessible to a limited number of attackers. An attacker would
719 additionally need online access to an unpatched system using the target
720 private key in a scenario with persistent DH parameters and a private
721 key that is shared between multiple clients. For example this can occur by
722 default in OpenSSL DHE based SSL/TLS ciphersuites.
724 This issue was reported to OpenSSL by Hanno Böck.
728 *) Certificate verify crash with missing PSS parameter
730 The signature verification routines will crash with a NULL pointer
731 dereference if presented with an ASN.1 signature using the RSA PSS
732 algorithm and absent mask generation function parameter. Since these
733 routines are used to verify certificate signature algorithms this can be
734 used to crash any certificate verification operation and exploited in a
735 DoS attack. Any application which performs certificate verification is
736 vulnerable including OpenSSL clients and servers which enable client
739 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
743 *) X509_ATTRIBUTE memory leak
745 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
746 memory. This structure is used by the PKCS#7 and CMS routines so any
747 application which reads PKCS#7 or CMS data from untrusted sources is
748 affected. SSL/TLS is not affected.
750 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
755 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
756 This changes the decoding behaviour for some invalid messages,
757 though the change is mostly in the more lenient direction, and
758 legacy behaviour is preserved as much as possible.
761 *) In DSA_generate_parameters_ex, if the provided seed is too short,
762 use a random seed, as already documented.
763 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
765 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
767 *) Alternate chains certificate forgery
769 During certificate verfification, OpenSSL will attempt to find an
770 alternative certificate chain if the first attempt to build such a chain
771 fails. An error in the implementation of this logic can mean that an
772 attacker could cause certain checks on untrusted certificates to be
773 bypassed, such as the CA flag, enabling them to use a valid leaf
774 certificate to act as a CA and "issue" an invalid certificate.
776 This issue was reported to OpenSSL by Adam Langley/David Benjamin
781 *) Race condition handling PSK identify hint
783 If PSK identity hints are received by a multi-threaded client then
784 the values are wrongly updated in the parent SSL_CTX structure. This can
785 result in a race condition potentially leading to a double free of the
790 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
792 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
793 incompatibility in the handling of HMAC. The previous ABI has now been
796 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
798 *) Malformed ECParameters causes infinite loop
800 When processing an ECParameters structure OpenSSL enters an infinite loop
801 if the curve specified is over a specially malformed binary polynomial
804 This can be used to perform denial of service against any
805 system which processes public keys, certificate requests or
806 certificates. This includes TLS clients and TLS servers with
807 client authentication enabled.
809 This issue was reported to OpenSSL by Joseph Barr-Pixton.
813 *) Exploitable out-of-bounds read in X509_cmp_time
815 X509_cmp_time does not properly check the length of the ASN1_TIME
816 string and can read a few bytes out of bounds. In addition,
817 X509_cmp_time accepts an arbitrary number of fractional seconds in the
820 An attacker can use this to craft malformed certificates and CRLs of
821 various sizes and potentially cause a segmentation fault, resulting in
822 a DoS on applications that verify certificates or CRLs. TLS clients
823 that verify CRLs are affected. TLS clients and servers with client
824 authentication enabled may be affected if they use custom verification
827 This issue was reported to OpenSSL by Robert Swiecki (Google), and
828 independently by Hanno Böck.
832 *) PKCS7 crash with missing EnvelopedContent
834 The PKCS#7 parsing code does not handle missing inner EncryptedContent
835 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
836 with missing content and trigger a NULL pointer dereference on parsing.
838 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
839 structures from untrusted sources are affected. OpenSSL clients and
840 servers are not affected.
842 This issue was reported to OpenSSL by Michal Zalewski (Google).
846 *) CMS verify infinite loop with unknown hash function
848 When verifying a signedData message the CMS code can enter an infinite loop
849 if presented with an unknown hash function OID. This can be used to perform
850 denial of service against any system which verifies signedData messages using
852 This issue was reported to OpenSSL by Johannes Bauer.
856 *) Race condition handling NewSessionTicket
858 If a NewSessionTicket is received by a multi-threaded client when attempting to
859 reuse a previous ticket then a race condition can occur potentially leading to
860 a double free of the ticket data.
864 *) Removed support for the two export grade static DH ciphersuites
865 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
866 were newly added (along with a number of other static DH ciphersuites) to
867 1.0.2. However the two export ones have *never* worked since they were
868 introduced. It seems strange in any case to be adding new export
869 ciphersuites, and given "logjam" it also does not seem correct to fix them.
872 *) Only support 256-bit or stronger elliptic curves with the
873 'ecdh_auto' setting (server) or by default (client). Of supported
874 curves, prefer P-256 (both).
877 *) Reject DH handshakes with parameters shorter than 768 bits.
878 [Kurt Roeckx and Emilia Kasper]
880 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
882 *) ClientHello sigalgs DoS fix
884 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
885 invalid signature algorithms extension a NULL pointer dereference will
886 occur. This can be exploited in a DoS attack against the server.
888 This issue was was reported to OpenSSL by David Ramos of Stanford
891 [Stephen Henson and Matt Caswell]
893 *) Multiblock corrupted pointer fix
895 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
896 feature only applies on 64 bit x86 architecture platforms that support AES
897 NI instructions. A defect in the implementation of "multiblock" can cause
898 OpenSSL's internal write buffer to become incorrectly set to NULL when
899 using non-blocking IO. Typically, when the user application is using a
900 socket BIO for writing, this will only result in a failed connection.
901 However if some other BIO is used then it is likely that a segmentation
902 fault will be triggered, thus enabling a potential DoS attack.
904 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
908 *) Segmentation fault in DTLSv1_listen fix
910 The DTLSv1_listen function is intended to be stateless and processes the
911 initial ClientHello from many peers. It is common for user code to loop
912 over the call to DTLSv1_listen until a valid ClientHello is received with
913 an associated cookie. A defect in the implementation of DTLSv1_listen means
914 that state is preserved in the SSL object from one invocation to the next
915 that can lead to a segmentation fault. Errors processing the initial
916 ClientHello can trigger this scenario. An example of such an error could be
917 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
920 This issue was reported to OpenSSL by Per Allansson.
924 *) Segmentation fault in ASN1_TYPE_cmp fix
926 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
927 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
928 certificate signature algorithm consistency this can be used to crash any
929 certificate verification operation and exploited in a DoS attack. Any
930 application which performs certificate verification is vulnerable including
931 OpenSSL clients and servers which enable client authentication.
935 *) Segmentation fault for invalid PSS parameters fix
937 The signature verification routines will crash with a NULL pointer
938 dereference if presented with an ASN.1 signature using the RSA PSS
939 algorithm and invalid parameters. Since these routines are used to verify
940 certificate signature algorithms this can be used to crash any
941 certificate verification operation and exploited in a DoS attack. Any
942 application which performs certificate verification is vulnerable including
943 OpenSSL clients and servers which enable client authentication.
945 This issue was was reported to OpenSSL by Brian Carpenter.
949 *) ASN.1 structure reuse memory corruption fix
951 Reusing a structure in ASN.1 parsing may allow an attacker to cause
952 memory corruption via an invalid write. Such reuse is and has been
953 strongly discouraged and is believed to be rare.
955 Applications that parse structures containing CHOICE or ANY DEFINED BY
956 components may be affected. Certificate parsing (d2i_X509 and related
957 functions) are however not affected. OpenSSL clients and servers are
962 *) PKCS7 NULL pointer dereferences fix
964 The PKCS#7 parsing code does not handle missing outer ContentInfo
965 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
966 missing content and trigger a NULL pointer dereference on parsing.
968 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
969 otherwise parse PKCS#7 structures from untrusted sources are
970 affected. OpenSSL clients and servers are not affected.
972 This issue was reported to OpenSSL by Michal Zalewski (Google).
976 *) DoS via reachable assert in SSLv2 servers fix
978 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
979 servers that both support SSLv2 and enable export cipher suites by sending
980 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
982 This issue was discovered by Sean Burford (Google) and Emilia Käsper
983 (OpenSSL development team).
987 *) Empty CKE with client auth and DHE fix
989 If client auth is used then a server can seg fault in the event of a DHE
990 ciphersuite being selected and a zero length ClientKeyExchange message
991 being sent by the client. This could be exploited in a DoS attack.
995 *) Handshake with unseeded PRNG fix
997 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
998 with an unseeded PRNG. The conditions are:
999 - The client is on a platform where the PRNG has not been seeded
1000 automatically, and the user has not seeded manually
1001 - A protocol specific client method version has been used (i.e. not
1002 SSL_client_methodv23)
1003 - A ciphersuite is used that does not require additional random data from
1004 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1006 If the handshake succeeds then the client random that has been used will
1007 have been generated from a PRNG with insufficient entropy and therefore the
1008 output may be predictable.
1010 For example using the following command with an unseeded openssl will
1011 succeed on an unpatched platform:
1013 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1017 *) Use After Free following d2i_ECPrivatekey error fix
1019 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1020 could cause a use after free condition. This, in turn, could cause a double
1021 free in several private key parsing functions (such as d2i_PrivateKey
1022 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1023 for applications that receive EC private keys from untrusted
1024 sources. This scenario is considered rare.
1026 This issue was discovered by the BoringSSL project and fixed in their
1031 *) X509_to_X509_REQ NULL pointer deref fix
1033 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1034 the certificate key is invalid. This function is rarely used in practice.
1036 This issue was discovered by Brian Carpenter.
1040 *) Removed the export ciphers from the DEFAULT ciphers
1043 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1045 *) Change RSA and DH/DSA key generation apps to generate 2048-bit
1049 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1050 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1051 So far those who have to target multiple plaforms would compromise
1052 and argue that binary targeting say ARMv5 would still execute on
1053 ARMv8. "Universal" build resolves this compromise by providing
1054 near-optimal performance even on newer platforms.
1057 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1058 (other platforms pending).
1059 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1061 *) Add support for the SignedCertificateTimestampList certificate and
1062 OCSP response extensions from RFC6962.
1065 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1066 for corner cases. (Certain input points at infinity could lead to
1067 bogus results, with non-infinity inputs mapped to infinity too.)
1070 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1071 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1072 common cases are optimized and there still is room for further
1073 improvements. Vector Permutation AES for Altivec is also added.
1076 *) Add support for little-endian ppc64 Linux target.
1077 [Marcelo Cerri (IBM)]
1079 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1080 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1081 are optimized and there still is room for further improvements.
1082 Both 32- and 64-bit modes are supported.
1083 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1085 *) Improved ARMv7 NEON support.
1088 *) Support for SPARC Architecture 2011 crypto extensions, first
1089 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1090 SHA256/512, MD5, GHASH and modular exponentiation.
1091 [Andy Polyakov, David Miller]
1093 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1095 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1097 *) Support for new and upcoming Intel processors, including AVX2,
1098 BMI and SHA ISA extensions. This includes additional "stitched"
1099 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1102 This work was sponsored by Intel Corp.
1105 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1106 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1107 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1110 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1111 this fixes a limiation in previous versions of OpenSSL.
1114 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1115 MGF1 digest and OAEP label.
1118 *) Add EVP support for key wrapping algorithms, to avoid problems with
1119 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1120 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1121 algorithms and include tests cases.
1124 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1126 [Douglas E. Engert, Steve Henson]
1128 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1129 difference in days and seconds between two tm or ASN1_TIME structures.
1132 *) Add -rev test option to s_server to just reverse order of characters
1133 received by client and send back to server. Also prints an abbreviated
1134 summary of the connection parameters.
1137 *) New option -brief for s_client and s_server to print out a brief summary
1138 of connection parameters.
1141 *) Add callbacks for arbitrary TLS extensions.
1142 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1144 *) New option -crl_download in several openssl utilities to download CRLs
1145 from CRLDP extension in certificates.
1148 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1151 *) New function X509_CRL_diff to generate a delta CRL from the difference
1152 of two full CRLs. Add support to "crl" utility.
1155 *) New functions to set lookup_crls function and to retrieve
1156 X509_STORE from X509_STORE_CTX.
1159 *) Print out deprecated issuer and subject unique ID fields in
1163 *) Extend OCSP I/O functions so they can be used for simple general purpose
1164 HTTP as well as OCSP. New wrapper function which can be used to download
1165 CRLs using the OCSP API.
1168 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1171 *) SSL_CONF* functions. These provide a common framework for application
1172 configuration using configuration files or command lines.
1175 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1176 message callback and prints the results. Needs compile time option
1177 "enable-ssl-trace". New options to s_client and s_server to enable
1181 *) New ctrl and macro to retrieve supported points extensions.
1182 Print out extension in s_server and s_client.
1185 *) New functions to retrieve certificate signature and signature
1189 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1193 *) New Suite B modes for TLS code. These use and enforce the requirements
1194 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1195 only use Suite B curves. The Suite B modes can be set by using the
1196 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1199 *) New chain verification flags for Suite B levels of security. Check
1200 algorithms are acceptable when flags are set in X509_verify_cert.
1203 *) Make tls1_check_chain return a set of flags indicating checks passed
1204 by a certificate chain. Add additional tests to handle client
1205 certificates: checks for matching certificate type and issuer name
1209 *) If an attempt is made to use a signature algorithm not in the peer
1210 preference list abort the handshake. If client has no suitable
1211 signature algorithms in response to a certificate request do not
1212 use the certificate.
1215 *) If server EC tmp key is not in client preference list abort handshake.
1218 *) Add support for certificate stores in CERT structure. This makes it
1219 possible to have different stores per SSL structure or one store in
1220 the parent SSL_CTX. Include distint stores for certificate chain
1221 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1222 to build and store a certificate chain in CERT structure: returing
1223 an error if the chain cannot be built: this will allow applications
1224 to test if a chain is correctly configured.
1226 Note: if the CERT based stores are not set then the parent SSL_CTX
1227 store is used to retain compatibility with existing behaviour.
1231 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1232 mask based on the current session, check mask when sending client
1233 hello and checking the requested ciphersuite.
1236 *) New ctrls to retrieve and set certificate types in a certificate
1237 request message. Print out received values in s_client. If certificate
1238 types is not set with custom values set sensible values based on
1239 supported signature algorithms.
1242 *) Support for distinct client and server supported signature algorithms.
1245 *) Add certificate callback. If set this is called whenever a certificate
1246 is required by client or server. An application can decide which
1247 certificate chain to present based on arbitrary criteria: for example
1248 supported signature algorithms. Add very simple example to s_server.
1249 This fixes many of the problems and restrictions of the existing client
1250 certificate callback: for example you can now clear an existing
1251 certificate and specify the whole chain.
1254 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1255 the certificate can be used for (if anything). Set valid_flags field
1256 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1257 to have similar checks in it.
1259 Add new "cert_flags" field to CERT structure and include a "strict mode".
1260 This enforces some TLS certificate requirements (such as only permitting
1261 certificate signature algorithms contained in the supported algorithms
1262 extension) which some implementations ignore: this option should be used
1263 with caution as it could cause interoperability issues.
1266 *) Update and tidy signature algorithm extension processing. Work out
1267 shared signature algorithms based on preferences and peer algorithms
1268 and print them out in s_client and s_server. Abort handshake if no
1269 shared signature algorithms.
1272 *) Add new functions to allow customised supported signature algorithms
1273 for SSL and SSL_CTX structures. Add options to s_client and s_server
1277 *) New function SSL_certs_clear() to delete all references to certificates
1278 from an SSL structure. Before this once a certificate had been added
1279 it couldn't be removed.
1282 *) Integrate hostname, email address and IP address checking with certificate
1283 verification. New verify options supporting checking in opensl utility.
1286 *) Fixes and wildcard matching support to hostname and email checking
1287 functions. Add manual page.
1288 [Florian Weimer (Red Hat Product Security Team)]
1290 *) New functions to check a hostname email or IP address against a
1291 certificate. Add options x509 utility to print results of checks against
1295 *) Fix OCSP checking.
1296 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1298 *) Initial experimental support for explicitly trusted non-root CAs.
1299 OpenSSL still tries to build a complete chain to a root but if an
1300 intermediate CA has a trust setting included that is used. The first
1301 setting is used: whether to trust (e.g., -addtrust option to the x509
1305 *) Add -trusted_first option which attempts to find certificates in the
1306 trusted store even if an untrusted chain is also supplied.
1309 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1310 platform support for Linux and Android.
1313 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1316 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1317 When in FIPS mode the approved implementations are used as normal,
1318 when not in FIPS mode the internal unapproved versions are used instead.
1319 This means that the FIPS capable OpenSSL isn't forced to use the
1320 (often lower perfomance) FIPS implementations outside FIPS mode.
1323 *) Transparently support X9.42 DH parameters when calling
1324 PEM_read_bio_DHparameters. This means existing applications can handle
1325 the new parameter format automatically.
1328 *) Initial experimental support for X9.42 DH parameter format: mainly
1329 to support use of 'q' parameter for RFC5114 parameters.
1332 *) Add DH parameters from RFC5114 including test data to dhtest.
1335 *) Support for automatic EC temporary key parameter selection. If enabled
1336 the most preferred EC parameters are automatically used instead of
1337 hardcoded fixed parameters. Now a server just has to call:
1338 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1339 support ECDH and use the most appropriate parameters.
1342 *) Enhance and tidy EC curve and point format TLS extension code. Use
1343 static structures instead of allocation if default values are used.
1344 New ctrls to set curves we wish to support and to retrieve shared curves.
1345 Print out shared curves in s_server. New options to s_server and s_client
1346 to set list of supported curves.
1349 *) New ctrls to retrieve supported signature algorithms and
1350 supported curve values as an array of NIDs. Extend openssl utility
1351 to print out received values.
1354 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1355 between NIDs and the more common NIST names such as "P-256". Enhance
1356 ecparam utility and ECC method to recognise the NIST names for curves.
1359 *) Enhance SSL/TLS certificate chain handling to support different
1360 chains for each certificate instead of one chain in the parent SSL_CTX.
1363 *) Support for fixed DH ciphersuite client authentication: where both
1364 server and client use DH certificates with common parameters.
1367 *) Support for fixed DH ciphersuites: those requiring DH server
1371 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1373 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1374 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1375 X509_CINF_get_signature were reverted post internal team review.
1377 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1379 *) Build fixes for the Windows and OpenVMS platforms
1380 [Matt Caswell and Richard Levitte]
1382 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1384 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1385 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1386 dereference. This could lead to a Denial Of Service attack. Thanks to
1387 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1391 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1392 dtls1_buffer_record function under certain conditions. In particular this
1393 could occur if an attacker sent repeated DTLS records with the same
1394 sequence number but for the next epoch. The memory leak could be exploited
1395 by an attacker in a Denial of Service attack through memory exhaustion.
1396 Thanks to Chris Mueller for reporting this issue.
1400 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1401 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1402 method would be set to NULL which could later result in a NULL pointer
1403 dereference. Thanks to Frank Schmirler for reporting this issue.
1407 *) Abort handshake if server key exchange message is omitted for ephemeral
1410 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1411 reporting this issue.
1415 *) Remove non-export ephemeral RSA code on client and server. This code
1416 violated the TLS standard by allowing the use of temporary RSA keys in
1417 non-export ciphersuites and could be used by a server to effectively
1418 downgrade the RSA key length used to a value smaller than the server
1419 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1420 INRIA or reporting this issue.
1424 *) Fixed issue where DH client certificates are accepted without verification.
1425 An OpenSSL server will accept a DH certificate for client authentication
1426 without the certificate verify message. This effectively allows a client to
1427 authenticate without the use of a private key. This only affects servers
1428 which trust a client certificate authority which issues certificates
1429 containing DH keys: these are extremely rare and hardly ever encountered.
1430 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1435 *) Ensure that the session ID context of an SSL is updated when its
1436 SSL_CTX is updated via SSL_set_SSL_CTX.
1438 The session ID context is typically set from the parent SSL_CTX,
1439 and can vary with the CTX.
1442 *) Fix various certificate fingerprint issues.
1444 By using non-DER or invalid encodings outside the signed portion of a
1445 certificate the fingerprint can be changed without breaking the signature.
1446 Although no details of the signed portion of the certificate can be changed
1447 this can cause problems with some applications: e.g. those using the
1448 certificate fingerprint for blacklists.
1450 1. Reject signatures with non zero unused bits.
1452 If the BIT STRING containing the signature has non zero unused bits reject
1453 the signature. All current signature algorithms require zero unused bits.
1455 2. Check certificate algorithm consistency.
1457 Check the AlgorithmIdentifier inside TBS matches the one in the
1458 certificate signature. NB: this will result in signature failure
1459 errors for some broken certificates.
1461 Thanks to Konrad Kraszewski from Google for reporting this issue.
1463 3. Check DSA/ECDSA signatures use DER.
1465 Reencode DSA/ECDSA signatures and compare with the original received
1466 signature. Return an error if there is a mismatch.
1468 This will reject various cases including garbage after signature
1469 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1470 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1471 (negative or with leading zeroes).
1473 Further analysis was conducted and fixes were developed by Stephen Henson
1474 of the OpenSSL core team.
1479 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1480 results on some platforms, including x86_64. This bug occurs at random
1481 with a very low probability, and is not known to be exploitable in any
1482 way, though its exact impact is difficult to determine. Thanks to Pieter
1483 Wuille (Blockstream) who reported this issue and also suggested an initial
1484 fix. Further analysis was conducted by the OpenSSL development team and
1485 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1486 the OpenSSL core team.
1490 *) Do not resume sessions on the server if the negotiated protocol
1491 version does not match the session's version. Resuming with a different
1492 version, while not strictly forbidden by the RFC, is of questionable
1493 sanity and breaks all known clients.
1494 [David Benjamin, Emilia Käsper]
1496 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1497 early CCS messages during renegotiation. (Note that because
1498 renegotiation is encrypted, this early CCS was not exploitable.)
1501 *) Tighten client-side session ticket handling during renegotiation:
1502 ensure that the client only accepts a session ticket if the server sends
1503 the extension anew in the ServerHello. Previously, a TLS client would
1504 reuse the old extension state and thus accept a session ticket if one was
1505 announced in the initial ServerHello.
1507 Similarly, ensure that the client requires a session ticket if one
1508 was advertised in the ServerHello. Previously, a TLS client would
1509 ignore a missing NewSessionTicket message.
1512 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1514 *) SRTP Memory Leak.
1516 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1517 sends a carefully crafted handshake message, to cause OpenSSL to fail
1518 to free up to 64k of memory causing a memory leak. This could be
1519 exploited in a Denial Of Service attack. This issue affects OpenSSL
1520 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1521 whether SRTP is used or configured. Implementations of OpenSSL that
1522 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1524 The fix was developed by the OpenSSL team.
1528 *) Session Ticket Memory Leak.
1530 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1531 integrity of that ticket is first verified. In the event of a session
1532 ticket integrity check failing, OpenSSL will fail to free memory
1533 causing a memory leak. By sending a large number of invalid session
1534 tickets an attacker could exploit this issue in a Denial Of Service
1539 *) Build option no-ssl3 is incomplete.
1541 When OpenSSL is configured with "no-ssl3" as a build option, servers
1542 could accept and complete a SSL 3.0 handshake, and clients could be
1543 configured to send them.
1545 [Akamai and the OpenSSL team]
1547 *) Add support for TLS_FALLBACK_SCSV.
1548 Client applications doing fallback retries should call
1549 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1551 [Adam Langley, Bodo Moeller]
1553 *) Add additional DigestInfo checks.
1555 Reencode DigestInto in DER and check against the original when
1556 verifying RSA signature: this will reject any improperly encoded
1557 DigestInfo structures.
1559 Note: this is a precautionary measure and no attacks are currently known.
1563 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1565 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1566 SRP code can be overrun an internal buffer. Add sanity check that
1567 g, A, B < N to SRP code.
1569 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1570 Group for discovering this issue.
1574 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1575 TLS 1.0 instead of higher protocol versions when the ClientHello message
1576 is badly fragmented. This allows a man-in-the-middle attacker to force a
1577 downgrade to TLS 1.0 even if both the server and the client support a
1578 higher protocol version, by modifying the client's TLS records.
1580 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1581 researching this issue.
1585 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1586 to a denial of service attack. A malicious server can crash the client
1587 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1588 ciphersuite and sending carefully crafted handshake messages.
1590 Thanks to Felix Gröbert (Google) for discovering and researching this
1595 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1596 to leak memory. This can be exploited through a Denial of Service attack.
1597 Thanks to Adam Langley for discovering and researching this issue.
1601 *) An attacker can force openssl to consume large amounts of memory whilst
1602 processing DTLS handshake messages. This can be exploited through a
1603 Denial of Service attack.
1604 Thanks to Adam Langley for discovering and researching this issue.
1608 *) An attacker can force an error condition which causes openssl to crash
1609 whilst processing DTLS packets due to memory being freed twice. This
1610 can be exploited through a Denial of Service attack.
1611 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1616 *) If a multithreaded client connects to a malicious server using a resumed
1617 session and the server sends an ec point format extension it could write
1618 up to 255 bytes to freed memory.
1620 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1625 *) A malicious server can crash an OpenSSL client with a null pointer
1626 dereference (read) by specifying an SRP ciphersuite even though it was not
1627 properly negotiated with the client. This can be exploited through a
1628 Denial of Service attack.
1630 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1631 discovering and researching this issue.
1635 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1636 X509_name_oneline, X509_name_print_ex et al. to leak some information
1637 from the stack. Applications may be affected if they echo pretty printing
1638 output to the attacker.
1640 Thanks to Ivan Fratric (Google) for discovering this issue.
1642 [Emilia Käsper, and Steve Henson]
1644 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1645 for corner cases. (Certain input points at infinity could lead to
1646 bogus results, with non-infinity inputs mapped to infinity too.)
1649 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1651 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1652 handshake can force the use of weak keying material in OpenSSL
1653 SSL/TLS clients and servers.
1655 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1656 researching this issue. (CVE-2014-0224)
1657 [KIKUCHI Masashi, Steve Henson]
1659 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1660 OpenSSL DTLS client the code can be made to recurse eventually crashing
1663 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1665 [Imre Rad, Steve Henson]
1667 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1668 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1669 client or server. This is potentially exploitable to run arbitrary
1670 code on a vulnerable client or server.
1672 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1673 [Jüri Aedla, Steve Henson]
1675 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1676 are subject to a denial of service attack.
1678 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1679 this issue. (CVE-2014-3470)
1680 [Felix Gröbert, Ivan Fratric, Steve Henson]
1682 *) Harmonize version and its documentation. -f flag is used to display
1684 [mancha <mancha1@zoho.com>]
1686 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1687 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1688 [mancha <mancha1@zoho.com>]
1690 *) Fix some double frees. These are not thought to be exploitable.
1691 [mancha <mancha1@zoho.com>]
1693 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1695 *) A missing bounds check in the handling of the TLS heartbeat extension
1696 can be used to reveal up to 64k of memory to a connected client or
1699 Thanks for Neel Mehta of Google Security for discovering this bug and to
1700 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1701 preparing the fix (CVE-2014-0160)
1702 [Adam Langley, Bodo Moeller]
1704 *) Fix for the attack described in the paper "Recovering OpenSSL
1705 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1706 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1707 http://eprint.iacr.org/2014/140
1709 Thanks to Yuval Yarom and Naomi Benger for discovering this
1710 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1711 [Yuval Yarom and Naomi Benger]
1713 *) TLS pad extension: draft-agl-tls-padding-03
1715 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1716 TLS client Hello record length value would otherwise be > 255 and
1717 less that 512 pad with a dummy extension containing zeroes so it
1718 is at least 512 bytes long.
1720 [Adam Langley, Steve Henson]
1722 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1724 *) Fix for TLS record tampering bug. A carefully crafted invalid
1725 handshake could crash OpenSSL with a NULL pointer exception.
1726 Thanks to Anton Johansson for reporting this issues.
1729 *) Keep original DTLS digest and encryption contexts in retransmission
1730 structures so we can use the previous session parameters if they need
1731 to be resent. (CVE-2013-6450)
1734 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1735 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1736 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1737 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1738 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1739 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1740 [Rob Stradling, Adam Langley]
1742 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1744 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1745 supporting platforms or when small records were transferred.
1746 [Andy Polyakov, Steve Henson]
1748 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1750 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1752 This addresses the flaw in CBC record processing discovered by
1753 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1754 at: http://www.isg.rhul.ac.uk/tls/
1756 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1757 Security Group at Royal Holloway, University of London
1758 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1759 Emilia Käsper for the initial patch.
1761 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1763 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1764 ciphersuites which can be exploited in a denial of service attack.
1765 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1766 and detecting this bug and to Wolfgang Ettlinger
1767 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1771 *) Return an error when checking OCSP signatures when key is NULL.
1772 This fixes a DoS attack. (CVE-2013-0166)
1775 *) Make openssl verify return errors.
1776 [Chris Palmer <palmer@google.com> and Ben Laurie]
1778 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1779 the right response is stapled. Also change SSL_get_certificate()
1780 so it returns the certificate actually sent.
1781 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1782 [Rob Stradling <rob.stradling@comodo.com>]
1784 *) Fix possible deadlock when decoding public keys.
1787 *) Don't use TLS 1.0 record version number in initial client hello
1791 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1793 *) Sanity check record length before skipping explicit IV in TLS
1794 1.2, 1.1 and DTLS to fix DoS attack.
1796 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1797 fuzzing as a service testing platform.
1801 *) Initialise tkeylen properly when encrypting CMS messages.
1802 Thanks to Solar Designer of Openwall for reporting this issue.
1805 *) In FIPS mode don't try to use composite ciphers as they are not
1809 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1811 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1812 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1813 mean any application compiled against OpenSSL 1.0.0 headers setting
1814 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1815 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1816 0x10000000L Any application which was previously compiled against
1817 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1818 will need to be recompiled as a result. Letting be results in
1819 inability to disable specifically TLS 1.1 and in client context,
1820 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1823 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1824 disable just protocol X, but all protocols above X *if* there are
1825 protocols *below* X still enabled. In more practical terms it means
1826 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1827 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1828 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1832 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1834 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1835 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1836 in CRYPTO_realloc_clean.
1838 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1839 issue and to Adam Langley <agl@chromium.org> for fixing it.
1841 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1843 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1846 *) Workarounds for some broken servers that "hang" if a client hello
1847 record length exceeds 255 bytes.
1849 1. Do not use record version number > TLS 1.0 in initial client
1850 hello: some (but not all) hanging servers will now work.
1851 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1852 the number of ciphers sent in the client hello. This should be
1853 set to an even number, such as 50, for example by passing:
1854 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1855 Most broken servers should now work.
1856 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1857 TLS 1.2 client support entirely.
1860 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1863 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1865 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1866 STRING form instead of a DigestInfo.
1869 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1870 and the RSA_sign/RSA_verify functions. This was made more apparent when
1871 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1872 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1873 the correct format in RSA_verify so both forms transparently work.
1876 *) Some servers which support TLS 1.0 can choke if we initially indicate
1877 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1878 encrypted premaster secret. As a workaround use the maximum pemitted
1879 client version in client hello, this should keep such servers happy
1880 and still work with previous versions of OpenSSL.
1883 *) Add support for TLS/DTLS heartbeats.
1884 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1886 *) Add support for SCTP.
1887 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1889 *) Improved PRNG seeding for VOS.
1890 [Paul Green <Paul.Green@stratus.com>]
1892 *) Extensive assembler packs updates, most notably:
1894 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1895 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1896 - x86_64: bit-sliced AES implementation;
1897 - ARM: NEON support, contemporary platforms optimizations;
1898 - s390x: z196 support;
1899 - *: GHASH and GF(2^m) multiplication implementations;
1903 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1904 (removal of unnecessary code)
1905 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1907 *) Add TLS key material exporter from RFC 5705.
1910 *) Add DTLS-SRTP negotiation from RFC 5764.
1913 *) Add Next Protocol Negotiation,
1914 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1915 disabled with a no-npn flag to config or Configure. Code donated
1917 [Adam Langley <agl@google.com> and Ben Laurie]
1919 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1920 NIST-P256, NIST-P521, with constant-time single point multiplication on
1921 typical inputs. Compiler support for the nonstandard type __uint128_t is
1922 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1923 Code made available under Apache License version 2.0.
1925 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1926 line to include this in your build of OpenSSL, and run "make depend" (or
1927 "make update"). This enables the following EC_METHODs:
1929 EC_GFp_nistp224_method()
1930 EC_GFp_nistp256_method()
1931 EC_GFp_nistp521_method()
1933 EC_GROUP_new_by_curve_name() will automatically use these (while
1934 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1936 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1938 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1939 all platforms. Move ssize_t definition from e_os.h to the public
1940 header file e_os2.h as it now appears in public header file cms.h
1943 *) New -sigopt option to the ca, req and x509 utilities. Additional
1944 signature parameters can be passed using this option and in
1948 *) Add RSA PSS signing function. This will generate and set the
1949 appropriate AlgorithmIdentifiers for PSS based on those in the
1950 corresponding EVP_MD_CTX structure. No application support yet.
1953 *) Support for companion algorithm specific ASN1 signing routines.
1954 New function ASN1_item_sign_ctx() signs a pre-initialised
1955 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1956 the appropriate parameters.
1959 *) Add new algorithm specific ASN1 verification initialisation function
1960 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1961 handling will be the same no matter what EVP_PKEY_METHOD is used.
1962 Add a PSS handler to support verification of PSS signatures: checked
1963 against a number of sample certificates.
1966 *) Add signature printing for PSS. Add PSS OIDs.
1967 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1969 *) Add algorithm specific signature printing. An individual ASN1 method
1970 can now print out signatures instead of the standard hex dump.
1972 More complex signatures (e.g. PSS) can print out more meaningful
1973 information. Include DSA version that prints out the signature
1977 *) Password based recipient info support for CMS library: implementing
1981 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1982 neatly separates the code into cipher and PBE sections and is required
1983 for some algorithms that split PBES2 into separate pieces (such as
1984 password based CMS).
1987 *) Session-handling fixes:
1988 - Fix handling of connections that are resuming with a session ID,
1989 but also support Session Tickets.
1990 - Fix a bug that suppressed issuing of a new ticket if the client
1991 presented a ticket with an expired session.
1992 - Try to set the ticket lifetime hint to something reasonable.
1993 - Make tickets shorter by excluding irrelevant information.
1994 - On the client side, don't ignore renewed tickets.
1995 [Adam Langley, Bodo Moeller (Google)]
1997 *) Fix PSK session representation.
2000 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2002 This work was sponsored by Intel.
2005 *) Add GCM support to TLS library. Some custom code is needed to split
2006 the IV between the fixed (from PRF) and explicit (from TLS record)
2007 portions. This adds all GCM ciphersuites supported by RFC5288 and
2008 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
2009 add a special AESGCM string for GCM only.
2012 *) Expand range of ctrls for AES GCM. Permit setting invocation
2013 field on decrypt and retrieval of invocation field only on encrypt.
2016 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2017 As required by RFC5289 these ciphersuites cannot be used if for
2018 versions of TLS earlier than 1.2.
2021 *) For FIPS capable OpenSSL interpret a NULL default public key method
2022 as unset and return the appopriate default but do *not* set the default.
2023 This means we can return the appopriate method in applications that
2024 swicth between FIPS and non-FIPS modes.
2027 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2028 ENGINE is used then we cannot handle that in the FIPS module so we
2029 keep original code iff non-FIPS operations are allowed.
2032 *) Add -attime option to openssl utilities.
2033 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2035 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2038 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2039 FIPS EC methods unconditionally for now.
2042 *) New build option no-ec2m to disable characteristic 2 code.
2045 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2046 all cases can be covered as some introduce binary incompatibilities.
2049 *) Redirect RSA operations to FIPS module including keygen,
2050 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2053 *) Add similar low level API blocking to ciphers.
2056 *) Low level digest APIs are not approved in FIPS mode: any attempt
2057 to use these will cause a fatal error. Applications that *really* want
2058 to use them can use the private_* version instead.
2061 *) Redirect cipher operations to FIPS module for FIPS builds.
2064 *) Redirect digest operations to FIPS module for FIPS builds.
2067 *) Update build system to add "fips" flag which will link in fipscanister.o
2068 for static and shared library builds embedding a signature if needed.
2071 *) Output TLS supported curves in preference order instead of numerical
2072 order. This is currently hardcoded for the highest order curves first.
2073 This should be configurable so applications can judge speed vs strength.
2076 *) Add TLS v1.2 server support for client authentication.
2079 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2083 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2084 FIPS modules versions.
2087 *) Add TLS v1.2 client side support for client authentication. Keep cache
2088 of handshake records longer as we don't know the hash algorithm to use
2089 until after the certificate request message is received.
2092 *) Initial TLS v1.2 client support. Add a default signature algorithms
2093 extension including all the algorithms we support. Parse new signature
2094 format in client key exchange. Relax some ECC signing restrictions for
2095 TLS v1.2 as indicated in RFC5246.
2098 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2099 to new signature format when needed using client digest preference.
2100 All server ciphersuites should now work correctly in TLS v1.2. No client
2101 support yet and no support for client certificates.
2104 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2105 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2106 ciphersuites. At present only RSA key exchange ciphersuites work with
2107 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2108 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2109 and version checking.
2112 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2113 with this defined it will not be affected by any changes to ssl internal
2114 structures. Add several utility functions to allow openssl application
2115 to work with OPENSSL_NO_SSL_INTERN defined.
2118 *) A long standing patch to add support for SRP from EdelWeb (Peter
2119 Sylvester and Christophe Renou) was integrated.
2120 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
2121 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
2124 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2127 *) Permit abbreviated handshakes when renegotiating using the function
2128 SSL_renegotiate_abbreviated().
2129 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2131 *) Add call to ENGINE_register_all_complete() to
2132 ENGINE_load_builtin_engines(), so some implementations get used
2133 automatically instead of needing explicit application support.
2136 *) Add support for TLS key exporter as described in RFC5705.
2137 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2139 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2140 a few changes are required:
2142 Add SSL_OP_NO_TLSv1_1 flag.
2143 Add TLSv1_1 methods.
2144 Update version checking logic to handle version 1.1.
2145 Add explicit IV handling (ported from DTLS code).
2146 Add command line options to s_client/s_server.
2149 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2151 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2152 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2153 content decryption and always return the same error. Note: this attack
2154 needs on average 2^20 messages so it only affects automated senders. The
2155 old behaviour can be reenabled in the CMS code by setting the
2156 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2157 an MMA defence is not necessary.
2158 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2159 this issue. (CVE-2012-0884)
2162 *) Fix CVE-2011-4619: make sure we really are receiving a
2163 client hello before rejecting multiple SGC restarts. Thanks to
2164 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2167 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2169 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2170 Thanks to Antonio Martin, Enterprise Secure Access Research and
2171 Development, Cisco Systems, Inc. for discovering this bug and
2172 preparing a fix. (CVE-2012-0050)
2175 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2177 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2178 of the Vaudenay padding oracle attack on CBC mode encryption
2179 which enables an efficient plaintext recovery attack against
2180 the OpenSSL implementation of DTLS. Their attack exploits timing
2181 differences arising during decryption processing. A research
2182 paper describing this attack can be found at:
2183 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2184 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2185 Security Group at Royal Holloway, University of London
2186 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2187 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2188 for preparing the fix. (CVE-2011-4108)
2189 [Robin Seggelmann, Michael Tuexen]
2191 *) Clear bytes used for block padding of SSL 3.0 records.
2193 [Adam Langley (Google)]
2195 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2196 Kadianakis <desnacked@gmail.com> for discovering this issue and
2197 Adam Langley for preparing the fix. (CVE-2011-4619)
2198 [Adam Langley (Google)]
2200 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2201 [Andrey Kulikov <amdeich@gmail.com>]
2203 *) Prevent malformed RFC3779 data triggering an assertion failure.
2204 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2205 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2206 [Rob Austein <sra@hactrn.net>]
2208 *) Improved PRNG seeding for VOS.
2209 [Paul Green <Paul.Green@stratus.com>]
2211 *) Fix ssl_ciph.c set-up race.
2212 [Adam Langley (Google)]
2214 *) Fix spurious failures in ecdsatest.c.
2215 [Emilia Käsper (Google)]
2217 *) Fix the BIO_f_buffer() implementation (which was mixing different
2218 interpretations of the '..._len' fields).
2219 [Adam Langley (Google)]
2221 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2222 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2223 threads won't reuse the same blinding coefficients.
2225 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2226 lock to call BN_BLINDING_invert_ex, and avoids one use of
2227 BN_BLINDING_update for each BN_BLINDING structure (previously,
2228 the last update always remained unused).
2229 [Emilia Käsper (Google)]
2231 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2232 [Bob Buckholz (Google)]
2234 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2236 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2237 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2238 [Kaspar Brand <ossl@velox.ch>]
2240 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2241 for multi-threaded use of ECDH. (CVE-2011-3210)
2242 [Adam Langley (Google)]
2244 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2247 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2248 signature public key algorithm by using OID xref utilities instead.
2249 Before this you could only use some ECC ciphersuites with SHA1 only.
2252 *) Add protection against ECDSA timing attacks as mentioned in the paper
2253 by Billy Bob Brumley and Nicola Tuveri, see:
2255 http://eprint.iacr.org/2011/232.pdf
2257 [Billy Bob Brumley and Nicola Tuveri]
2259 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2261 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2262 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2264 *) Fix bug in string printing code: if *any* escaping is enabled we must
2265 escape the escape character (backslash) or the resulting string is
2269 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2271 *) Disable code workaround for ancient and obsolete Netscape browsers
2272 and servers: an attacker can use it in a ciphersuite downgrade attack.
2273 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2276 *) Fixed J-PAKE implementation error, originally discovered by
2277 Sebastien Martini, further info and confirmation from Stefan
2278 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2281 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2283 *) Fix extension code to avoid race conditions which can result in a buffer
2284 overrun vulnerability: resumed sessions must not be modified as they can
2285 be shared by multiple threads. CVE-2010-3864
2288 *) Fix WIN32 build system to correctly link an ENGINE directory into
2292 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2294 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2296 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2298 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2300 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2301 context. The operation can be customised via the ctrl mechanism in
2302 case ENGINEs want to include additional functionality.
2305 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2308 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2309 output hashes compatible with older versions of OpenSSL.
2310 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2312 *) Fix compression algorithm handling: if resuming a session use the
2313 compression algorithm of the resumed session instead of determining
2314 it from client hello again. Don't allow server to change algorithm.
2317 *) Add load_crls() function to apps tidying load_certs() too. Add option
2318 to verify utility to allow additional CRLs to be included.
2321 *) Update OCSP request code to permit adding custom headers to the request:
2322 some responders need this.
2325 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2327 [Julia Lawall <julia@diku.dk>]
2329 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2330 needlessly dereferenced structures, used obsolete functions and
2331 didn't handle all updated verify codes correctly.
2334 *) Disable MD2 in the default configuration.
2337 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2338 indicate the initial BIO being pushed or popped. This makes it possible
2339 to determine whether the BIO is the one explicitly called or as a result
2340 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2341 it handles reference counts correctly and doesn't zero out the I/O bio
2342 when it is not being explicitly popped. WARNING: applications which
2343 included workarounds for the old buggy behaviour will need to be modified
2344 or they could free up already freed BIOs.
2347 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2348 renaming to all platforms (within the 0.9.8 branch, this was
2349 done conditionally on Netware platforms to avoid a name clash).
2350 [Guenter <lists@gknw.net>]
2352 *) Add ECDHE and PSK support to DTLS.
2353 [Michael Tuexen <tuexen@fh-muenster.de>]
2355 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2359 *) Add "missing" function EVP_MD_flags() (without this the only way to
2360 retrieve a digest flags is by accessing the structure directly. Update
2361 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2362 or cipher is registered as in the "from" argument. Print out all
2363 registered digests in the dgst usage message instead of manually
2364 attempting to work them out.
2367 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2368 this allows the use of compression and extensions. Change default cipher
2369 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2370 by default unless an application cipher string requests it.
2373 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2374 key ids to find matching certificates and keys but some PKCS#12 files
2375 don't follow the (somewhat unwritten) rules and this strategy fails.
2376 Now just gather all certificates together and the first private key
2377 then look for the first certificate that matches the key.
2380 *) Support use of registered digest and cipher names for dgst and cipher
2381 commands instead of having to add each one as a special case. So now
2388 openssl dgst -sha256 foo
2390 and this works for ENGINE based algorithms too.
2394 *) Update Gost ENGINE to support parameter files.
2395 [Victor B. Wagner <vitus@cryptocom.ru>]
2397 *) Support GeneralizedTime in ca utility.
2398 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2400 *) Enhance the hash format used for certificate directory links. The new
2401 form uses the canonical encoding (meaning equivalent names will work
2402 even if they aren't identical) and uses SHA1 instead of MD5. This form
2403 is incompatible with the older format and as a result c_rehash should
2404 be used to rebuild symbolic links.
2407 *) Make PKCS#8 the default write format for private keys, replacing the
2408 traditional format. This form is standardised, more secure and doesn't
2409 include an implicit MD5 dependency.
2412 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2413 committed to OpenSSL should pass this lot as a minimum.
2416 *) Add session ticket override functionality for use by EAP-FAST.
2417 [Jouni Malinen <j@w1.fi>]
2419 *) Modify HMAC functions to return a value. Since these can be implemented
2420 in an ENGINE errors can occur.
2423 *) Type-checked OBJ_bsearch_ex.
2426 *) Type-checked OBJ_bsearch. Also some constification necessitated
2427 by type-checking. Still to come: TXT_DB, bsearch(?),
2428 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2432 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2433 seconds to a tm structure directly, instead of going through OS
2434 specific date routines. This avoids any issues with OS routines such
2435 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2436 and X509_time_adj_ex() to cover the extended range. The existing
2437 X509_time_adj() is still usable and will no longer have any date issues.
2440 *) Delta CRL support. New use deltas option which will attempt to locate
2441 and search any appropriate delta CRLs available.
2443 This work was sponsored by Google.
2446 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2447 code and add additional score elements. Validate alternate CRL paths
2448 as part of the CRL checking and indicate a new error "CRL path validation
2449 error" in this case. Applications wanting additional details can use
2450 the verify callback and check the new "parent" field. If this is not
2451 NULL CRL path validation is taking place. Existing applications wont
2452 see this because it requires extended CRL support which is off by
2455 This work was sponsored by Google.
2458 *) Support for freshest CRL extension.
2460 This work was sponsored by Google.
2463 *) Initial indirect CRL support. Currently only supported in the CRLs
2464 passed directly and not via lookup. Process certificate issuer
2465 CRL entry extension and lookup CRL entries by bother issuer name
2466 and serial number. Check and process CRL issuer entry in IDP extension.
2468 This work was sponsored by Google.
2471 *) Add support for distinct certificate and CRL paths. The CRL issuer
2472 certificate is validated separately in this case. Only enabled if
2473 an extended CRL support flag is set: this flag will enable additional
2474 CRL functionality in future.
2476 This work was sponsored by Google.
2479 *) Add support for policy mappings extension.
2481 This work was sponsored by Google.
2484 *) Fixes to pathlength constraint, self issued certificate handling,
2485 policy processing to align with RFC3280 and PKITS tests.
2487 This work was sponsored by Google.
2490 *) Support for name constraints certificate extension. DN, email, DNS
2491 and URI types are currently supported.
2493 This work was sponsored by Google.
2496 *) To cater for systems that provide a pointer-based thread ID rather
2497 than numeric, deprecate the current numeric thread ID mechanism and
2498 replace it with a structure and associated callback type. This
2499 mechanism allows a numeric "hash" to be extracted from a thread ID in
2500 either case, and on platforms where pointers are larger than 'long',
2501 mixing is done to help ensure the numeric 'hash' is usable even if it
2502 can't be guaranteed unique. The default mechanism is to use "&errno"
2503 as a pointer-based thread ID to distinguish between threads.
2505 Applications that want to provide their own thread IDs should now use
2506 CRYPTO_THREADID_set_callback() to register a callback that will call
2507 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2509 Note that ERR_remove_state() is now deprecated, because it is tied
2510 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2511 to free the current thread's error state should be replaced by
2512 ERR_remove_thread_state(NULL).
2514 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2515 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2516 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2517 application was previously providing a numeric thread callback that
2518 was inappropriate for distinguishing threads, then uniqueness might
2519 have been obtained with &errno that happened immediately in the
2520 intermediate development versions of OpenSSL; this is no longer the
2521 case, the numeric thread callback will now override the automatic use
2523 [Geoff Thorpe, with help from Bodo Moeller]
2525 *) Initial support for different CRL issuing certificates. This covers a
2526 simple case where the self issued certificates in the chain exist and
2527 the real CRL issuer is higher in the existing chain.
2529 This work was sponsored by Google.
2532 *) Removed effectively defunct crypto/store from the build.
2535 *) Revamp of STACK to provide stronger type-checking. Still to come:
2536 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2537 ASN1_STRING, CONF_VALUE.
2540 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2541 RAM on SSL connections. This option can save about 34k per idle SSL.
2544 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2545 STACK, TXT_DB, bsearch, qsort.
2548 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2549 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2550 support for data, signedData, compressedData, digestedData and
2551 encryptedData, envelopedData types included. Scripts to check against
2552 RFC4134 examples draft and interop and consistency checks of many
2553 content types and variants.
2556 *) Add options to enc utility to support use of zlib compression BIO.
2559 *) Extend mk1mf to support importing of options and assembly language
2560 files from Configure script, currently only included in VC-WIN32.
2561 The assembly language rules can now optionally generate the source
2562 files from the associated perl scripts.
2565 *) Implement remaining functionality needed to support GOST ciphersuites.
2566 Interop testing has been performed using CryptoPro implementations.
2567 [Victor B. Wagner <vitus@cryptocom.ru>]
2569 *) s390x assembler pack.
2572 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2576 *) Implement Opaque PRF Input TLS extension as specified in
2577 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2578 official specification yet and no extension type assignment by
2579 IANA exists, this extension (for now) will have to be explicitly
2580 enabled when building OpenSSL by providing the extension number
2581 to use. For example, specify an option
2583 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2585 to the "config" or "Configure" script to enable the extension,
2586 assuming extension number 0x9527 (which is a completely arbitrary
2587 and unofficial assignment based on the MD5 hash of the Internet
2588 Draft). Note that by doing so, you potentially lose
2589 interoperability with other TLS implementations since these might
2590 be using the same extension number for other purposes.
2592 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2593 opaque PRF input value to use in the handshake. This will create
2594 an interal copy of the length-'len' string at 'src', and will
2595 return non-zero for success.
2597 To get more control and flexibility, provide a callback function
2600 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2601 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2605 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2608 Callback function 'cb' will be called in handshakes, and is
2609 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2610 Argument 'arg' is for application purposes (the value as given to
2611 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2612 be provided to the callback function). The callback function
2613 has to return non-zero to report success: usually 1 to use opaque
2614 PRF input just if possible, or 2 to enforce use of the opaque PRF
2615 input. In the latter case, the library will abort the handshake
2616 if opaque PRF input is not successfully negotiated.
2618 Arguments 'peerinput' and 'len' given to the callback function
2619 will always be NULL and 0 in the case of a client. A server will
2620 see the client's opaque PRF input through these variables if
2621 available (NULL and 0 otherwise). Note that if the server
2622 provides an opaque PRF input, the length must be the same as the
2623 length of the client's opaque PRF input.
2625 Note that the callback function will only be called when creating
2626 a new session (session resumption can resume whatever was
2627 previously negotiated), and will not be called in SSL 2.0
2628 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2629 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2630 for applications that need to enforce opaque PRF input.
2634 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2637 [Victor B. Wagner <vitus@cryptocom.ru>]
2639 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2640 RFC4507bis. The encrypted ticket format is an encrypted encoded
2641 SSL_SESSION structure, that way new session features are automatically
2644 If a client application caches session in an SSL_SESSION structure
2645 support is transparent because tickets are now stored in the encoded
2648 The SSL_CTX structure automatically generates keys for ticket
2649 protection in servers so again support should be possible
2650 with no application modification.
2652 If a client or server wishes to disable RFC4507 support then the option
2653 SSL_OP_NO_TICKET can be set.
2655 Add a TLS extension debugging callback to allow the contents of any client
2656 or server extensions to be examined.
2658 This work was sponsored by Google.
2661 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2662 OpenSSL should now compile cleanly on gcc 4.2
2663 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2665 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2666 support including streaming MAC support: this is required for GOST
2667 ciphersuite support.
2668 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2670 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2671 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2672 to output in BER and PEM format.
2675 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2676 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2677 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2678 ENGINE support for HMAC keys which are unextractable. New -mac and
2679 -macopt options to dgst utility.
2682 *) New option -sigopt to dgst utility. Update dgst to use
2683 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2684 alternative signing paramaters such as X9.31 or PSS in the dgst
2688 *) Change ssl_cipher_apply_rule(), the internal function that does
2689 the work each time a ciphersuite string requests enabling
2690 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2691 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2692 the order of disabled ciphersuites such that those ciphersuites
2693 that most recently went from enabled to disabled not only stay
2694 in order with respect to each other, but also have higher priority
2695 than other disabled ciphersuites the next time ciphersuites are
2698 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2699 the same ciphersuites as with "HIGH" alone, but in a specific
2700 order where the PSK ciphersuites come first (since they are the
2701 most recently disabled ciphersuites when "HIGH" is parsed).
2703 Also, change ssl_create_cipher_list() (using this new
2704 funcionality) such that between otherwise identical
2705 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2709 *) Change ssl_create_cipher_list() so that it automatically
2710 arranges the ciphersuites in reasonable order before starting
2711 to process the rule string. Thus, the definition for "DEFAULT"
2712 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2713 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2714 This makes it much easier to arrive at a reasonable default order
2715 in applications for which anonymous ciphers are OK (meaning
2716 that you can't actually use DEFAULT).
2717 [Bodo Moeller; suggested by Victor Duchovni]
2719 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2720 processing) into multiple integers instead of setting
2721 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2722 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2723 (These masks as well as the individual bit definitions are hidden
2724 away into the non-exported interface ssl/ssl_locl.h, so this
2725 change to the definition of the SSL_CIPHER structure shouldn't
2726 affect applications.) This give us more bits for each of these
2727 categories, so there is no longer a need to coagulate AES128 and
2728 AES256 into a single algorithm bit, and to coagulate Camellia128
2729 and Camellia256 into a single algorithm bit, which has led to all
2732 Thus, among other things, the kludge introduced in 0.9.7m and
2733 0.9.8e for masking out AES256 independently of AES128 or masking
2734 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2736 With the change, we also introduce new ciphersuite aliases that
2737 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2741 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2742 Use the leftmost N bytes of the signature input if the input is
2743 larger than the prime q (with N being the size in bytes of q).
2746 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2747 it yet and it is largely untested.
2750 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2753 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2754 some compilers (gcc 4.2 and later) reject their use. Safestack is
2755 reimplemented. Update ASN1 to avoid use of legacy functions.
2758 *) Win32/64 targets are linked with Winsock2.
2761 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2762 to external functions. This can be used to increase CRL handling
2763 efficiency especially when CRLs are very large by (for example) storing
2764 the CRL revoked certificates in a database.
2767 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2768 new CRLs added to a directory can be used. New command line option
2769 -verify_return_error to s_client and s_server. This causes real errors
2770 to be returned by the verify callback instead of carrying on no matter
2771 what. This reflects the way a "real world" verify callback would behave.
2774 *) GOST engine, supporting several GOST algorithms and public key formats.
2775 Kindly donated by Cryptocom.
2778 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2779 partitioned by DP are handled but no indirect CRL or reason partitioning
2780 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2781 selected via a scoring technique which handles IDP and AKID in CRLs.
2784 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2785 will ultimately be used for all verify operations: this will remove the
2786 X509_STORE dependency on certificate verification and allow alternative
2787 lookup methods. X509_STORE based implementations of these two callbacks.
2790 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2791 Modify get_crl() to find a valid (unexpired) CRL if possible.
2794 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2795 this would be called X509_CRL_cmp() but that name is already used by
2796 a function that just compares CRL issuer names. Cache several CRL
2797 extensions in X509_CRL structure and cache CRLDP in X509.
2800 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2801 this maps equivalent X509_NAME structures into a consistent structure.
2802 Name comparison can then be performed rapidly using memcmp().
2805 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2809 *) Allow digests to supply their own micalg string for S/MIME type using
2810 the ctrl EVP_MD_CTRL_MICALG.
2813 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2814 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2815 ctrl. It can then customise the structure before and/or after signing
2819 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2820 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2821 to free up any added signature OIDs.
2824 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2825 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2826 digest and cipher tables. New options added to openssl utility:
2827 list-message-digest-algorithms and list-cipher-algorithms.
2830 *) Change the array representation of binary polynomials: the list
2831 of degrees of non-zero coefficients is now terminated with -1.
2832 Previously it was terminated with 0, which was also part of the
2833 value; thus, the array representation was not applicable to
2834 polynomials where t^0 has coefficient zero. This change makes
2835 the array representation useful in a more general context.
2838 *) Various modifications and fixes to SSL/TLS cipher string
2839 handling. For ECC, the code now distinguishes between fixed ECDH
2840 with RSA certificates on the one hand and with ECDSA certificates
2841 on the other hand, since these are separate ciphersuites. The
2842 unused code for Fortezza ciphersuites has been removed.
2844 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2845 (not "ECDHE"). For consistency with the code for DH
2846 certificates, use of ECDH certificates is now considered ECDH
2847 authentication, not RSA or ECDSA authentication (the latter is
2848 merely the CA's signing algorithm and not actively used in the
2851 The temporary ciphersuite alias "ECCdraft" is no longer
2852 available, and ECC ciphersuites are no longer excluded from "ALL"
2853 and "DEFAULT". The following aliases now exist for RFC 4492
2854 ciphersuites, most of these by analogy with the DH case:
2856 kECDHr - ECDH cert, signed with RSA
2857 kECDHe - ECDH cert, signed with ECDSA
2858 kECDH - ECDH cert (signed with either RSA or ECDSA)
2859 kEECDH - ephemeral ECDH
2860 ECDH - ECDH cert or ephemeral ECDH
2866 AECDH - anonymous ECDH
2867 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2871 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2872 Use correct micalg parameters depending on digest(s) in signed message.
2875 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2876 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2879 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2880 an engine to register a method. Add ENGINE lookups for methods and
2881 functional reference processing.
2884 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2885 EVP_{Sign,Verify}* which allow an application to customise the signature
2889 *) New -resign option to smime utility. This adds one or more signers
2890 to an existing PKCS#7 signedData structure. Also -md option to use an
2891 alternative message digest algorithm for signing.
2894 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2895 create PKCS7 structures containing multiple signers. Update smime
2896 application to support multiple signers.
2899 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2903 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2904 Reorganize PBE internals to lookup from a static table using NIDs,
2905 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2906 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2907 PRF which will be automatically used with PBES2.
2910 *) Replace the algorithm specific calls to generate keys in "req" with the
2914 *) Update PKCS#7 enveloped data routines to use new API. This is now
2915 supported by any public key method supporting the encrypt operation. A
2916 ctrl is added to allow the public key algorithm to examine or modify
2917 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2921 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2922 a default digest type to use. In most cases this will be SHA1 but some
2923 algorithms (such as GOST) need to specify an alternative digest. The
2924 return value indicates how strong the prefernce is 1 means optional and
2925 2 is mandatory (that is it is the only supported type). Modify
2926 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2927 use the default md. Update openssl utilities to use the default digest
2928 type for signing if it is not explicitly indicated.
2931 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2932 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2933 signing method from the key type. This effectively removes the link
2934 between digests and public key types.
2937 *) Add an OID cross reference table and utility functions. Its purpose is to
2938 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2939 rsaEncryption. This will allow some of the algorithm specific hackery
2940 needed to use the correct OID to be removed.
2943 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2944 structures for PKCS7_sign(). They are now set up by the relevant public
2948 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2951 *) Add support for key derivation (agreement) in the API, DH method and
2955 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2956 public and private key formats. As a side effect these add additional
2957 command line functionality not previously available: DSA signatures can be
2958 generated and verified using pkeyutl and DH key support and generation in
2963 [Oliver Tappe <zooey@hirschkaefer.de>]
2965 *) New make target "install_html_docs" installs HTML renditions of the
2967 [Oliver Tappe <zooey@hirschkaefer.de>]
2969 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
2970 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2971 support key and parameter generation and add initial key generation
2972 functionality for RSA.
2975 *) Add functions for main EVP_PKEY_method operations. The undocumented
2976 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2977 EVP_PKEY_{encrypt,decrypt}_old.
2980 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2981 key API, doesn't do much yet.
2984 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2985 public key algorithms. New option to openssl utility:
2986 "list-public-key-algorithms" to print out info.
2989 *) Implement the Supported Elliptic Curves Extension for
2990 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2993 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2994 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2997 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2998 utilities such as rsa, dsa, dsaparam etc except they process any key
3002 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
3003 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
3004 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
3008 *) Initial support for pluggable public key ASN1.
3009 De-spaghettify the public key ASN1 handling. Move public and private
3010 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
3011 algorithm specific handling to a single module within the relevant
3012 algorithm directory. Add functions to allow (near) opaque processing
3013 of public and private key structures.
3016 *) Implement the Supported Point Formats Extension for
3017 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3020 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
3021 for the psk identity [hint] and the psk callback functions to the
3022 SSL_SESSION, SSL and SSL_CTX structure.
3025 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
3029 SSL_CTX_use_psk_identity_hint
3030 SSL_get_psk_identity_hint
3031 SSL_get_psk_identity
3032 SSL_use_psk_identity_hint
3034 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3036 *) Add RFC 3161 compliant time stamp request creation, response generation
3037 and response verification functionality.
3038 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3040 *) Add initial support for TLS extensions, specifically for the server_name
3041 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3042 have new members for a host name. The SSL data structure has an
3043 additional member SSL_CTX *initial_ctx so that new sessions can be
3044 stored in that context to allow for session resumption, even after the
3045 SSL has been switched to a new SSL_CTX in reaction to a client's
3046 server_name extension.
3048 New functions (subject to change):
3050 SSL_get_servername()
3051 SSL_get_servername_type()
3054 New CTRL codes and macros (subject to change):
3056 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3057 - SSL_CTX_set_tlsext_servername_callback()
3058 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3059 - SSL_CTX_set_tlsext_servername_arg()
3060 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3062 openssl s_client has a new '-servername ...' option.
3064 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3065 '-key2 ...', '-servername_fatal' (subject to change). This allows
3066 testing the HostName extension for a specific single host name ('-cert'
3067 and '-key' remain fallbacks for handshakes without HostName
3068 negotiation). If the unrecogninzed_name alert has to be sent, this by
3069 default is a warning; it becomes fatal with the '-servername_fatal'
3072 [Peter Sylvester, Remy Allais, Christophe Renou]
3074 *) Whirlpool hash implementation is added.
3077 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3078 bn(64,32). Because of instruction set limitations it doesn't have
3079 any negative impact on performance. This was done mostly in order
3080 to make it possible to share assembler modules, such as bn_mul_mont
3081 implementations, between 32- and 64-bit builds without hassle.
3084 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3085 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3089 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3090 dedicated Montgomery multiplication procedure, is introduced.
3091 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3092 "64-bit" performance on certain 32-bit targets.
3095 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3096 in SSL structures. New SSL ctrl to set maximum send fragment size.
3097 Save memory by seeting the I/O buffer sizes dynamically instead of
3098 using the maximum available value.
3101 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3102 in addition to the text details.
3105 *) Very, very preliminary EXPERIMENTAL support for printing of general
3106 ASN1 structures. This currently produces rather ugly output and doesn't
3107 handle several customised structures at all.
3110 *) Integrated support for PVK file format and some related formats such
3111 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3112 these in the 'rsa' and 'dsa' utilities.
3115 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3118 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3119 place for the (very old) "NETSCAPE" format certificates which are now
3120 handled using new ASN1 code equivalents.
3123 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3124 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3125 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3128 *) Modify CRL distribution points extension code to print out previously
3129 unsupported fields. Enhance extension setting code to allow setting of
3133 *) Add print and set support for Issuing Distribution Point CRL extension.
3136 *) Change 'Configure' script to enable Camellia by default.
3139 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3141 *) When rejecting SSL/TLS records due to an incorrect version number, never
3142 update s->server with a new major version number. As of
3143 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3144 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3145 the previous behavior could result in a read attempt at NULL when
3146 receiving specific incorrect SSL/TLS records once record payload
3147 protection is active. (CVE-2010-0740)
3148 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3150 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3151 could be crashed if the relevant tables were not present (e.g. chrooted).
3152 [Tomas Hoger <thoger@redhat.com>]
3154 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3156 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3157 [Martin Olsson, Neel Mehta]
3159 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3160 accommodate for stack sorting, always a write lock!).
3163 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3164 excessive delays in the RAND_poll(): over a minute. As a workaround
3165 include a time check in the inner Heap32Next loop too.
3168 *) The code that handled flushing of data in SSL/TLS originally used the
3169 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3170 the problem outlined in PR#1949. The fix suggested there however can
3171 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3172 of Apache). So instead simplify the code to flush unconditionally.
3173 This should be fine since flushing with no data to flush is a no op.
3176 *) Handle TLS versions 2.0 and later properly and correctly use the
3177 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3178 off ancient servers have a habit of sticking around for a while...
3181 *) Modify compression code so it frees up structures without using the
3182 ex_data callbacks. This works around a problem where some applications
3183 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3184 restarting) then use compression (e.g. SSL with compression) later.
3185 This results in significant per-connection memory leaks and
3186 has caused some security issues including CVE-2008-1678 and
3190 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3191 change when encrypting or decrypting.
3194 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3195 connect and renegotiate with servers which do not support RI.
3196 Until RI is more widely deployed this option is enabled by default.
3199 *) Add "missing" ssl ctrls to clear options and mode.
3202 *) If client attempts to renegotiate and doesn't support RI respond with
3203 a no_renegotiation alert as required by RFC5746. Some renegotiating
3204 TLS clients will continue a connection gracefully when they receive
3205 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3206 waiting for a server hello which it will never receive. Now we treat a
3207 received no_renegotiation alert as a fatal error. This is because
3208 applications requesting a renegotiation might well expect it to succeed
3209 and would have no code in place to handle the server denying it so the
3210 only safe thing to do is to terminate the connection.
3213 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3214 peer supports secure renegotiation and 0 otherwise. Print out peer
3215 renegotiation support in s_client/s_server.
3218 *) Replace the highly broken and deprecated SPKAC certification method with
3219 the updated NID creation version. This should correctly handle UTF8.
3222 *) Implement RFC5746. Re-enable renegotiation but require the extension
3223 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3224 turns out to be a bad idea. It has been replaced by
3225 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3226 SSL_CTX_set_options(). This is really not recommended unless you
3227 know what you are doing.
3228 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3230 *) Fixes to stateless session resumption handling. Use initial_ctx when
3231 issuing and attempting to decrypt tickets in case it has changed during
3232 servername handling. Use a non-zero length session ID when attempting
3233 stateless session resumption: this makes it possible to determine if
3234 a resumption has occurred immediately after receiving server hello
3235 (several places in OpenSSL subtly assume this) instead of later in
3239 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3240 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3241 fixes for a few places where the return code is not checked
3243 [Julia Lawall <julia@diku.dk>]
3245 *) Add --strict-warnings option to Configure script to include devteam
3246 warnings in other configurations.
3249 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3250 makes it possible to install openssl libraries in locations which
3251 have names other than "lib", for example "/usr/lib64" which some
3253 [Steve Henson, based on patch from Jeremy Utley]
3255 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3256 X690 8.9.12 and can produce some misleading textual output of OIDs.
3257 [Steve Henson, reported by Dan Kaminsky]
3259 *) Delete MD2 from algorithm tables. This follows the recommendation in
3260 several standards that it is not used in new applications due to
3261 several cryptographic weaknesses. For binary compatibility reasons
3262 the MD2 API is still compiled in by default.
3265 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3269 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3270 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3272 [Guenter <lists@gknw.net>]
3274 *) Fix the server certificate chain building code to use X509_verify_cert(),
3275 it used to have an ad-hoc builder which was unable to cope with anything
3276 other than a simple chain.
3277 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3279 *) Don't check self signed certificate signatures in X509_verify_cert()
3280 by default (a flag can override this): it just wastes time without
3281 adding any security. As a useful side effect self signed root CAs
3282 with non-FIPS digests are now usable in FIPS mode.
3285 *) In dtls1_process_out_of_seq_message() the check if the current message
3286 is already buffered was missing. For every new message was memory
3287 allocated, allowing an attacker to perform an denial of service attack
3288 with sending out of seq handshake messages until there is no memory
3289 left. Additionally every future messege was buffered, even if the
3290 sequence number made no sense and would be part of another handshake.
3291 So only messages with sequence numbers less than 10 in advance will be
3292 buffered. (CVE-2009-1378)
3293 [Robin Seggelmann, discovered by Daniel Mentz]
3295 *) Records are buffered if they arrive with a future epoch to be
3296 processed after finishing the corresponding handshake. There is
3297 currently no limitation to this buffer allowing an attacker to perform
3298 a DOS attack with sending records with future epochs until there is no
3299 memory left. This patch adds the pqueue_size() function to detemine
3300 the size of a buffer and limits the record buffer to 100 entries.
3302 [Robin Seggelmann, discovered by Daniel Mentz]
3304 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3305 parent structure is freed. (CVE-2009-1379)
3308 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3309 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3312 [Ilya O. <vrghost@gmail.com>]
3314 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3316 *) Disable renegotiation completely - this fixes a severe security
3317 problem (CVE-2009-3555) at the cost of breaking all
3318 renegotiation. Renegotiation can be re-enabled by setting
3319 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3320 run-time. This is really not recommended unless you know what
3324 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3326 *) Don't set val to NULL when freeing up structures, it is freed up by
3327 underlying code. If sizeof(void *) > sizeof(long) this can result in
3328 zeroing past the valid field. (CVE-2009-0789)
3329 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3331 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3332 checked correctly. This would allow some invalid signed attributes to
3333 appear to verify correctly. (CVE-2009-0591)
3334 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3336 *) Reject UniversalString and BMPString types with invalid lengths. This
3337 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3338 a legal length. (CVE-2009-0590)
3341 *) Set S/MIME signing as the default purpose rather than setting it
3342 unconditionally. This allows applications to override it at the store
3346 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3347 to handle some structures.
3350 *) Improve efficiency of mem_gets: don't search whole buffer each time
3352 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3354 *) New -hex option for openssl rand.
3357 *) Print out UTF8String and NumericString when parsing ASN1.
3360 *) Support NumericString type for name components.
3363 *) Allow CC in the environment to override the automatically chosen
3364 compiler. Note that nothing is done to ensure flags work with the
3368 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3370 *) Properly check EVP_VerifyFinal() and similar return values
3372 [Ben Laurie, Bodo Moeller, Google Security Team]
3374 *) Enable TLS extensions by default.
3377 *) Allow the CHIL engine to be loaded, whether the application is
3378 multithreaded or not. (This does not release the developer from the
3379 obligation to set up the dynamic locking callbacks.)
3380 [Sander Temme <sander@temme.net>]
3382 *) Use correct exit code if there is an error in dgst command.
3383 [Steve Henson; problem pointed out by Roland Dirlewanger]
3385 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3386 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3389 *) Add experimental JPAKE support, including demo authentication in
3390 s_client and s_server.
3393 *) Set the comparison function in v3_addr_canonize().
3394 [Rob Austein <sra@hactrn.net>]
3396 *) Add support for XMPP STARTTLS in s_client.
3397 [Philip Paeps <philip@freebsd.org>]
3399 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3400 to ensure that even with this option, only ciphersuites in the
3401 server's preference list will be accepted. (Note that the option
3402 applies only when resuming a session, so the earlier behavior was
3403 just about the algorithm choice for symmetric cryptography.)
3406 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3408 *) Fix NULL pointer dereference if a DTLS server received
3409 ChangeCipherSpec as first record (CVE-2009-1386).
3412 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
3413 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3416 *) The fix in 0.9.8c that supposedly got rid of unsafe
3417 double-checked locking was incomplete for RSA blinding,
3418 addressing just one layer of what turns out to have been
3419 doubly unsafe triple-checked locking.
3421 So now fix this for real by retiring the MONT_HELPER macro
3422 in crypto/rsa/rsa_eay.c.
3424 [Bodo Moeller; problem pointed out by Marius Schilder]
3426 *) Various precautionary measures:
3428 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3430 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3431 (NB: This would require knowledge of the secret session ticket key
3432 to exploit, in which case you'd be SOL either way.)
3434 - Change bn_nist.c so that it will properly handle input BIGNUMs
3435 outside the expected range.
3437 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3440 [Neel Mehta, Bodo Moeller]
3442 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3443 the load fails. Useful for distros.
3444 [Ben Laurie and the FreeBSD team]
3446 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3449 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3452 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3454 This work was sponsored by Logica.
3457 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3458 keystores. Support for SSL/TLS client authentication too.
3459 Not compiled unless enable-capieng specified to Configure.
3461 This work was sponsored by Logica.
3464 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3465 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3466 attribute creation routines such as certifcate requests and PKCS#12
3470 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3472 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3473 handshake which could lead to a cilent crash as found using the
3474 Codenomicon TLS test suite (CVE-2008-1672)
3475 [Steve Henson, Mark Cox]
3477 *) Fix double free in TLS server name extensions which could lead to
3478 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3481 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3483 Clear the error queue to ensure that error entries left from
3484 older function calls do not interfere with the correct operation.
3485 [Lutz Jaenicke, Erik de Castro Lopo]
3487 *) Remove root CA certificates of commercial CAs:
3489 The OpenSSL project does not recommend any specific CA and does not
3490 have any policy with respect to including or excluding any CA.
3491 Therefore it does not make any sense to ship an arbitrary selection
3492 of root CA certificates with the OpenSSL software.
3495 *) RSA OAEP patches to fix two separate invalid memory reads.
3496 The first one involves inputs when 'lzero' is greater than
3497 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3498 before the beginning of from). The second one involves inputs where
3499 the 'db' section contains nothing but zeroes (there is a one-byte
3500 invalid read after the end of 'db').
3501 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3503 *) Partial backport from 0.9.9-dev:
3505 Introduce bn_mul_mont (dedicated Montgomery multiplication
3506 procedure) as a candidate for BIGNUM assembler implementation.
3507 While 0.9.9-dev uses assembler for various architectures, only
3508 x86_64 is available by default here in the 0.9.8 branch, and
3509 32-bit x86 is available through a compile-time setting.
3511 To try the 32-bit x86 assembler implementation, use Configure
3512 option "enable-montasm" (which exists only for this backport).
3514 As "enable-montasm" for 32-bit x86 disclaims code stability
3515 anyway, in this constellation we activate additional code
3516 backported from 0.9.9-dev for further performance improvements,
3517 namely BN_from_montgomery_word. (To enable this otherwise,
3518 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3520 [Andy Polyakov (backport partially by Bodo Moeller)]
3522 *) Add TLS session ticket callback. This allows an application to set
3523 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3524 values. This is useful for key rollover for example where several key
3525 sets may exist with different names.
3528 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3529 This was broken until now in 0.9.8 releases, such that the only way
3530 a registered ENGINE could be used (assuming it initialises
3531 successfully on the host) was to explicitly set it as the default
3532 for the relevant algorithms. This is in contradiction with 0.9.7
3533 behaviour and the documentation. With this fix, when an ENGINE is
3534 registered into a given algorithm's table of implementations, the
3535 'uptodate' flag is reset so that auto-discovery will be used next
3536 time a new context for that algorithm attempts to select an
3538 [Ian Lister (tweaked by Geoff Thorpe)]
3540 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3541 implemention in the following ways:
3543 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3546 Lack of BER streaming support means one pass streaming processing is
3547 only supported if data is detached: setting the streaming flag is
3548 ignored for embedded content.
3550 CMS support is disabled by default and must be explicitly enabled
3551 with the enable-cms configuration option.
3554 *) Update the GMP engine glue to do direct copies between BIGNUM and
3555 mpz_t when openssl and GMP use the same limb size. Otherwise the
3556 existing "conversion via a text string export" trick is still used.
3557 [Paul Sheer <paulsheer@gmail.com>]
3559 *) Zlib compression BIO. This is a filter BIO which compressed and
3560 uncompresses any data passed through it.
3563 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3564 RFC3394 compatible AES key wrapping.
3567 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3568 sets string data without copying. X509_ALGOR_set0() and
3569 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3570 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3571 from an X509_ATTRIBUTE structure optionally checking it occurs only
3572 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3576 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3577 to get the expected BN_FLG_CONSTTIME behavior.
3578 [Bodo Moeller (Google)]
3582 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3583 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3584 - added some more tests to do_tests.pl
3585 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3586 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3587 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3588 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3589 - various changes to netware.pl to enable gcc-cross builds on Win32
3591 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3592 - various changes to fix missing prototype warnings
3593 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3594 - added AES, WHIRLPOOL and CPUID assembler code to build files
3595 - added missing AES assembler make rules to mk1mf.pl
3596 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3597 [Guenter Knauf <eflash@gmx.net>]
3599 *) Implement certificate status request TLS extension defined in RFC3546.
3600 A client can set the appropriate parameters and receive the encoded
3601 OCSP response via a callback. A server can query the supplied parameters
3602 and set the encoded OCSP response in the callback. Add simplified examples
3603 to s_client and s_server.
3606 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3608 *) Fix various bugs:
3609 + Binary incompatibility of ssl_ctx_st structure
3610 + DTLS interoperation with non-compliant servers
3611 + Don't call get_session_cb() without proposed session
3612 + Fix ia64 assembler code
3613 [Andy Polyakov, Steve Henson]
3615 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3617 *) DTLS Handshake overhaul. There were longstanding issues with
3618 OpenSSL DTLS implementation, which were making it impossible for
3619 RFC 4347 compliant client to communicate with OpenSSL server.
3620 Unfortunately just fixing these incompatibilities would "cut off"
3621 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3622 server keeps tolerating non RFC compliant syntax. The opposite is
3623 not true, 0.9.8f client can not communicate with earlier server.
3624 This update even addresses CVE-2007-4995.
3627 *) Changes to avoid need for function casts in OpenSSL: some compilers
3628 (gcc 4.2 and later) reject their use.
3629 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3632 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3633 RFC4507bis. The encrypted ticket format is an encrypted encoded
3634 SSL_SESSION structure, that way new session features are automatically
3637 If a client application caches session in an SSL_SESSION structure
3638 support is transparent because tickets are now stored in the encoded
3641 The SSL_CTX structure automatically generates keys for ticket
3642 protection in servers so again support should be possible
3643 with no application modification.
3645 If a client or server wishes to disable RFC4507 support then the option
3646 SSL_OP_NO_TICKET can be set.
3648 Add a TLS extension debugging callback to allow the contents of any client
3649 or server extensions to be examined.
3651 This work was sponsored by Google.
3654 *) Add initial support for TLS extensions, specifically for the server_name
3655 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3656 have new members for a host name. The SSL data structure has an
3657 additional member SSL_CTX *initial_ctx so that new sessions can be
3658 stored in that context to allow for session resumption, even after the
3659 SSL has been switched to a new SSL_CTX in reaction to a client's
3660 server_name extension.
3662 New functions (subject to change):
3664 SSL_get_servername()
3665 SSL_get_servername_type()
3668 New CTRL codes and macros (subject to change):
3670 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3671 - SSL_CTX_set_tlsext_servername_callback()
3672 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3673 - SSL_CTX_set_tlsext_servername_arg()
3674 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3676 openssl s_client has a new '-servername ...' option.
3678 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3679 '-key2 ...', '-servername_fatal' (subject to change). This allows
3680 testing the HostName extension for a specific single host name ('-cert'
3681 and '-key' remain fallbacks for handshakes without HostName
3682 negotiation). If the unrecogninzed_name alert has to be sent, this by
3683 default is a warning; it becomes fatal with the '-servername_fatal'
3686 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3688 *) Add AES and SSE2 assembly language support to VC++ build.
3691 *) Mitigate attack on final subtraction in Montgomery reduction.
3694 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3695 (which previously caused an internal error).
3698 *) Squeeze another 10% out of IGE mode when in != out.
3701 *) AES IGE mode speedup.
3702 [Dean Gaudet (Google)]
3704 *) Add the Korean symmetric 128-bit cipher SEED (see
3705 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3706 add SEED ciphersuites from RFC 4162:
3708 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3709 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3710 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3711 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3713 To minimize changes between patchlevels in the OpenSSL 0.9.8
3714 series, SEED remains excluded from compilation unless OpenSSL
3715 is configured with 'enable-seed'.
3716 [KISA, Bodo Moeller]
3718 *) Mitigate branch prediction attacks, which can be practical if a
3719 single processor is shared, allowing a spy process to extract
3720 information. For detailed background information, see
3721 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3722 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3723 and Necessary Software Countermeasures"). The core of the change
3724 are new versions BN_div_no_branch() and
3725 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3726 respectively, which are slower, but avoid the security-relevant
3727 conditional branches. These are automatically called by BN_div()
3728 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3729 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3730 remove a conditional branch.
3732 BN_FLG_CONSTTIME is the new name for the previous
3733 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3734 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3735 in the exponent causes BN_mod_exp_mont() to use the alternative
3736 implementation in BN_mod_exp_mont_consttime().) The old name
3737 remains as a deprecated alias.
3739 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3740 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3741 constant-time implementations for more than just exponentiation.
3742 Here too the old name is kept as a deprecated alias.
3744 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3745 the BN_BLINDING structure gets an independent copy of the
3746 modulus. This means that the previous "BIGNUM *m" argument to
3747 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3748 essentially becomes "const BIGNUM *m", although we can't actually
3749 change this in the header file before 0.9.9. It allows
3750 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3751 enable BN_FLG_CONSTTIME.
3753 [Matthew D Wood (Intel Corp)]
3755 *) In the SSL/TLS server implementation, be strict about session ID
3756 context matching (which matters if an application uses a single
3757 external cache for different purposes). Previously,
3758 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3759 set. This did ensure strict client verification, but meant that,
3760 with applications using a single external cache for quite
3761 different requirements, clients could circumvent ciphersuite
3762 restrictions for a given session ID context by starting a session
3763 in a different context.
3766 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3767 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3768 authentication-only ciphersuites.
3771 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3772 not complete and could lead to a possible single byte overflow
3773 (CVE-2007-5135) [Ben Laurie]
3775 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3777 *) Since AES128 and AES256 (and similarly Camellia128 and
3778 Camellia256) share a single mask bit in the logic of
3779 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3780 kludge to work properly if AES128 is available and AES256 isn't
3781 (or if Camellia128 is available and Camellia256 isn't).
3784 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3785 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3786 When a point or a seed is encoded in a BIT STRING, we need to
3787 prevent the removal of trailing zero bits to get the proper DER
3788 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3789 of a NamedBitList, for which trailing 0 bits need to be removed.)
3792 *) Have SSL/TLS server implementation tolerate "mismatched" record
3793 protocol version while receiving ClientHello even if the
3794 ClientHello is fragmented. (The server can't insist on the
3795 particular protocol version it has chosen before the ServerHello
3796 message has informed the client about his choice.)
3799 *) Add RFC 3779 support.
3800 [Rob Austein for ARIN, Ben Laurie]
3802 *) Load error codes if they are not already present instead of using a
3803 static variable. This allows them to be cleanly unloaded and reloaded.
3804 Improve header file function name parsing.
3807 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3808 or CAPABILITY handshake as required by RFCs.
3811 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3813 *) Introduce limits to prevent malicious keys being able to
3814 cause a denial of service. (CVE-2006-2940)
3815 [Steve Henson, Bodo Moeller]
3817 *) Fix ASN.1 parsing of certain invalid structures that can result
3818 in a denial of service. (CVE-2006-2937) [Steve Henson]
3820 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3821 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3823 *) Fix SSL client code which could crash if connecting to a
3824 malicious SSLv2 server. (CVE-2006-4343)
3825 [Tavis Ormandy and Will Drewry, Google Security Team]
3827 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3828 match only those. Before that, "AES256-SHA" would be interpreted
3829 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3830 the same strength classification in 0.9.7h) as we currently only
3831 have a single AES bit in the ciphersuite description bitmap.
3832 That change, however, also applied to ciphersuite strings such as
3833 "RC4-MD5" that intentionally matched multiple ciphersuites --
3834 namely, SSL 2.0 ciphersuites in addition to the more common ones
3835 from SSL 3.0/TLS 1.0.
3837 So we change the selection algorithm again: Naming an explicit
3838 ciphersuite selects this one ciphersuite, and any other similar
3839 ciphersuite (same bitmap) from *other* protocol versions.
3840 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3841 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3843 Since SSL 2.0 does not have any ciphersuites for which the
3844 128/256 bit distinction would be relevant, this works for now.
3845 The proper fix will be to use different bits for AES128 and
3846 AES256, which would have avoided the problems from the beginning;
3847 however, bits are scarce, so we can only do this in a new release
3848 (not just a patchlevel) when we can change the SSL_CIPHER
3849 definition to split the single 'unsigned long mask' bitmap into
3850 multiple values to extend the available space.
3854 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3856 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3857 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3859 *) Add AES IGE and biIGE modes.
3862 *) Change the Unix randomness entropy gathering to use poll() when
3863 possible instead of select(), since the latter has some
3864 undesirable limitations.
3865 [Darryl Miles via Richard Levitte and Bodo Moeller]
3867 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3868 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3869 cannot be implicitly activated as part of, e.g., the "AES" alias.
3870 However, please upgrade to OpenSSL 0.9.9[-dev] for
3871 non-experimental use of the ECC ciphersuites to get TLS extension
3872 support, which is required for curve and point format negotiation
3873 to avoid potential handshake problems.
3876 *) Disable rogue ciphersuites:
3878 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3879 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3880 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3882 The latter two were purportedly from
3883 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3886 Also deactivate the remaining ciphersuites from
3887 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3888 unofficial, and the ID has long expired.
3891 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3892 dual-core machines) and other potential thread-safety issues.
3895 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3896 versions), which is now available for royalty-free use
3897 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3898 Also, add Camellia TLS ciphersuites from RFC 4132.
3900 To minimize changes between patchlevels in the OpenSSL 0.9.8
3901 series, Camellia remains excluded from compilation unless OpenSSL
3902 is configured with 'enable-camellia'.
3905 *) Disable the padding bug check when compression is in use. The padding
3906 bug check assumes the first packet is of even length, this is not
3907 necessarily true if compresssion is enabled and can result in false
3908 positives causing handshake failure. The actual bug test is ancient
3909 code so it is hoped that implementations will either have fixed it by
3910 now or any which still have the bug do not support compression.
3913 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3915 *) When applying a cipher rule check to see if string match is an explicit
3916 cipher suite and only match that one cipher suite if it is.
3919 *) Link in manifests for VC++ if needed.
3920 [Austin Ziegler <halostatue@gmail.com>]
3922 *) Update support for ECC-based TLS ciphersuites according to
3923 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3924 TLS extensions, which are supported starting with the 0.9.9
3925 branch, not in the OpenSSL 0.9.8 branch).
3928 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3929 opaque EVP_CIPHER_CTX handling.
3932 *) Fixes and enhancements to zlib compression code. We now only use
3933 "zlib1.dll" and use the default __cdecl calling convention on Win32
3934 to conform with the standards mentioned here:
3935 http://www.zlib.net/DLL_FAQ.txt
3936 Static zlib linking now works on Windows and the new --with-zlib-include
3937 --with-zlib-lib options to Configure can be used to supply the location
3938 of the headers and library. Gracefully handle case where zlib library
3942 *) Several fixes and enhancements to the OID generation code. The old code
3943 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3944 handle numbers larger than ULONG_MAX, truncated printing and had a
3945 non standard OBJ_obj2txt() behaviour.
3948 *) Add support for building of engines under engine/ as shared libraries
3949 under VC++ build system.
3952 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3953 Hopefully, we will not see any false combination of paths any more.
3956 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3958 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3959 (part of SSL_OP_ALL). This option used to disable the
3960 countermeasure against man-in-the-middle protocol-version
3961 rollback in the SSL 2.0 server implementation, which is a bad
3962 idea. (CVE-2005-2969)
3964 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3965 for Information Security, National Institute of Advanced Industrial
3966 Science and Technology [AIST], Japan)]
3968 *) Add two function to clear and return the verify parameter flags.
3971 *) Keep cipherlists sorted in the source instead of sorting them at
3972 runtime, thus removing the need for a lock.
3975 *) Avoid some small subgroup attacks in Diffie-Hellman.
3976 [Nick Mathewson and Ben Laurie]
3978 *) Add functions for well-known primes.
3981 *) Extended Windows CE support.
3982 [Satoshi Nakamura and Andy Polyakov]
3984 *) Initialize SSL_METHOD structures at compile time instead of during
3985 runtime, thus removing the need for a lock.
3988 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3989 attempting to decrypt each encrypted key in turn. Add support to
3993 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3995 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3998 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
4001 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
4002 key into the same file any more.
4005 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
4008 *) Add -utf8 command line and config file option to 'ca'.
4009 [Stefan <stf@udoma.org]
4011 *) Removed the macro des_crypt(), as it seems to conflict with some
4012 libraries. Use DES_crypt().
4015 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
4016 involves renaming the source and generated shared-libs for
4017 both. The engines will accept the corrected or legacy ids
4018 ('ncipher' and '4758_cca' respectively) when binding. NB,
4019 this only applies when building 'shared'.
4020 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
4022 *) Add attribute functions to EVP_PKEY structure. Modify
4023 PKCS12_create() to recognize a CSP name attribute and
4024 use it. Make -CSP option work again in pkcs12 utility.
4027 *) Add new functionality to the bn blinding code:
4028 - automatic re-creation of the BN_BLINDING parameters after
4029 a fixed number of uses (currently 32)
4030 - add new function for parameter creation
4031 - introduce flags to control the update behaviour of the
4032 BN_BLINDING parameters
4033 - hide BN_BLINDING structure
4034 Add a second BN_BLINDING slot to the RSA structure to improve
4035 performance when a single RSA object is shared among several
4039 *) Add support for DTLS.
4040 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
4042 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
4043 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
4046 *) Remove buggy and incompletet DH cert support from
4047 ssl/ssl_rsa.c and ssl/s3_both.c
4050 *) Use SHA-1 instead of MD5 as the default digest algorithm for
4051 the apps/openssl applications.
4054 *) Compile clean with "-Wall -Wmissing-prototypes
4055 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
4056 DEBUG_SAFESTACK must also be set.
4059 *) Change ./Configure so that certain algorithms can be disabled by default.
4060 The new counterpiece to "no-xxx" is "enable-xxx".
4062 The patented RC5 and MDC2 algorithms will now be disabled unless
4063 "enable-rc5" and "enable-mdc2", respectively, are specified.
4065 (IDEA remains enabled despite being patented. This is because IDEA
4066 is frequently required for interoperability, and there is no license
4067 fee for non-commercial use. As before, "no-idea" can be used to
4068 avoid this algorithm.)
4072 *) Add processing of proxy certificates (see RFC 3820). This work was
4073 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
4074 EGEE (Enabling Grids for E-science in Europe).
4077 *) RC4 performance overhaul on modern architectures/implementations, such
4078 as Intel P4, IA-64 and AMD64.
4081 *) New utility extract-section.pl. This can be used specify an alternative
4082 section number in a pod file instead of having to treat each file as
4083 a separate case in Makefile. This can be done by adding two lines to the
4086 =for comment openssl_section:XXX
4088 The blank line is mandatory.
4092 *) New arguments -certform, -keyform and -pass for s_client and s_server
4093 to allow alternative format key and certificate files and passphrase
4097 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4098 update associated structures and add various utility functions.
4100 Add new policy related verify parameters, include policy checking in
4101 standard verify code. Enhance 'smime' application with extra parameters
4102 to support policy checking and print out.
4105 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4106 Nehemiah processors. These extensions support AES encryption in hardware
4107 as well as RNG (though RNG support is currently disabled).
4108 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4110 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4113 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4114 [Andy Polyakov and a number of other people]
4116 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4117 implementation contributed by IBM.
4118 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4120 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4121 exponent rather than 'unsigned long'. There is a corresponding change to
4122 the new 'rsa_keygen' element of the RSA_METHOD structure.
4123 [Jelte Jansen, Geoff Thorpe]
4125 *) Functionality for creating the initial serial number file is now
4126 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4128 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4129 number file to 1, which is bound to cause problems. To avoid
4130 the problems while respecting compatibility between different 0.9.7
4131 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4132 CA.pl for serial number initialization. With the new release 0.9.8,
4133 we can fix the problem directly in the 'ca' utility.)
4136 *) Reduced header interdepencies by declaring more opaque objects in
4137 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4138 give fewer recursive includes, which could break lazy source code - so
4139 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4140 developers should define this symbol when building and using openssl to
4141 ensure they track the recommended behaviour, interfaces, [etc], but
4142 backwards-compatible behaviour prevails when this isn't defined.
4145 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4148 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4149 This will generate a random key of the appropriate length based on the
4150 cipher context. The EVP_CIPHER can provide its own random key generation
4151 routine to support keys of a specific form. This is used in the des and
4152 3des routines to generate a key of the correct parity. Update S/MIME
4153 code to use new functions and hence generate correct parity DES keys.
4154 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4155 valid (weak or incorrect parity).
4158 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4159 as looking them up. This is useful when the verified structure may contain
4160 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4161 present unless the new PKCS7_NO_CRL flag is asserted.
4164 *) Extend ASN1 oid configuration module. It now additionally accepts the
4167 shortName = some long name, 1.2.3.4
4170 *) Reimplemented the BN_CTX implementation. There is now no more static
4171 limitation on the number of variables it can handle nor the depth of the
4172 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4173 information can now expand as required, and rather than having a single
4174 static array of bignums, BN_CTX now uses a linked-list of such arrays
4175 allowing it to expand on demand whilst maintaining the usefulness of
4176 BN_CTX's "bundling".
4179 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4180 to allow all RSA operations to function using a single BN_CTX.
4183 *) Preliminary support for certificate policy evaluation and checking. This
4184 is initially intended to pass the tests outlined in "Conformance Testing
4185 of Relying Party Client Certificate Path Processing Logic" v1.07.
4188 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4189 remained unused and not that useful. A variety of other little bignum
4190 tweaks and fixes have also been made continuing on from the audit (see
4194 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4195 associated ASN1, EVP and SSL functions and old ASN1 macros.
4198 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4199 and this should never fail. So the return value from the use of
4200 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4201 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4204 *) BN_CTX_get() should return zero-valued bignums, providing the same
4205 initialised value as BN_new().
4206 [Geoff Thorpe, suggested by Ulf Möller]
4208 *) Support for inhibitAnyPolicy certificate extension.
4211 *) An audit of the BIGNUM code is underway, for which debugging code is
4212 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4213 is considered valid when processing BIGNUMs, and causes execution to
4214 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4215 further steps are taken to deliberately pollute unused data in BIGNUM
4216 structures to try and expose faulty code further on. For now, openssl will
4217 (in its default mode of operation) continue to tolerate the inconsistent
4218 forms that it has tolerated in the past, but authors and packagers should
4219 consider trying openssl and their own applications when compiled with
4220 these debugging symbols defined. It will help highlight potential bugs in
4221 their own code, and will improve the test coverage for OpenSSL itself. At
4222 some point, these tighter rules will become openssl's default to improve
4223 maintainability, though the assert()s and other overheads will remain only
4224 in debugging configurations. See bn.h for more details.
4225 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4227 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4228 that can only be obtained through BN_CTX_new() (which implicitly
4229 initialises it). The presence of this function only made it possible
4230 to overwrite an existing structure (and cause memory leaks).
4233 *) Because of the callback-based approach for implementing LHASH as a
4234 template type, lh_insert() adds opaque objects to hash-tables and
4235 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4236 to clean up those corresponding objects before destroying the hash table
4237 (and losing the object pointers). So some over-zealous constifications in
4238 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4239 objects as "const" and the lh_doall[_arg] callback wrappers are not
4240 prototyped to have "const" restrictions on the object pointers they are
4241 given (and so aren't required to cast them away any more).
4244 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4245 (speed) prefers to use its own implementation. The two implementations
4246 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4247 its object type properly exposed (MS_TM) instead of casting to/from "char
4248 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4249 aren't necessarily the greatest nomenclatures - but this is what was used
4250 internally to the implementation so I've used that for now.
4253 *) Ensure that deprecated functions do not get compiled when
4254 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4255 the self-tests were still using deprecated key-generation functions so
4256 these have been updated also.
4259 *) Reorganise PKCS#7 code to separate the digest location functionality
4260 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
4261 New function PKCS7_set_digest() to set the digest type for PKCS#7
4262 digestedData type. Add additional code to correctly generate the
4263 digestedData type and add support for this type in PKCS7 initialization
4267 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4268 structure of type "other".
4271 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4272 sure the loop does correctly stop and breaking ("division by zero")
4273 modulus operations are not performed. The (pre-generated) prime
4274 table crypto/bn/bn_prime.h was already correct, but it could not be
4275 re-generated on some platforms because of the "division by zero"
4276 situation in the script.
4277 [Ralf S. Engelschall]
4279 *) Update support for ECC-based TLS ciphersuites according to
4280 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4281 SHA-1 now is only used for "small" curves (where the
4282 representation of a field element takes up to 24 bytes); for
4283 larger curves, the field element resulting from ECDH is directly
4284 used as premaster secret.
4285 [Douglas Stebila (Sun Microsystems Laboratories)]
4287 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4288 curve secp160r1 to the tests.
4289 [Douglas Stebila (Sun Microsystems Laboratories)]
4291 *) Add the possibility to load symbols globally with DSO.
4292 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4294 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4295 control of the error stack.
4298 *) Add support for STORE in ENGINE.
4301 *) Add the STORE type. The intention is to provide a common interface
4302 to certificate and key stores, be they simple file-based stores, or
4303 HSM-type store, or LDAP stores, or...
4304 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4307 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4308 pass a list of arguments to any function as well as provide a way
4309 for a function to pass data back to the caller.
4312 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4313 works like BUF_strdup() but can be used to duplicate a portion of
4314 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4318 *) Add the function sk_find_ex() which works like sk_find(), but will
4319 return an index to an element even if an exact match couldn't be
4320 found. The index is guaranteed to point at the element where the
4321 searched-for key would be inserted to preserve sorting order.
4324 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4325 takes an extra flags argument for optional functionality. Currently,
4326 the following flags are defined:
4328 OBJ_BSEARCH_VALUE_ON_NOMATCH
4329 This one gets OBJ_bsearch_ex() to return a pointer to the first
4330 element where the comparing function returns a negative or zero
4333 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4334 This one gets OBJ_bsearch_ex() to return a pointer to the first
4335 element where the comparing function returns zero. This is useful
4336 if there are more than one element where the comparing function
4340 *) Make it possible to create self-signed certificates with 'openssl ca'
4341 in such a way that the self-signed certificate becomes part of the
4342 CA database and uses the same mechanisms for serial number generation
4343 as all other certificate signing. The new flag '-selfsign' enables
4344 this functionality. Adapt CA.sh and CA.pl.in.
4347 *) Add functionality to check the public key of a certificate request
4348 against a given private. This is useful to check that a certificate
4349 request can be signed by that key (self-signing).
4352 *) Make it possible to have multiple active certificates with the same
4353 subject in the CA index file. This is done only if the keyword
4354 'unique_subject' is set to 'no' in the main CA section (default
4355 if 'CA_default') of the configuration file. The value is saved
4356 with the database itself in a separate index attribute file,
4357 named like the index file with '.attr' appended to the name.
4360 *) Generate muti valued AVAs using '+' notation in config files for
4364 *) Support for nameConstraints certificate extension.
4367 *) Support for policyConstraints certificate extension.
4370 *) Support for policyMappings certificate extension.
4373 *) Make sure the default DSA_METHOD implementation only uses its
4374 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4375 and change its own handlers to be NULL so as to remove unnecessary
4376 indirection. This lets alternative implementations fallback to the
4377 default implementation more easily.
4380 *) Support for directoryName in GeneralName related extensions
4384 *) Make it possible to link applications using Makefile.shared.
4385 Make that possible even when linking against static libraries!
4388 *) Support for single pass processing for S/MIME signing. This now
4389 means that S/MIME signing can be done from a pipe, in addition
4390 cleartext signing (multipart/signed type) is effectively streaming
4391 and the signed data does not need to be all held in memory.
4393 This is done with a new flag PKCS7_STREAM. When this flag is set
4394 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4395 is done after the data is output (and digests calculated) in
4396 SMIME_write_PKCS7().
4399 *) Add full support for -rpath/-R, both in shared libraries and
4400 applications, at least on the platforms where it's known how
4404 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4405 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4406 will now compute a table of multiples of the generator that
4407 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4408 faster (notably in the case of a single point multiplication,
4409 scalar * generator).
4410 [Nils Larsch, Bodo Moeller]
4412 *) IPv6 support for certificate extensions. The various extensions
4413 which use the IP:a.b.c.d can now take IPv6 addresses using the
4414 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4418 *) Added an ENGINE that implements RSA by performing private key
4419 exponentiations with the GMP library. The conversions to and from
4420 GMP's mpz_t format aren't optimised nor are any montgomery forms
4421 cached, and on x86 it appears OpenSSL's own performance has caught up.
4422 However there are likely to be other architectures where GMP could
4423 provide a boost. This ENGINE is not built in by default, but it can be
4424 specified at Configure time and should be accompanied by the necessary
4425 linker additions, eg;
4426 ./config -DOPENSSL_USE_GMP -lgmp
4429 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4430 testing availability of engines with "-t" - the old behaviour is
4431 produced by increasing the feature's verbosity with "-tt".
4434 *) ECDSA routines: under certain error conditions uninitialized BN objects
4435 could be freed. Solution: make sure initialization is performed early
4436 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4440 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4441 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4442 software implementations. For DSA and DH, parameter generation can
4443 also be overriden by providing the appropriate method callbacks.
4446 *) Change the "progress" mechanism used in key-generation and
4447 primality testing to functions that take a new BN_GENCB pointer in
4448 place of callback/argument pairs. The new API functions have "_ex"
4449 postfixes and the older functions are reimplemented as wrappers for
4450 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4451 declarations of the old functions to help (graceful) attempts to
4452 migrate to the new functions. Also, the new key-generation API
4453 functions operate on a caller-supplied key-structure and return
4454 success/failure rather than returning a key or NULL - this is to
4455 help make "keygen" another member function of RSA_METHOD etc.
4457 Example for using the new callback interface:
4459 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4463 BN_GENCB_set(&my_cb, my_callback, my_arg);
4465 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4466 /* For the meaning of a, b in calls to my_callback(), see the
4467 * documentation of the function that calls the callback.
4468 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4469 * my_callback should return 1 if it wants BN_is_prime_ex()
4470 * to continue, or 0 to stop.
4475 *) Change the ZLIB compression method to be stateful, and make it
4476 available to TLS with the number defined in
4477 draft-ietf-tls-compression-04.txt.
4480 *) Add the ASN.1 structures and functions for CertificatePair, which
4481 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4483 CertificatePair ::= SEQUENCE {
4484 forward [0] Certificate OPTIONAL,
4485 reverse [1] Certificate OPTIONAL,
4486 -- at least one of the pair shall be present -- }
4488 Also implement the PEM functions to read and write certificate
4489 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4491 This needed to be defined, mostly for the sake of the LDAP
4492 attribute crossCertificatePair, but may prove useful elsewhere as
4496 *) Make it possible to inhibit symlinking of shared libraries in
4497 Makefile.shared, for Cygwin's sake.
4500 *) Extend the BIGNUM API by creating a function
4501 void BN_set_negative(BIGNUM *a, int neg);
4502 and a macro that behave like
4503 int BN_is_negative(const BIGNUM *a);
4505 to avoid the need to access 'a->neg' directly in applications.
4508 *) Implement fast modular reduction for pseudo-Mersenne primes
4509 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4510 EC_GROUP_new_curve_GFp() will now automatically use this
4512 [Nils Larsch <nla@trustcenter.de>]
4514 *) Add new lock type (CRYPTO_LOCK_BN).
4517 *) Change the ENGINE framework to automatically load engines
4518 dynamically from specific directories unless they could be
4519 found to already be built in or loaded. Move all the
4520 current engines except for the cryptodev one to a new
4522 The engines in engines/ are built as shared libraries if
4523 the "shared" options was given to ./Configure or ./config.
4524 Otherwise, they are inserted in libcrypto.a.
4525 /usr/local/ssl/engines is the default directory for dynamic
4526 engines, but that can be overriden at configure time through
4527 the usual use of --prefix and/or --openssldir, and at run
4528 time with the environment variable OPENSSL_ENGINES.
4529 [Geoff Thorpe and Richard Levitte]
4531 *) Add Makefile.shared, a helper makefile to build shared
4532 libraries. Addapt Makefile.org.
4535 *) Add version info to Win32 DLLs.
4536 [Peter 'Luna' Runestig" <peter@runestig.com>]
4538 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4539 can be added using this API to created arbitrary PKCS#12
4540 files while avoiding the low level API.
4542 New options to PKCS12_create(), key or cert can be NULL and
4543 will then be omitted from the output file. The encryption
4544 algorithm NIDs can be set to -1 for no encryption, the mac
4545 iteration count can be set to 0 to omit the mac.
4547 Enhance pkcs12 utility by making the -nokeys and -nocerts
4548 options work when creating a PKCS#12 file. New option -nomac
4549 to omit the mac, NONE can be set for an encryption algorithm.
4550 New code is modified to use the enhanced PKCS12_create()
4551 instead of the low level API.
4554 *) Extend ASN1 encoder to support indefinite length constructed
4555 encoding. This can output sequences tags and octet strings in
4556 this form. Modify pk7_asn1.c to support indefinite length
4557 encoding. This is experimental and needs additional code to
4558 be useful, such as an ASN1 bio and some enhanced streaming
4561 Extend template encode functionality so that tagging is passed
4562 down to the template encoder.
4565 *) Let 'openssl req' fail if an argument to '-newkey' is not
4566 recognized instead of using RSA as a default.
4569 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4570 As these are not official, they are not included in "ALL";
4571 the "ECCdraft" ciphersuite group alias can be used to select them.
4572 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4574 *) Add ECDH engine support.
4575 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4577 *) Add ECDH in new directory crypto/ecdh/.
4578 [Douglas Stebila (Sun Microsystems Laboratories)]
4580 *) Let BN_rand_range() abort with an error after 100 iterations
4581 without success (which indicates a broken PRNG).
4584 *) Change BN_mod_sqrt() so that it verifies that the input value
4585 is really the square of the return value. (Previously,
4586 BN_mod_sqrt would show GIGO behaviour.)
4589 *) Add named elliptic curves over binary fields from X9.62, SECG,
4590 and WAP/WTLS; add OIDs that were still missing.
4592 [Sheueling Chang Shantz and Douglas Stebila
4593 (Sun Microsystems Laboratories)]
4595 *) Extend the EC library for elliptic curves over binary fields
4596 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4599 EC_GF2m_simple_method
4603 EC_GROUP_new_curve_GF2m
4604 EC_GROUP_set_curve_GF2m
4605 EC_GROUP_get_curve_GF2m
4606 EC_POINT_set_affine_coordinates_GF2m
4607 EC_POINT_get_affine_coordinates_GF2m
4608 EC_POINT_set_compressed_coordinates_GF2m
4610 Point compression for binary fields is disabled by default for
4611 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4614 As binary polynomials are represented as BIGNUMs, various members
4615 of the EC_GROUP and EC_POINT data structures can be shared
4616 between the implementations for prime fields and binary fields;
4617 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4618 are essentially identical to their ..._GFp counterparts.
4619 (For simplicity, the '..._GFp' prefix has been dropped from
4620 various internal method names.)
4622 An internal 'field_div' method (similar to 'field_mul' and
4623 'field_sqr') has been added; this is used only for binary fields.
4625 [Sheueling Chang Shantz and Douglas Stebila
4626 (Sun Microsystems Laboratories)]
4628 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4629 through methods ('mul', 'precompute_mult').
4631 The generic implementations (now internally called 'ec_wNAF_mul'
4632 and 'ec_wNAF_precomputed_mult') remain the default if these
4633 methods are undefined.
4635 [Sheueling Chang Shantz and Douglas Stebila
4636 (Sun Microsystems Laboratories)]
4638 *) New function EC_GROUP_get_degree, which is defined through
4639 EC_METHOD. For curves over prime fields, this returns the bit
4640 length of the modulus.
4642 [Sheueling Chang Shantz and Douglas Stebila
4643 (Sun Microsystems Laboratories)]
4645 *) New functions EC_GROUP_dup, EC_POINT_dup.
4646 (These simply call ..._new and ..._copy).
4648 [Sheueling Chang Shantz and Douglas Stebila
4649 (Sun Microsystems Laboratories)]
4651 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4652 Polynomials are represented as BIGNUMs (where the sign bit is not
4653 used) in the following functions [macros]:
4656 BN_GF2m_sub [= BN_GF2m_add]
4657 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4658 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4659 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4661 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4662 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4663 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4664 BN_GF2m_cmp [= BN_ucmp]
4666 (Note that only the 'mod' functions are actually for fields GF(2^m).
4667 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4669 For some functions, an the irreducible polynomial defining a
4670 field can be given as an 'unsigned int[]' with strictly
4671 decreasing elements giving the indices of those bits that are set;
4672 i.e., p[] represents the polynomial
4673 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4675 p[0] > p[1] > ... > p[k] = 0.
4676 This applies to the following functions:
4681 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4682 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4684 BN_GF2m_mod_sqrt_arr
4685 BN_GF2m_mod_solve_quad_arr
4689 Conversion can be performed by the following functions:
4694 bntest.c has additional tests for binary polynomial arithmetic.
4696 Two implementations for BN_GF2m_mod_div() are available.
4697 The default algorithm simply uses BN_GF2m_mod_inv() and
4698 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4699 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4700 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4702 [Sheueling Chang Shantz and Douglas Stebila
4703 (Sun Microsystems Laboratories)]
4705 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4706 functionality is disabled at compile-time.
4707 [Douglas Stebila <douglas.stebila@sun.com>]
4709 *) Change default behaviour of 'openssl asn1parse' so that more
4710 information is visible when viewing, e.g., a certificate:
4712 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4713 mode the content of non-printable OCTET STRINGs is output in a
4714 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4715 avoid the appearance of a printable string.
4716 [Nils Larsch <nla@trustcenter.de>]
4718 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4720 EC_GROUP_set_asn1_flag()
4721 EC_GROUP_get_asn1_flag()
4722 EC_GROUP_set_point_conversion_form()
4723 EC_GROUP_get_point_conversion_form()
4724 These control ASN1 encoding details:
4725 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4726 has been set to OPENSSL_EC_NAMED_CURVE.
4727 - Points are encoded in uncompressed form by default; options for
4728 asn1_for are as for point2oct, namely
4729 POINT_CONVERSION_COMPRESSED
4730 POINT_CONVERSION_UNCOMPRESSED
4731 POINT_CONVERSION_HYBRID
4733 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4736 EC_GROUP_get0_seed()
4737 EC_GROUP_get_seed_len()
4738 This is used only for ASN1 purposes (so far).
4739 [Nils Larsch <nla@trustcenter.de>]
4741 *) Add 'field_type' member to EC_METHOD, which holds the NID
4742 of the appropriate field type OID. The new function
4743 EC_METHOD_get_field_type() returns this value.
4744 [Nils Larsch <nla@trustcenter.de>]
4749 EC_POINT_point2hex()
4750 EC_POINT_hex2point()
4751 providing useful interfaces to EC_POINT_point2oct() and
4752 EC_POINT_oct2point().
4753 [Nils Larsch <nla@trustcenter.de>]
4755 *) Change internals of the EC library so that the functions
4756 EC_GROUP_set_generator()
4757 EC_GROUP_get_generator()
4758 EC_GROUP_get_order()
4759 EC_GROUP_get_cofactor()
4760 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4761 to methods, which would lead to unnecessary code duplication when
4762 adding different types of curves.
4763 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4765 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4766 arithmetic, and such that modified wNAFs are generated
4767 (which avoid length expansion in many cases).
4770 *) Add a function EC_GROUP_check_discriminant() (defined via
4771 EC_METHOD) that verifies that the curve discriminant is non-zero.
4773 Add a function EC_GROUP_check() that makes some sanity tests
4774 on a EC_GROUP, its generator and order. This includes
4775 EC_GROUP_check_discriminant().
4776 [Nils Larsch <nla@trustcenter.de>]
4778 *) Add ECDSA in new directory crypto/ecdsa/.
4780 Add applications 'openssl ecparam' and 'openssl ecdsa'
4781 (these are based on 'openssl dsaparam' and 'openssl dsa').
4783 ECDSA support is also included in various other files across the
4784 library. Most notably,
4785 - 'openssl req' now has a '-newkey ecdsa:file' option;
4786 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4787 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4788 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4789 them suitable for ECDSA where domain parameters must be
4790 extracted before the specific public key;
4791 - ECDSA engine support has been added.
4792 [Nils Larsch <nla@trustcenter.de>]
4794 *) Include some named elliptic curves, and add OIDs from X9.62,
4795 SECG, and WAP/WTLS. Each curve can be obtained from the new
4797 EC_GROUP_new_by_curve_name(),
4798 and the list of available named curves can be obtained with
4799 EC_get_builtin_curves().
4800 Also add a 'curve_name' member to EC_GROUP objects, which can be
4802 EC_GROUP_set_curve_name()
4803 EC_GROUP_get_curve_name()
4804 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4806 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4807 was actually never needed) and in BN_mul(). The removal in BN_mul()
4808 required a small change in bn_mul_part_recursive() and the addition
4809 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4810 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4811 bn_sub_words() and bn_add_words() except they take arrays with
4815 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4817 *) Cleanse PEM buffers before freeing them since they may contain
4819 [Benjamin Bennett <ben@psc.edu>]
4821 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4822 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4823 authentication-only ciphersuites.
4826 *) Since AES128 and AES256 share a single mask bit in the logic of
4827 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4828 kludge to work properly if AES128 is available and AES256 isn't.
4831 *) Expand security boundary to match 1.1.1 module.
4834 *) Remove redundant features: hash file source, editing of test vectors
4835 modify fipsld to use external fips_premain.c signature.
4838 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4839 run algorithm test programs.
4842 *) Make algorithm test programs more tolerant of whitespace.
4845 *) Have SSL/TLS server implementation tolerate "mismatched" record
4846 protocol version while receiving ClientHello even if the
4847 ClientHello is fragmented. (The server can't insist on the
4848 particular protocol version it has chosen before the ServerHello
4849 message has informed the client about his choice.)
4852 *) Load error codes if they are not already present instead of using a
4853 static variable. This allows them to be cleanly unloaded and reloaded.
4856 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4858 *) Introduce limits to prevent malicious keys being able to
4859 cause a denial of service. (CVE-2006-2940)
4860 [Steve Henson, Bodo Moeller]
4862 *) Fix ASN.1 parsing of certain invalid structures that can result
4863 in a denial of service. (CVE-2006-2937) [Steve Henson]
4865 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4866 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4868 *) Fix SSL client code which could crash if connecting to a
4869 malicious SSLv2 server. (CVE-2006-4343)
4870 [Tavis Ormandy and Will Drewry, Google Security Team]
4872 *) Change ciphersuite string processing so that an explicit
4873 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4874 will no longer include "AES128-SHA"), and any other similar
4875 ciphersuite (same bitmap) from *other* protocol versions (so that
4876 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4877 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4878 changes from 0.9.8b and 0.9.8d.
4881 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4883 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4884 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4886 *) Change the Unix randomness entropy gathering to use poll() when
4887 possible instead of select(), since the latter has some
4888 undesirable limitations.
4889 [Darryl Miles via Richard Levitte and Bodo Moeller]
4891 *) Disable rogue ciphersuites:
4893 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4894 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4895 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4897 The latter two were purportedly from
4898 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4901 Also deactive the remaining ciphersuites from
4902 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4903 unofficial, and the ID has long expired.
4906 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4907 dual-core machines) and other potential thread-safety issues.
4910 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4912 *) Adapt fipsld and the build system to link against the validated FIPS
4913 module in FIPS mode.
4916 *) Fixes for VC++ 2005 build under Windows.
4919 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4920 from a Windows bash shell such as MSYS. It is autodetected from the
4921 "config" script when run from a VC++ environment. Modify standard VC++
4922 build to use fipscanister.o from the GNU make build.
4925 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4927 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4928 The value now differs depending on if you build for FIPS or not.
4929 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4930 safely run with a non-FIPSed libcrypto, as it may crash because of
4931 the difference induced by this change.
4934 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4936 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4937 (part of SSL_OP_ALL). This option used to disable the
4938 countermeasure against man-in-the-middle protocol-version
4939 rollback in the SSL 2.0 server implementation, which is a bad
4940 idea. (CVE-2005-2969)
4942 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4943 for Information Security, National Institute of Advanced Industrial
4944 Science and Technology [AIST], Japan)]
4946 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4947 mainly for FIPS compliance and not fully integrated at this stage.
4950 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4951 the exponentiation using a fixed-length exponent. (Otherwise,
4952 the information leaked through timing could expose the secret key
4953 after many signatures; cf. Bleichenbacher's attack on DSA with
4957 *) Make a new fixed-window mod_exp implementation the default for
4958 RSA, DSA, and DH private-key operations so that the sequence of
4959 squares and multiplies and the memory access pattern are
4960 independent of the particular secret key. This will mitigate
4961 cache-timing and potential related attacks.
4963 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4964 and this is automatically used by BN_mod_exp_mont() if the new flag
4965 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4966 will use this BN flag for private exponents unless the flag
4967 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4968 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4970 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4972 *) Change the client implementation for SSLv23_method() and
4973 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4974 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4975 (Previously, the SSL 2.0 backwards compatible Client Hello
4976 message format would be used even with SSL_OP_NO_SSLv2.)
4979 *) Add support for smime-type MIME parameter in S/MIME messages which some
4983 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4984 a threadsafe manner. Modify rsa code to use new function and add calls
4985 to dsa and dh code (which had race conditions before).
4988 *) Include the fixed error library code in the C error file definitions
4989 instead of fixing them up at runtime. This keeps the error code
4990 structures constant.
4993 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4995 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4998 *) Fixes for newer kerberos headers. NB: the casts are needed because
4999 the 'length' field is signed on one version and unsigned on another
5000 with no (?) obvious way to tell the difference, without these VC++
5001 complains. Also the "definition" of FAR (blank) is no longer included
5002 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
5003 some needed definitions.
5006 *) Undo Cygwin change.
5009 *) Added support for proxy certificates according to RFC 3820.
5010 Because they may be a security thread to unaware applications,
5011 they must be explicitely allowed in run-time. See
5012 docs/HOWTO/proxy_certificates.txt for further information.
5015 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
5017 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
5018 server and client random values. Previously
5019 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
5020 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
5022 This change has negligible security impact because:
5024 1. Server and client random values still have 24 bytes of pseudo random
5027 2. Server and client random values are sent in the clear in the initial
5030 3. The master secret is derived using the premaster secret (48 bytes in
5031 size for static RSA ciphersuites) as well as client server and random
5034 The OpenSSL team would like to thank the UK NISCC for bringing this issue
5037 [Stephen Henson, reported by UK NISCC]
5039 *) Use Windows randomness collection on Cygwin.
5042 *) Fix hang in EGD/PRNGD query when communication socket is closed
5043 prematurely by EGD/PRNGD.
5044 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
5046 *) Prompt for pass phrases when appropriate for PKCS12 input format.
5049 *) Back-port of selected performance improvements from development
5050 branch, as well as improved support for PowerPC platforms.
5053 *) Add lots of checks for memory allocation failure, error codes to indicate
5054 failure and freeing up memory if a failure occurs.
5055 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
5057 *) Add new -passin argument to dgst.
5060 *) Perform some character comparisons of different types in X509_NAME_cmp:
5061 this is needed for some certificates that reencode DNs into UTF8Strings
5062 (in violation of RFC3280) and can't or wont issue name rollover
5066 *) Make an explicit check during certificate validation to see that
5067 the CA setting in each certificate on the chain is correct. As a
5068 side effect always do the following basic checks on extensions,
5069 not just when there's an associated purpose to the check:
5071 - if there is an unhandled critical extension (unless the user
5072 has chosen to ignore this fault)
5073 - if the path length has been exceeded (if one is set at all)
5074 - that certain extensions fit the associated purpose (if one has
5078 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5080 *) Avoid a race condition when CRLs are checked in a multi threaded
5081 environment. This would happen due to the reordering of the revoked
5082 entries during signature checking and serial number lookup. Now the
5083 encoding is cached and the serial number sort performed under a lock.
5084 Add new STACK function sk_is_sorted().
5087 *) Add Delta CRL to the extension code.
5090 *) Various fixes to s3_pkt.c so alerts are sent properly.
5091 [David Holmes <d.holmes@f5.com>]
5093 *) Reduce the chances of duplicate issuer name and serial numbers (in
5094 violation of RFC3280) using the OpenSSL certificate creation utilities.
5095 This is done by creating a random 64 bit value for the initial serial
5096 number when a serial number file is created or when a self signed
5097 certificate is created using 'openssl req -x509'. The initial serial
5098 number file is created using 'openssl x509 -next_serial' in CA.pl
5099 rather than being initialized to 1.
5102 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5104 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5105 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5106 [Joe Orton, Steve Henson]
5108 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5110 [Joe Orton, Steve Henson]
5112 *) Make it possible to have multiple active certificates with the same
5113 subject in the CA index file. This is done only if the keyword
5114 'unique_subject' is set to 'no' in the main CA section (default
5115 if 'CA_default') of the configuration file. The value is saved
5116 with the database itself in a separate index attribute file,
5117 named like the index file with '.attr' appended to the name.
5120 *) X509 verify fixes. Disable broken certificate workarounds when
5121 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5122 keyUsage extension present. Don't accept CRLs with unhandled critical
5123 extensions: since verify currently doesn't process CRL extensions this
5124 rejects a CRL with *any* critical extensions. Add new verify error codes
5128 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5129 A clarification of RFC2560 will require the use of OCTET STRINGs and
5130 some implementations cannot handle the current raw format. Since OpenSSL
5131 copies and compares OCSP nonces as opaque blobs without any attempt at
5132 parsing them this should not create any compatibility issues.
5135 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5136 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5137 this HMAC (and other) operations are several times slower than OpenSSL
5141 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5142 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5144 *) Use the correct content when signing type "other".
5147 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5149 *) Fix various bugs revealed by running the NISCC test suite:
5151 Stop out of bounds reads in the ASN1 code when presented with
5152 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5154 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5156 If verify callback ignores invalid public key errors don't try to check
5157 certificate signature with the NULL public key.
5161 *) New -ignore_err option in ocsp application to stop the server
5162 exiting on the first error in a request.
5165 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5166 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5170 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5171 extra data after the compression methods not only for TLS 1.0
5172 but also for SSL 3.0 (as required by the specification).
5173 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5175 *) Change X509_certificate_type() to mark the key as exported/exportable
5176 when it's 512 *bits* long, not 512 bytes.
5179 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5180 blocks during encryption.
5183 *) Various fixes to base64 BIO and non blocking I/O. On write
5184 flushes were not handled properly if the BIO retried. On read
5185 data was not being buffered properly and had various logic bugs.
5186 This also affects blocking I/O when the data being decoded is a
5190 *) Various S/MIME bugfixes and compatibility changes:
5191 output correct application/pkcs7 MIME type if
5192 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5193 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5194 of files as .eml work). Correctly handle very long lines in MIME
5198 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5200 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5201 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5202 a protocol version number mismatch like a decryption error
5203 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5206 *) Turn on RSA blinding by default in the default implementation
5207 to avoid a timing attack. Applications that don't want it can call
5208 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5209 They would be ill-advised to do so in most cases.
5210 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5212 *) Change RSA blinding code so that it works when the PRNG is not
5213 seeded (in this case, the secret RSA exponent is abused as
5214 an unpredictable seed -- if it is not unpredictable, there
5215 is no point in blinding anyway). Make RSA blinding thread-safe
5216 by remembering the creator's thread ID in rsa->blinding and
5217 having all other threads use local one-time blinding factors
5218 (this requires more computation than sharing rsa->blinding, but
5219 avoids excessive locking; and if an RSA object is not shared
5220 between threads, blinding will still be very fast).
5223 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5224 ENGINE as defaults for all supported algorithms irrespective of
5225 the 'flags' parameter. 'flags' is now honoured, so applications
5226 should make sure they are passing it correctly.
5229 *) Target "mingw" now allows native Windows code to be generated in
5230 the Cygwin environment as well as with the MinGW compiler.
5233 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5235 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5236 via timing by performing a MAC computation even if incorrrect
5237 block cipher padding has been found. This is a countermeasure
5238 against active attacks where the attacker has to distinguish
5239 between bad padding and a MAC verification error. (CVE-2003-0078)
5241 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5242 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5243 Martin Vuagnoux (EPFL, Ilion)]
5245 *) Make the no-err option work as intended. The intention with no-err
5246 is not to have the whole error stack handling routines removed from
5247 libcrypto, it's only intended to remove all the function name and
5248 reason texts, thereby removing some of the footprint that may not
5249 be interesting if those errors aren't displayed anyway.
5251 NOTE: it's still possible for any application or module to have it's
5252 own set of error texts inserted. The routines are there, just not
5253 used by default when no-err is given.
5256 *) Add support for FreeBSD on IA64.
5257 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5259 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5260 Kerberos function mit_des_cbc_cksum(). Before this change,
5261 the value returned by DES_cbc_cksum() was like the one from
5262 mit_des_cbc_cksum(), except the bytes were swapped.
5263 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5265 *) Allow an application to disable the automatic SSL chain building.
5266 Before this a rather primitive chain build was always performed in
5267 ssl3_output_cert_chain(): an application had no way to send the
5268 correct chain if the automatic operation produced an incorrect result.
5270 Now the chain builder is disabled if either:
5272 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5274 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5276 The reasoning behind this is that an application would not want the
5277 auto chain building to take place if extra chain certificates are
5278 present and it might also want a means of sending no additional
5279 certificates (for example the chain has two certificates and the
5283 *) Add the possibility to build without the ENGINE framework.
5284 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5286 *) Under Win32 gmtime() can return NULL: check return value in
5287 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5290 *) DSA routines: under certain error conditions uninitialized BN objects
5291 could be freed. Solution: make sure initialization is performed early
5292 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5293 Nils Larsch <nla@trustcenter.de> via PR#459)
5296 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5297 checked on reconnect on the client side, therefore session resumption
5298 could still fail with a "ssl session id is different" error. This
5299 behaviour is masked when SSL_OP_ALL is used due to
5300 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5301 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5302 followup to PR #377.
5305 *) IA-32 assembler support enhancements: unified ELF targets, support
5306 for SCO/Caldera platforms, fix for Cygwin shared build.
5309 *) Add support for FreeBSD on sparc64. As a consequence, support for
5310 FreeBSD on non-x86 processors is separate from x86 processors on
5311 the config script, much like the NetBSD support.
5312 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5314 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5316 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5319 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5320 code (06) was taken as the first octet of the session ID and the last
5321 octet was ignored consequently. As a result SSLv2 client side session
5322 caching could not have worked due to the session ID mismatch between
5324 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5328 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5329 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5333 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5334 seems that in spite of existing for more than a year, many application
5335 author have done nothing to provide the necessary callbacks, which
5336 means that this particular engine will not work properly anywhere.
5337 This is a very unfortunate situation which forces us, in the name
5338 of usability, to give the hw_ncipher.c a static lock, which is part
5340 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5341 appear in 0.9.8 or later. We EXPECT application authors to have
5342 dealt properly with this when 0.9.8 is released (unless we actually
5343 make such changes in the libcrypto locking code that changes will
5344 have to be made anyway).
5347 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5348 octets have been read, EOF or an error occurs. Without this change
5349 some truncated ASN1 structures will not produce an error.
5352 *) Disable Heimdal support, since it hasn't been fully implemented.
5353 Still give the possibility to force the use of Heimdal, but with
5354 warnings and a request that patches get sent to openssl-dev.
5357 *) Add the VC-CE target, introduce the WINCE sysname, and add
5358 INSTALL.WCE and appropriate conditionals to make it build.
5359 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5361 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5362 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5363 edit numbers of the version.
5364 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5366 *) Introduce safe string copy and catenation functions
5367 (BUF_strlcpy() and BUF_strlcat()).
5368 [Ben Laurie (CHATS) and Richard Levitte]
5370 *) Avoid using fixed-size buffers for one-line DNs.
5371 [Ben Laurie (CHATS)]
5373 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5374 resizing buffers containing secrets, and use where appropriate.
5375 [Ben Laurie (CHATS)]
5377 *) Avoid using fixed size buffers for configuration file location.
5378 [Ben Laurie (CHATS)]
5380 *) Avoid filename truncation for various CA files.
5381 [Ben Laurie (CHATS)]
5383 *) Use sizeof in preference to magic numbers.
5384 [Ben Laurie (CHATS)]
5386 *) Avoid filename truncation in cert requests.
5387 [Ben Laurie (CHATS)]
5389 *) Add assertions to check for (supposedly impossible) buffer
5391 [Ben Laurie (CHATS)]
5393 *) Don't cache truncated DNS entries in the local cache (this could
5394 potentially lead to a spoofing attack).
5395 [Ben Laurie (CHATS)]
5397 *) Fix various buffers to be large enough for hex/decimal
5398 representations in a platform independent manner.
5399 [Ben Laurie (CHATS)]
5401 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5402 resizing buffers containing secrets, and use where appropriate.
5403 [Ben Laurie (CHATS)]
5405 *) Add BIO_indent() to avoid much slightly worrying code to do
5407 [Ben Laurie (CHATS)]
5409 *) Convert sprintf()/BIO_puts() to BIO_printf().
5410 [Ben Laurie (CHATS)]
5412 *) buffer_gets() could terminate with the buffer only half
5414 [Ben Laurie (CHATS)]
5416 *) Add assertions to prevent user-supplied crypto functions from
5417 overflowing internal buffers by having large block sizes, etc.
5418 [Ben Laurie (CHATS)]
5420 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5422 [Ben Laurie (CHATS)]
5424 *) Eliminate unused copy of key in RC4.
5425 [Ben Laurie (CHATS)]
5427 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5428 [Ben Laurie (CHATS)]
5430 *) Fix off-by-one error in EGD path.
5431 [Ben Laurie (CHATS)]
5433 *) If RANDFILE path is too long, ignore instead of truncating.
5434 [Ben Laurie (CHATS)]
5436 *) Eliminate unused and incorrectly sized X.509 structure
5438 [Ben Laurie (CHATS)]
5440 *) Eliminate unused and dangerous function knumber().
5441 [Ben Laurie (CHATS)]
5443 *) Eliminate unused and dangerous structure, KSSL_ERR.
5444 [Ben Laurie (CHATS)]
5446 *) Protect against overlong session ID context length in an encoded
5447 session object. Since these are local, this does not appear to be
5449 [Ben Laurie (CHATS)]
5451 *) Change from security patch (see 0.9.6e below) that did not affect
5452 the 0.9.6 release series:
5454 Remote buffer overflow in SSL3 protocol - an attacker could
5455 supply an oversized master key in Kerberos-enabled versions.
5457 [Ben Laurie (CHATS)]
5459 *) Change the SSL kerb5 codes to match RFC 2712.
5462 *) Make -nameopt work fully for req and add -reqopt switch.
5463 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5465 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5466 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5468 *) Make sure tests can be performed even if the corresponding algorithms
5469 have been removed entirely. This was also the last step to make
5470 OpenSSL compilable with DJGPP under all reasonable conditions.
5471 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5473 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5474 to allow version independent disabling of normally unselected ciphers,
5475 which may be activated as a side-effect of selecting a single cipher.
5477 (E.g., cipher list string "RSA" enables ciphersuites that are left
5478 out of "ALL" because they do not provide symmetric encryption.
5479 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5480 [Lutz Jaenicke, Bodo Moeller]
5482 *) Add appropriate support for separate platform-dependent build
5483 directories. The recommended way to make a platform-dependent
5484 build directory is the following (tested on Linux), maybe with
5487 # Place yourself outside of the OpenSSL source tree. In
5488 # this example, the environment variable OPENSSL_SOURCE
5489 # is assumed to contain the absolute OpenSSL source directory.
5490 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5491 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5492 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5493 mkdir -p `dirname $F`
5494 ln -s $OPENSSL_SOURCE/$F $F
5497 To be absolutely sure not to disturb the source tree, a "make clean"
5498 is a good thing. If it isn't successfull, don't worry about it,
5499 it probably means the source directory is very clean.
5502 *) Make sure any ENGINE control commands make local copies of string
5503 pointers passed to them whenever necessary. Otherwise it is possible
5504 the caller may have overwritten (or deallocated) the original string
5505 data when a later ENGINE operation tries to use the stored values.
5506 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5508 *) Improve diagnostics in file reading and command-line digests.
5509 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5511 *) Add AES modes CFB and OFB to the object database. Correct an
5512 error in AES-CFB decryption.
5515 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5516 allows existing EVP_CIPHER_CTX structures to be reused after
5517 calling EVP_*Final(). This behaviour is used by encryption
5518 BIOs and some applications. This has the side effect that
5519 applications must explicitly clean up cipher contexts with
5520 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5523 *) Check the values of dna and dnb in bn_mul_recursive before calling
5524 bn_mul_comba (a non zero value means the a or b arrays do not contain
5525 n2 elements) and fallback to bn_mul_normal if either is not zero.
5528 *) Fix escaping of non-ASCII characters when using the -subj option
5529 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5532 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5533 form for "surname", serialNumber has no short form.
5534 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5535 therefore remove "mail" short name for "internet 7".
5536 The OID for unique identifiers in X509 certificates is
5537 x500UniqueIdentifier, not uniqueIdentifier.
5538 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5541 *) Add an "init" command to the ENGINE config module and auto initialize
5542 ENGINEs. Without any "init" command the ENGINE will be initialized
5543 after all ctrl commands have been executed on it. If init=1 the
5544 ENGINE is initailized at that point (ctrls before that point are run
5545 on the uninitialized ENGINE and after on the initialized one). If
5546 init=0 then the ENGINE will not be iniatialized at all.
5549 *) Fix the 'app_verify_callback' interface so that the user-defined
5550 argument is actually passed to the callback: In the
5551 SSL_CTX_set_cert_verify_callback() prototype, the callback
5552 declaration has been changed from
5555 int (*cb)(X509_STORE_CTX *,void *);
5556 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5557 i=s->ctx->app_verify_callback(&ctx)
5558 has been changed into
5559 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5561 To update applications using SSL_CTX_set_cert_verify_callback(),
5562 a dummy argument can be added to their callback functions.
5563 [D. K. Smetters <smetters@parc.xerox.com>]
5565 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5566 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5568 *) Add and OPENSSL_LOAD_CONF define which will cause
5569 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5570 This allows older applications to transparently support certain
5571 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5572 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5573 load the config file and OPENSSL_add_all_algorithms_conf() which will
5574 always load it have also been added.
5577 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5578 Adjust NIDs and EVP layer.
5579 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5581 *) Config modules support in openssl utility.
5583 Most commands now load modules from the config file,
5584 though in a few (such as version) this isn't done
5585 because it couldn't be used for anything.
5587 In the case of ca and req the config file used is
5588 the same as the utility itself: that is the -config
5589 command line option can be used to specify an
5593 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5594 use "openssl_conf" if filename is NULL use default openssl config file.
5597 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5598 config section name. Add a new flag to tolerate a missing config file
5599 and move code to CONF_modules_load_file().
5602 *) Support for crypto accelerator cards from Accelerated Encryption
5603 Processing, www.aep.ie. (Use engine 'aep')
5604 The support was copied from 0.9.6c [engine] and adapted/corrected
5605 to work with the new engine framework.
5606 [AEP Inc. and Richard Levitte]
5608 *) Support for SureWare crypto accelerator cards from Baltimore
5609 Technologies. (Use engine 'sureware')
5610 The support was copied from 0.9.6c [engine] and adapted
5611 to work with the new engine framework.
5614 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5615 make the newer ENGINE framework commands for the CHIL engine work.
5616 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5618 *) Make it possible to produce shared libraries on ReliantUNIX.
5619 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5621 *) Add the configuration target debug-linux-ppro.
5622 Make 'openssl rsa' use the general key loading routines
5623 implemented in apps.c, and make those routines able to
5624 handle the key format FORMAT_NETSCAPE and the variant
5626 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5628 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5629 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5631 *) Add -keyform to rsautl, and document -engine.
5632 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5634 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5635 BIO_R_NO_SUCH_FILE error code rather than the generic
5636 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5639 *) Add new functions
5641 ERR_peek_last_error_line
5642 ERR_peek_last_error_line_data.
5643 These are similar to
5646 ERR_peek_error_line_data,
5647 but report on the latest error recorded rather than the first one
5648 still in the error queue.
5649 [Ben Laurie, Bodo Moeller]
5651 *) default_algorithms option in ENGINE config module. This allows things
5653 default_algorithms = ALL
5654 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5657 *) Prelminary ENGINE config module.
5660 *) New experimental application configuration code.
5663 *) Change the AES code to follow the same name structure as all other
5664 symmetric ciphers, and behave the same way. Move everything to
5665 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5666 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5668 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5669 [Ben Laurie and Theo de Raadt]
5671 *) Add option to output public keys in req command.
5672 [Massimiliano Pala madwolf@openca.org]
5674 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5675 (up to about 10% better than before for P-192 and P-224).
5678 *) New functions/macros
5680 SSL_CTX_set_msg_callback(ctx, cb)
5681 SSL_CTX_set_msg_callback_arg(ctx, arg)
5682 SSL_set_msg_callback(ssl, cb)
5683 SSL_set_msg_callback_arg(ssl, arg)
5685 to request calling a callback function
5687 void cb(int write_p, int version, int content_type,
5688 const void *buf, size_t len, SSL *ssl, void *arg)
5690 whenever a protocol message has been completely received
5691 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5692 protocol version according to which the SSL library interprets
5693 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5694 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5695 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5696 specification (change_cipher_spec(20), alert(21), handshake(22)).
5697 'buf' and 'len' point to the actual message, 'ssl' to the
5698 SSL object, and 'arg' is the application-defined value set by
5699 SSL[_CTX]_set_msg_callback_arg().
5701 'openssl s_client' and 'openssl s_server' have new '-msg' options
5702 to enable a callback that displays all protocol messages.
5705 *) Change the shared library support so shared libraries are built as
5706 soon as the corresponding static library is finished, and thereby get
5707 openssl and the test programs linked against the shared library.
5708 This still only happens when the keyword "shard" has been given to
5709 the configuration scripts.
5711 NOTE: shared library support is still an experimental thing, and
5712 backward binary compatibility is still not guaranteed.
5713 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5715 *) Add support for Subject Information Access extension.
5716 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5718 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5719 additional bytes when new memory had to be allocated, not just
5720 when reusing an existing buffer.
5723 *) New command line and configuration option 'utf8' for the req command.
5724 This allows field values to be specified as UTF8 strings.
5727 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5728 runs for the former and machine-readable output for the latter.
5731 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5732 of the e-mail address in the DN (i.e., it will go into a certificate
5733 extension only). The new configuration file option 'email_in_dn = no'
5734 has the same effect.
5735 [Massimiliano Pala madwolf@openca.org]
5737 *) Change all functions with names starting with des_ to be starting
5738 with DES_ instead. Add wrappers that are compatible with libdes,
5739 but are named _ossl_old_des_*. Finally, add macros that map the
5740 des_* symbols to the corresponding _ossl_old_des_* if libdes
5741 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5742 desired, the des_* symbols will be mapped to DES_*, with one
5745 Since we provide two compatibility mappings, the user needs to
5746 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5747 compatibility is desired. The default (i.e., when that macro
5748 isn't defined) is OpenSSL 0.9.6c compatibility.
5750 There are also macros that enable and disable the support of old
5751 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5752 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5753 are defined, the default will apply: to support the old des routines.
5755 In either case, one must include openssl/des.h to get the correct
5756 definitions. Do not try to just include openssl/des_old.h, that
5759 NOTE: This is a major break of an old API into a new one. Software
5760 authors are encouraged to switch to the DES_ style functions. Some
5761 time in the future, des_old.h and the libdes compatibility functions
5762 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5763 default), and then completely removed.
5766 *) Test for certificates which contain unsupported critical extensions.
5767 If such a certificate is found during a verify operation it is
5768 rejected by default: this behaviour can be overridden by either
5769 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5770 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5771 X509_supported_extension() has also been added which returns 1 if a
5772 particular extension is supported.
5775 *) Modify the behaviour of EVP cipher functions in similar way to digests
5776 to retain compatibility with existing code.
5779 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5780 compatibility with existing code. In particular the 'ctx' parameter does
5781 not have to be to be initialized before the call to EVP_DigestInit() and
5782 it is tidied up after a call to EVP_DigestFinal(). New function
5783 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5784 EVP_MD_CTX_copy() changed to not require the destination to be
5785 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5786 requires the destination to be valid.
5788 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5789 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5792 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5793 so that complete 'Handshake' protocol structures are kept in memory
5794 instead of overwriting 'msg_type' and 'length' with 'body' data.
5797 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5798 [Massimo Santin via Richard Levitte]
5800 *) Major restructuring to the underlying ENGINE code. This includes
5801 reduction of linker bloat, separation of pure "ENGINE" manipulation
5802 (initialisation, etc) from functionality dealing with implementations
5803 of specific crypto iterfaces. This change also introduces integrated
5804 support for symmetric ciphers and digest implementations - so ENGINEs
5805 can now accelerate these by providing EVP_CIPHER and EVP_MD
5806 implementations of their own. This is detailed in crypto/engine/README
5807 as it couldn't be adequately described here. However, there are a few
5808 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5809 were changed in the original introduction of ENGINE code have now
5810 reverted back - the hooking from this code to ENGINE is now a good
5811 deal more passive and at run-time, operations deal directly with
5812 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5813 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5814 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5815 they were not being used by the framework as there is no concept of a
5816 BIGNUM_METHOD and they could not be generalised to the new
5817 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5818 ENGINE_cpy() has been removed as it cannot be consistently defined in
5822 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5825 *) Change mkdef.pl to sort symbols that get the same entry number,
5826 and make sure the automatically generated functions ERR_load_*
5827 become part of libeay.num as well.
5830 *) New function SSL_renegotiate_pending(). This returns true once
5831 renegotiation has been requested (either SSL_renegotiate() call
5832 or HelloRequest/ClientHello receveived from the peer) and becomes
5833 false once a handshake has been completed.
5834 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5835 sends a HelloRequest, but does not ensure that a handshake takes
5836 place. SSL_renegotiate_pending() is useful for checking if the
5837 client has followed the request.)
5840 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5841 By default, clients may request session resumption even during
5842 renegotiation (if session ID contexts permit); with this option,
5843 session resumption is possible only in the first handshake.
5845 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5846 more bits available for options that should not be part of
5847 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5850 *) Add some demos for certificate and certificate request creation.
5853 *) Make maximum certificate chain size accepted from the peer application
5854 settable (SSL*_get/set_max_cert_list()), as proposed by
5855 "Douglas E. Engert" <deengert@anl.gov>.
5858 *) Add support for shared libraries for Unixware-7
5859 (Boyd Lynn Gerber <gerberb@zenez.com>).
5862 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5863 be done prior to destruction. Use this to unload error strings from
5864 ENGINEs that load their own error strings. NB: This adds two new API
5865 functions to "get" and "set" this destroy handler in an ENGINE.
5868 *) Alter all existing ENGINE implementations (except "openssl" and
5869 "openbsd") to dynamically instantiate their own error strings. This
5870 makes them more flexible to be built both as statically-linked ENGINEs
5871 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5872 Also, add stub code to each that makes building them as self-contained
5873 shared-libraries easier (see README.ENGINE).
5876 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5877 implementations into applications that are completely implemented in
5878 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5879 commands that can be used to configure what shared-library to load and
5880 to control aspects of the way it is handled. Also, made an update to
5881 the README.ENGINE file that brings its information up-to-date and
5882 provides some information and instructions on the "dynamic" ENGINE
5883 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5886 *) Make it possible to unload ranges of ERR strings with a new
5887 "ERR_unload_strings" function.
5890 *) Add a copy() function to EVP_MD.
5893 *) Make EVP_MD routines take a context pointer instead of just the
5894 md_data void pointer.
5897 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5898 that the digest can only process a single chunk of data
5899 (typically because it is provided by a piece of
5900 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5901 is only going to provide a single chunk of data, and hence the
5902 framework needn't accumulate the data for oneshot drivers.
5905 *) As with "ERR", make it possible to replace the underlying "ex_data"
5906 functions. This change also alters the storage and management of global
5907 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5908 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5909 index counters. The API functions that use this state have been changed
5910 to take a "class_index" rather than pointers to the class's local STACK
5911 and counter, and there is now an API function to dynamically create new
5912 classes. This centralisation allows us to (a) plug a lot of the
5913 thread-safety problems that existed, and (b) makes it possible to clean
5914 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5915 such data would previously have always leaked in application code and
5916 workarounds were in place to make the memory debugging turn a blind eye
5917 to it. Application code that doesn't use this new function will still
5918 leak as before, but their memory debugging output will announce it now
5919 rather than letting it slide.
5921 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5922 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5923 has a return value to indicate success or failure.
5926 *) Make it possible to replace the underlying "ERR" functions such that the
5927 global state (2 LHASH tables and 2 locks) is only used by the "default"
5928 implementation. This change also adds two functions to "get" and "set"
5929 the implementation prior to it being automatically set the first time
5930 any other ERR function takes place. Ie. an application can call "get",
5931 pass the return value to a module it has just loaded, and that module
5932 can call its own "set" function using that value. This means the
5933 module's "ERR" operations will use (and modify) the error state in the
5934 application and not in its own statically linked copy of OpenSSL code.
5937 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5938 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5939 the operation, and provides a more encapsulated way for external code
5940 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5941 to use these functions rather than manually incrementing the counts.
5943 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5946 *) Add EVP test program.
5949 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5952 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5953 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5954 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5955 These allow a CRL to be built without having to access X509_CRL fields
5956 directly. Modify 'ca' application to use new functions.
5959 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5960 bug workarounds. Rollback attack detection is a security feature.
5961 The problem will only arise on OpenSSL servers when TLSv1 is not
5962 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5963 Software authors not wanting to support TLSv1 will have special reasons
5964 for their choice and can explicitly enable this option.
5965 [Bodo Moeller, Lutz Jaenicke]
5967 *) Rationalise EVP so it can be extended: don't include a union of
5968 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5969 (similar to those existing for EVP_CIPHER_CTX).
5974 EVP_MD_CTX_init(&md); /* new function call */
5975 EVP_DigestInit(&md, EVP_sha1());
5976 EVP_DigestUpdate(&md, in, len);
5977 EVP_DigestFinal(&md, out, NULL);
5978 EVP_MD_CTX_cleanup(&md); /* new function call */
5982 *) Make DES key schedule conform to the usual scheme, as well as
5983 correcting its structure. This means that calls to DES functions
5984 now have to pass a pointer to a des_key_schedule instead of a
5985 plain des_key_schedule (which was actually always a pointer
5988 des_key_schedule ks;
5990 des_set_key_checked(..., &ks);
5991 des_ncbc_encrypt(..., &ks, ...);
5993 (Note that a later change renames 'des_...' into 'DES_...'.)
5996 *) Initial reduction of linker bloat: the use of some functions, such as
5997 PEM causes large amounts of unused functions to be linked in due to
5998 poor organisation. For example pem_all.c contains every PEM function
5999 which has a knock on effect of linking in large amounts of (unused)
6000 ASN1 code. Grouping together similar functions and splitting unrelated
6001 functions prevents this.
6004 *) Cleanup of EVP macros.
6007 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
6008 correct _ecb suffix.
6011 *) Add initial OCSP responder support to ocsp application. The
6012 revocation information is handled using the text based index
6013 use by the ca application. The responder can either handle
6014 requests generated internally, supplied in files (for example
6015 via a CGI script) or using an internal minimal server.
6018 *) Add configuration choices to get zlib compression for TLS.
6021 *) Changes to Kerberos SSL for RFC 2712 compliance:
6022 1. Implemented real KerberosWrapper, instead of just using
6023 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
6024 2. Implemented optional authenticator field of KerberosWrapper.
6026 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
6027 and authenticator structs; see crypto/krb5/.
6029 Generalized Kerberos calls to support multiple Kerberos libraries.
6030 [Vern Staats <staatsvr@asc.hpc.mil>,
6031 Jeffrey Altman <jaltman@columbia.edu>
6032 via Richard Levitte]
6034 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
6035 already does with RSA. testdsa.h now has 'priv_key/pub_key'
6036 values for each of the key sizes rather than having just
6037 parameters (and 'speed' generating keys each time).
6040 *) Speed up EVP routines.
6043 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
6044 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
6045 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
6046 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
6048 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
6049 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
6050 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
6053 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
6055 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
6058 *) Added the OS2-EMX target.
6059 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
6061 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
6062 to support NCONF routines in extension code. New function CONF_set_nconf()
6063 to allow functions which take an NCONF to also handle the old LHASH
6064 structure: this means that the old CONF compatible routines can be
6065 retained (in particular wrt extensions) without having to duplicate the
6066 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
6069 *) Enhance the general user interface with mechanisms for inner control
6070 and with possibilities to have yes/no kind of prompts.
6073 *) Change all calls to low level digest routines in the library and
6074 applications to use EVP. Add missing calls to HMAC_cleanup() and
6075 don't assume HMAC_CTX can be copied using memcpy().
6076 [Verdon Walker <VWalker@novell.com>, Steve Henson]
6078 *) Add the possibility to control engines through control names but with
6079 arbitrary arguments instead of just a string.
6080 Change the key loaders to take a UI_METHOD instead of a callback
6081 function pointer. NOTE: this breaks binary compatibility with earlier
6082 versions of OpenSSL [engine].
6083 Adapt the nCipher code for these new conditions and add a card insertion
6087 *) Enhance the general user interface with mechanisms to better support
6088 dialog box interfaces, application-defined prompts, the possibility
6089 to use defaults (for example default passwords from somewhere else)
6090 and interrupts/cancellations.
6093 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6094 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6097 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6098 tidy up some unnecessarily weird code in 'sk_new()').
6099 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6101 *) Change the key loading routines for ENGINEs to use the same kind
6102 callback (pem_password_cb) as all other routines that need this
6106 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6107 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6108 than this minimum value is recommended.
6111 *) New random seeder for OpenVMS, using the system process statistics
6112 that are easily reachable.
6115 *) Windows apparently can't transparently handle global
6116 variables defined in DLLs. Initialisations such as:
6118 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6120 wont compile. This is used by the any applications that need to
6121 declare their own ASN1 modules. This was fixed by adding the option
6122 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6123 needed for static libraries under Win32.
6126 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6127 setting of purpose and trust fields. New X509_STORE trust and
6128 purpose functions and tidy up setting in other SSL functions.
6131 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6132 structure. These are inherited by X509_STORE_CTX when it is
6133 initialised. This allows various defaults to be set in the
6134 X509_STORE structure (such as flags for CRL checking and custom
6135 purpose or trust settings) for functions which only use X509_STORE_CTX
6136 internally such as S/MIME.
6138 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6139 trust settings if they are not set in X509_STORE. This allows X509_STORE
6140 purposes and trust (in S/MIME for example) to override any set by default.
6142 Add command line options for CRL checking to smime, s_client and s_server
6146 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6147 are set then the CRL is looked up in the X509_STORE structure and
6148 its validity and signature checked, then if the certificate is found
6149 in the CRL the verify fails with a revoked error.
6151 Various new CRL related callbacks added to X509_STORE_CTX structure.
6153 Command line options added to 'verify' application to support this.
6155 This needs some additional work, such as being able to handle multiple
6156 CRLs with different times, extension based lookup (rather than just
6157 by subject name) and ultimately more complete V2 CRL extension
6161 *) Add a general user interface API (crypto/ui/). This is designed
6162 to replace things like des_read_password and friends (backward
6163 compatibility functions using this new API are provided).
6164 The purpose is to remove prompting functions from the DES code
6165 section as well as provide for prompting through dialog boxes in
6166 a window system and the like.
6169 *) Add "ex_data" support to ENGINE so implementations can add state at a
6170 per-structure level rather than having to store it globally.
6173 *) Make it possible for ENGINE structures to be copied when retrieved by
6174 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6175 This causes the "original" ENGINE structure to act like a template,
6176 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6177 operational state can be localised to each ENGINE structure, despite the
6178 fact they all share the same "methods". New ENGINE structures returned in
6179 this case have no functional references and the return value is the single
6180 structural reference. This matches the single structural reference returned
6181 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6185 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6186 needs to match any other type at all we need to manually clear the
6190 *) Changes to the "openssl engine" utility to include;
6191 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6192 about an ENGINE's available control commands.
6193 - executing control commands from command line arguments using the
6194 '-pre' and '-post' switches. '-post' is only used if '-t' is
6195 specified and the ENGINE is successfully initialised. The syntax for
6196 the individual commands are colon-separated, for example;
6197 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6200 *) New dynamic control command support for ENGINEs. ENGINEs can now
6201 declare their own commands (numbers), names (strings), descriptions,
6202 and input types for run-time discovery by calling applications. A
6203 subset of these commands are implicitly classed as "executable"
6204 depending on their input type, and only these can be invoked through
6205 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6206 can be based on user input, config files, etc). The distinction is
6207 that "executable" commands cannot return anything other than a boolean
6208 result and can only support numeric or string input, whereas some
6209 discoverable commands may only be for direct use through
6210 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6211 pointers, or other custom uses. The "executable" commands are to
6212 support parameterisations of ENGINE behaviour that can be
6213 unambiguously defined by ENGINEs and used consistently across any
6214 OpenSSL-based application. Commands have been added to all the
6215 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6216 control over shared-library paths without source code alterations.
6219 *) Changed all ENGINE implementations to dynamically allocate their
6220 ENGINEs rather than declaring them statically. Apart from this being
6221 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6222 this also allows the implementations to compile without using the
6223 internal engine_int.h header.
6226 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6227 'const' value. Any code that should be able to modify a RAND_METHOD
6228 should already have non-const pointers to it (ie. they should only
6229 modify their own ones).
6232 *) Made a variety of little tweaks to the ENGINE code.
6233 - "atalla" and "ubsec" string definitions were moved from header files
6234 to C code. "nuron" string definitions were placed in variables
6235 rather than hard-coded - allowing parameterisation of these values
6236 later on via ctrl() commands.
6237 - Removed unused "#if 0"'d code.
6238 - Fixed engine list iteration code so it uses ENGINE_free() to release
6239 structural references.
6240 - Constified the RAND_METHOD element of ENGINE structures.
6241 - Constified various get/set functions as appropriate and added
6242 missing functions (including a catch-all ENGINE_cpy that duplicates
6243 all ENGINE values onto a new ENGINE except reference counts/state).
6244 - Removed NULL parameter checks in get/set functions. Setting a method
6245 or function to NULL is a way of cancelling out a previously set
6246 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6247 and doesn't justify the extra error symbols and code.
6248 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6249 flags from engine_int.h to engine.h.
6250 - Changed prototypes for ENGINE handler functions (init(), finish(),
6251 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6254 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6255 to the algorithm using long division. The binary algorithm can be
6256 used only if the modulus is odd. On 32-bit systems, it is faster
6257 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6258 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6259 up to 450 bits. In 64-bit environments, the binary algorithm
6260 appears to be advantageous for much longer moduli; here we use it
6261 for moduli up to 2048 bits.
6264 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6265 could not support the combine flag in choice fields.
6268 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6269 extensions from a certificate request to the certificate.
6272 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6273 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6274 file: this allows the display of the certificate about to be
6275 signed to be customised, to allow certain fields to be included
6276 or excluded and extension details. The old system didn't display
6277 multicharacter strings properly, omitted fields not in the policy
6278 and couldn't display additional details such as extensions.
6281 *) Function EC_POINTs_mul for multiple scalar multiplication
6282 of an arbitrary number of elliptic curve points
6283 \sum scalars[i]*points[i],
6284 optionally including the generator defined for the EC_GROUP:
6285 scalar*generator + \sum scalars[i]*points[i].
6287 EC_POINT_mul is a simple wrapper function for the typical case
6288 that the point list has just one item (besides the optional
6292 *) First EC_METHODs for curves over GF(p):
6294 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6295 operations and provides various method functions that can also
6296 operate with faster implementations of modular arithmetic.
6298 EC_GFp_mont_method() reuses most functions that are part of
6299 EC_GFp_simple_method, but uses Montgomery arithmetic.
6301 [Bodo Moeller; point addition and point doubling
6302 implementation directly derived from source code provided by
6303 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6305 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6306 crypto/ec/ec_lib.c):
6308 Curves are EC_GROUP objects (with an optional group generator)
6309 based on EC_METHODs that are built into the library.
6311 Points are EC_POINT objects based on EC_GROUP objects.
6313 Most of the framework would be able to handle curves over arbitrary
6314 finite fields, but as there are no obvious types for fields other
6315 than GF(p), some functions are limited to that for now.
6318 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6319 that the file contains a complete HTTP response.
6322 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6323 change the def and num file printf format specifier from "%-40sXXX"
6324 to "%-39s XXX". The latter will always guarantee a space after the
6325 field while the former will cause them to run together if the field
6326 is 40 of more characters long.
6329 *) Constify the cipher and digest 'method' functions and structures
6330 and modify related functions to take constant EVP_MD and EVP_CIPHER
6334 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6335 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6338 *) Modify EVP_Digest*() routines so they now return values. Although the
6339 internal software routines can never fail additional hardware versions
6343 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6345 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6346 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6349 ERR_R_NESTED_ASN1_ERROR
6351 ERR_R_MISSING_ASN1_EOS
6352 were 4 .. 9, conflicting with
6353 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6355 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6356 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6358 Add new error code 'ERR_R_INTERNAL_ERROR'.
6361 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6365 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6366 sets the subject name for a new request or supersedes the
6367 subject name in a given request. Formats that can be parsed are
6368 'CN=Some Name, OU=myOU, C=IT'
6370 'CN=Some Name/OU=myOU/C=IT'.
6372 Add options '-batch' and '-verbose' to 'openssl req'.
6373 [Massimiliano Pala <madwolf@hackmasters.net>]
6375 *) Introduce the possibility to access global variables through
6376 functions on platform were that's the best way to handle exporting
6377 global variables in shared libraries. To enable this functionality,
6378 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6379 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6380 is normally done by Configure or something similar).
6382 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6383 in the source file (foo.c) like this:
6385 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6386 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6388 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6389 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6391 OPENSSL_DECLARE_GLOBAL(int,foo);
6392 #define foo OPENSSL_GLOBAL_REF(foo)
6393 OPENSSL_DECLARE_GLOBAL(double,bar);
6394 #define bar OPENSSL_GLOBAL_REF(bar)
6396 The #defines are very important, and therefore so is including the
6397 header file everywhere where the defined globals are used.
6399 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6400 of ASN.1 items, but that structure is a bit different.
6402 The largest change is in util/mkdef.pl which has been enhanced with
6403 better and easier to understand logic to choose which symbols should
6404 go into the Windows .def files as well as a number of fixes and code
6405 cleanup (among others, algorithm keywords are now sorted
6406 lexicographically to avoid constant rewrites).
6409 *) In BN_div() keep a copy of the sign of 'num' before writing the
6410 result to 'rm' because if rm==num the value will be overwritten
6411 and produce the wrong result if 'num' is negative: this caused
6412 problems with BN_mod() and BN_nnmod().
6415 *) Function OCSP_request_verify(). This checks the signature on an
6416 OCSP request and verifies the signer certificate. The signer
6417 certificate is just checked for a generic purpose and OCSP request
6421 *) Add OCSP_check_validity() function to check the validity of OCSP
6422 responses. OCSP responses are prepared in real time and may only
6423 be a few seconds old. Simply checking that the current time lies
6424 between thisUpdate and nextUpdate max reject otherwise valid responses
6425 caused by either OCSP responder or client clock inaccuracy. Instead
6426 we allow thisUpdate and nextUpdate to fall within a certain period of
6427 the current time. The age of the response can also optionally be
6428 checked. Two new options -validity_period and -status_age added to
6432 *) If signature or public key algorithm is unrecognized print out its
6433 OID rather that just UNKNOWN.
6436 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6437 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6438 ID to be generated from the issuer certificate alone which can then be
6439 passed to OCSP_id_issuer_cmp().
6442 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6443 ASN1 modules to export functions returning ASN1_ITEM pointers
6444 instead of the ASN1_ITEM structures themselves. This adds several
6445 new macros which allow the underlying ASN1 function/structure to
6446 be accessed transparently. As a result code should not use ASN1_ITEM
6447 references directly (such as &X509_it) but instead use the relevant
6448 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6449 use of the new ASN1 code on platforms where exporting structures
6450 is problematical (for example in shared libraries) but exporting
6451 functions returning pointers to structures is not.
6454 *) Add support for overriding the generation of SSL/TLS session IDs.
6455 These callbacks can be registered either in an SSL_CTX or per SSL.
6456 The purpose of this is to allow applications to control, if they wish,
6457 the arbitrary values chosen for use as session IDs, particularly as it
6458 can be useful for session caching in multiple-server environments. A
6459 command-line switch for testing this (and any client code that wishes
6460 to use such a feature) has been added to "s_server".
6461 [Geoff Thorpe, Lutz Jaenicke]
6463 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6464 of the form '#if defined(...) || defined(...) || ...' and
6465 '#if !defined(...) && !defined(...) && ...'. This also avoids
6466 the growing number of special cases it was previously handling.
6469 *) Make all configuration macros available for application by making
6470 sure they are available in opensslconf.h, by giving them names starting
6471 with "OPENSSL_" to avoid conflicts with other packages and by making
6472 sure e_os2.h will cover all platform-specific cases together with
6474 Additionally, it is now possible to define configuration/platform-
6475 specific names (called "system identities"). In the C code, these
6476 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6477 macro with the name beginning with "OPENSSL_SYS_", which is determined
6478 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6482 *) New option -set_serial to 'req' and 'x509' this allows the serial
6483 number to use to be specified on the command line. Previously self
6484 signed certificates were hard coded with serial number 0 and the
6485 CA options of 'x509' had to use a serial number in a file which was
6489 *) New options to 'ca' utility to support V2 CRL entry extensions.
6490 Currently CRL reason, invalidity date and hold instruction are
6491 supported. Add new CRL extensions to V3 code and some new objects.
6494 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6495 disable standard block padding (aka PKCS#5 padding) in the EVP
6496 API, which was previously mandatory. This means that the data is
6497 not padded in any way and so the total length much be a multiple
6498 of the block size, otherwise an error occurs.
6501 *) Initial (incomplete) OCSP SSL support.
6504 *) New function OCSP_parse_url(). This splits up a URL into its host,
6505 port and path components: primarily to parse OCSP URLs. New -url
6506 option to ocsp utility.
6509 *) New nonce behavior. The return value of OCSP_check_nonce() now
6510 reflects the various checks performed. Applications can decide
6511 whether to tolerate certain situations such as an absent nonce
6512 in a response when one was present in a request: the ocsp application
6513 just prints out a warning. New function OCSP_add1_basic_nonce()
6514 this is to allow responders to include a nonce in a response even if
6515 the request is nonce-less.
6518 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6519 skipped when using openssl x509 multiple times on a single input file,
6520 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6523 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6524 set string type: to handle setting ASN1_TIME structures. Fix ca
6525 utility to correctly initialize revocation date of CRLs.
6528 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6529 the clients preferred ciphersuites and rather use its own preferences.
6530 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6531 Internet Explorer by ensuring unchanged hash method during stepup.
6532 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6535 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6536 to aes and add a new 'exist' option to print out symbols that don't
6540 *) Additional options to ocsp utility to allow flags to be set and
6541 additional certificates supplied.
6544 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6545 OCSP client a number of certificate to only verify the response
6549 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6550 handle the new API. Currently only ECB, CBC modes supported. Add new
6553 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6554 Encryption Standard (AES) Ciphersuites for Transport Layer
6555 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6556 not enabled by default and were not part of the "ALL" ciphersuite
6557 alias because they were not yet official; they could be
6558 explicitly requested by specifying the "AESdraft" ciphersuite
6559 group alias. In the final release of OpenSSL 0.9.7, the group
6560 alias is called "AES" and is part of "ALL".)
6561 [Ben Laurie, Steve Henson, Bodo Moeller]
6563 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6564 request to response.
6567 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6568 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6569 extract information from a certificate request. OCSP_response_create()
6570 creates a response and optionally adds a basic response structure.
6571 OCSP_basic_add1_status() adds a complete single response to a basic
6572 response and returns the OCSP_SINGLERESP structure just added (to allow
6573 extensions to be included for example). OCSP_basic_add1_cert() adds a
6574 certificate to a basic response and OCSP_basic_sign() signs a basic
6575 response with various flags. New helper functions ASN1_TIME_check()
6576 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6577 (converts ASN1_TIME to GeneralizedTime).
6580 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6581 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6582 structure from a certificate. X509_pubkey_digest() digests the public_key
6583 contents: this is used in various key identifiers.
6586 *) Make sk_sort() tolerate a NULL argument.
6587 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6589 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6590 passed by the function are trusted implicitly. If any of them signed the
6591 response then it is assumed to be valid and is not verified.
6594 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6595 to data. This was previously part of the PKCS7 ASN1 code. This
6596 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6597 [Steve Henson, reported by Kenneth R. Robinette
6598 <support@securenetterm.com>]
6600 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6601 routines: without these tracing memory leaks is very painful.
6602 Fix leaks in PKCS12 and PKCS7 routines.
6605 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6606 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6607 effectively meant GeneralizedTime would never be used. Now it
6608 is initialised to -1 but X509_time_adj() now has to check the value
6609 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6610 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6611 [Steve Henson, reported by Kenneth R. Robinette
6612 <support@securenetterm.com>]
6614 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6615 result in a zero length in the ASN1_INTEGER structure which was
6616 not consistent with the structure when d2i_ASN1_INTEGER() was used
6617 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6618 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6619 where it did not print out a minus for negative ASN1_INTEGER.
6622 *) Add summary printout to ocsp utility. The various functions which
6623 convert status values to strings have been renamed to:
6624 OCSP_response_status_str(), OCSP_cert_status_str() and
6625 OCSP_crl_reason_str() and are no longer static. New options
6626 to verify nonce values and to disable verification. OCSP response
6627 printout format cleaned up.
6630 *) Add additional OCSP certificate checks. These are those specified
6631 in RFC2560. This consists of two separate checks: the CA of the
6632 certificate being checked must either be the OCSP signer certificate
6633 or the issuer of the OCSP signer certificate. In the latter case the
6634 OCSP signer certificate must contain the OCSP signing extended key
6635 usage. This check is performed by attempting to match the OCSP
6636 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6637 in the OCSP_CERTID structures of the response.
6640 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6641 and related routines. This uses the standard OpenSSL certificate
6642 verify routines to perform initial checks (just CA validity) and
6643 to obtain the certificate chain. Then additional checks will be
6644 performed on the chain. Currently the root CA is checked to see
6645 if it is explicitly trusted for OCSP signing. This is used to set
6646 a root CA as a global signing root: that is any certificate that
6647 chains to that CA is an acceptable OCSP signing certificate.
6650 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6651 extensions from a separate configuration file.
6652 As when reading extensions from the main configuration file,
6653 the '-extensions ...' option may be used for specifying the
6655 [Massimiliano Pala <madwolf@comune.modena.it>]
6657 *) New OCSP utility. Allows OCSP requests to be generated or
6658 read. The request can be sent to a responder and the output
6659 parsed, outputed or printed in text form. Not complete yet:
6660 still needs to check the OCSP response validity.
6663 *) New subcommands for 'openssl ca':
6664 'openssl ca -status <serial>' prints the status of the cert with
6665 the given serial number (according to the index file).
6666 'openssl ca -updatedb' updates the expiry status of certificates
6668 [Massimiliano Pala <madwolf@comune.modena.it>]
6670 *) New '-newreq-nodes' command option to CA.pl. This is like
6671 '-newreq', but calls 'openssl req' with the '-nodes' option
6672 so that the resulting key is not encrypted.
6673 [Damien Miller <djm@mindrot.org>]
6675 *) New configuration for the GNU Hurd.
6676 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6678 *) Initial code to implement OCSP basic response verify. This
6679 is currently incomplete. Currently just finds the signer's
6680 certificate and verifies the signature on the response.
6683 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6684 value of OPENSSLDIR. This is available via the new '-d' option
6685 to 'openssl version', and is also included in 'openssl version -a'.
6688 *) Allowing defining memory allocation callbacks that will be given
6689 file name and line number information in additional arguments
6690 (a const char* and an int). The basic functionality remains, as
6691 well as the original possibility to just replace malloc(),
6692 realloc() and free() by functions that do not know about these
6693 additional arguments. To register and find out the current
6694 settings for extended allocation functions, the following
6695 functions are provided:
6697 CRYPTO_set_mem_ex_functions
6698 CRYPTO_set_locked_mem_ex_functions
6699 CRYPTO_get_mem_ex_functions
6700 CRYPTO_get_locked_mem_ex_functions
6702 These work the same way as CRYPTO_set_mem_functions and friends.
6703 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6704 extended allocation function is enabled.
6705 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6706 a conventional allocation function is enabled.
6707 [Richard Levitte, Bodo Moeller]
6709 *) Finish off removing the remaining LHASH function pointer casts.
6710 There should no longer be any prototype-casting required when using
6711 the LHASH abstraction, and any casts that remain are "bugs". See
6712 the callback types and macros at the head of lhash.h for details
6713 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6716 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6717 If /dev/[u]random devices are not available or do not return enough
6718 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6720 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6721 /etc/entropy will be queried once each in this sequence, quering stops
6722 when enough entropy was collected without querying more sockets.
6725 *) Change the Unix RAND_poll() variant to be able to poll several
6726 random devices, as specified by DEVRANDOM, until a sufficient amount
6727 of data has been collected. We spend at most 10 ms on each file
6728 (select timeout) and read in non-blocking mode. DEVRANDOM now
6729 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6730 (previously it was just the string "/dev/urandom"), so on typical
6731 platforms the 10 ms delay will never occur.
6732 Also separate out the Unix variant to its own file, rand_unix.c.
6733 For VMS, there's a currently-empty rand_vms.c.
6736 *) Move OCSP client related routines to ocsp_cl.c. These
6737 provide utility functions which an application needing
6738 to issue a request to an OCSP responder and analyse the
6739 response will typically need: as opposed to those which an
6740 OCSP responder itself would need which will be added later.
6742 OCSP_request_sign() signs an OCSP request with an API similar
6743 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6744 response. OCSP_response_get1_basic() extracts basic response
6745 from response. OCSP_resp_find_status(): finds and extracts status
6746 information from an OCSP_CERTID structure (which will be created
6747 when the request structure is built). These are built from lower
6748 level functions which work on OCSP_SINGLERESP structures but
6749 wont normally be used unless the application wishes to examine
6750 extensions in the OCSP response for example.
6752 Replace nonce routines with a pair of functions.
6753 OCSP_request_add1_nonce() adds a nonce value and optionally
6754 generates a random value. OCSP_check_nonce() checks the
6755 validity of the nonce in an OCSP response.
6758 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6759 This doesn't copy the supplied OCSP_CERTID and avoids the
6760 need to free up the newly created id. Change return type
6761 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6762 This can then be used to add extensions to the request.
6763 Deleted OCSP_request_new(), since most of its functionality
6764 is now in OCSP_REQUEST_new() (and the case insensitive name
6765 clash) apart from the ability to set the request name which
6766 will be added elsewhere.
6769 *) Update OCSP API. Remove obsolete extensions argument from
6770 various functions. Extensions are now handled using the new
6771 OCSP extension code. New simple OCSP HTTP function which
6772 can be used to send requests and parse the response.
6775 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6776 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6777 uses the special reorder version of SET OF to sort the attributes
6778 and reorder them to match the encoded order. This resolves a long
6779 standing problem: a verify on a PKCS7 structure just after signing
6780 it used to fail because the attribute order did not match the
6781 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6782 it uses the received order. This is necessary to tolerate some broken
6783 software that does not order SET OF. This is handled by encoding
6784 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6785 to produce the required SET OF.
6788 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6789 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6790 files to get correct declarations of the ASN.1 item variables.
6793 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6794 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6795 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6796 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6797 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6798 ASN1_ITEM and no wrapper functions.
6801 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6802 replace the old function pointer based I/O routines. Change most of
6803 the *_d2i_bio() and *_d2i_fp() functions to use these.
6806 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6807 lines, recognice more "algorithms" that can be deselected, and make
6808 it complain about algorithm deselection that isn't recognised.
6811 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6812 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6813 to use new functions. Add NO_ASN1_OLD which can be set to remove
6814 some old style ASN1 functions: this can be used to determine if old
6815 code will still work when these eventually go away.
6818 *) New extension functions for OCSP structures, these follow the
6819 same conventions as certificates and CRLs.
6822 *) New function X509V3_add1_i2d(). This automatically encodes and
6823 adds an extension. Its behaviour can be customised with various
6824 flags to append, replace or delete. Various wrappers added for
6825 certifcates and CRLs.
6828 *) Fix to avoid calling the underlying ASN1 print routine when
6829 an extension cannot be parsed. Correct a typo in the
6830 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6833 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
6834 entries for variables.
6837 *) Add functionality to apps/openssl.c for detecting locking
6838 problems: As the program is single-threaded, all we have
6839 to do is register a locking callback using an array for
6840 storing which locks are currently held by the program.
6843 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6844 SSL_get_ex_data_X509_STORE_idx(), which is used in
6845 ssl_verify_cert_chain() and thus can be called at any time
6846 during TLS/SSL handshakes so that thread-safety is essential.
6847 Unfortunately, the ex_data design is not at all suited
6848 for multi-threaded use, so it probably should be abolished.
6851 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6852 [Broadcom, tweaked and integrated by Geoff Thorpe]
6854 *) Move common extension printing code to new function
6855 X509V3_print_extensions(). Reorganise OCSP print routines and
6856 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6859 *) New function X509_signature_print() to remove duplication in some
6863 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6864 set (this was treated exactly the same as SET OF previously). This
6865 is used to reorder the STACK representing the structure to match the
6866 encoding. This will be used to get round a problem where a PKCS7
6867 structure which was signed could not be verified because the STACK
6868 order did not reflect the encoded order.
6871 *) Reimplement the OCSP ASN1 module using the new code.
6874 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6875 for its ASN1 operations. The old style function pointers still exist
6876 for now but they will eventually go away.
6879 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6880 completely replaces the old ASN1 functionality with a table driven
6881 encoder and decoder which interprets an ASN1_ITEM structure describing
6882 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6883 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6884 has also been converted to the new form.
6887 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6888 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6889 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6890 for negative moduli.
6893 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6894 of not touching the result's sign bit.
6897 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6901 *) Changed the LHASH code to use prototypes for callbacks, and created
6902 macros to declare and implement thin (optionally static) functions
6903 that provide type-safety and avoid function pointer casting for the
6904 type-specific callbacks.
6907 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6909 [Veers Staats <staatsvr@asc.hpc.mil>,
6910 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6912 *) Reformat the FAQ so the different questions and answers can be divided
6913 in sections depending on the subject.
6916 *) Have the zlib compression code load ZLIB.DLL dynamically under
6920 *) New function BN_mod_sqrt for computing square roots modulo a prime
6921 (using the probabilistic Tonelli-Shanks algorithm unless
6922 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6923 be handled deterministically).
6924 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6926 *) Make BN_mod_inverse faster by explicitly handling small quotients
6927 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6928 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6931 *) New function BN_kronecker.
6934 *) Fix BN_gcd so that it works on negative inputs; the result is
6935 positive unless both parameters are zero.
6936 Previously something reasonably close to an infinite loop was
6937 possible because numbers could be growing instead of shrinking
6938 in the implementation of Euclid's algorithm.
6941 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6942 sign of the number in question.
6944 Fix BN_is_word(a,w) to work correctly for w == 0.
6946 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6947 because its test if the absolute value of 'a' equals 'w'.
6948 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6949 it exists mostly for use in the implementations of BN_is_zero(),
6950 BN_is_one(), and BN_is_word().
6953 *) New function BN_swap.
6956 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6957 the exponentiation functions are more likely to produce reasonable
6958 results on negative inputs.
6961 *) Change BN_mod_mul so that the result is always non-negative.
6962 Previously, it could be negative if one of the factors was negative;
6963 I don't think anyone really wanted that behaviour.
6966 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6967 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6968 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6969 and add new functions:
6978 BN_mod_lshift1_quick
6982 These functions always generate non-negative results.
6984 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6985 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6987 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6988 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6989 be reduced modulo m.
6990 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6993 The following entry accidentily appeared in the CHANGES file
6994 distributed with OpenSSL 0.9.7. The modifications described in
6995 it do *not* apply to OpenSSL 0.9.7.
6997 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6998 was actually never needed) and in BN_mul(). The removal in BN_mul()
6999 required a small change in bn_mul_part_recursive() and the addition
7000 of the functions bn_cmp_part_words(), bn_sub_part_words() and
7001 bn_add_part_words(), which do the same thing as bn_cmp_words(),
7002 bn_sub_words() and bn_add_words() except they take arrays with
7007 *) In 'openssl passwd', verify passwords read from the terminal
7008 unless the '-salt' option is used (which usually means that
7009 verification would just waste user's time since the resulting
7010 hash is going to be compared with some given password hash)
7011 or the new '-noverify' option is used.
7013 This is an incompatible change, but it does not affect
7014 non-interactive use of 'openssl passwd' (passwords on the command
7015 line, '-stdin' option, '-in ...' option) and thus should not
7019 *) Remove all references to RSAref, since there's no more need for it.
7022 *) Make DSO load along a path given through an environment variable
7023 (SHLIB_PATH) with shl_load().
7026 *) Constify the ENGINE code as a result of BIGNUM constification.
7027 Also constify the RSA code and most things related to it. In a
7028 few places, most notable in the depth of the ASN.1 code, ugly
7029 casts back to non-const were required (to be solved at a later
7033 *) Make it so the openssl application has all engines loaded by default.
7036 *) Constify the BIGNUM routines a little more.
7039 *) Add the following functions:
7041 ENGINE_load_cswift()
7043 ENGINE_load_atalla()
7045 ENGINE_load_builtin_engines()
7047 That way, an application can itself choose if external engines that
7048 are built-in in OpenSSL shall ever be used or not. The benefit is
7049 that applications won't have to be linked with libdl or other dso
7050 libraries unless it's really needed.
7052 Changed 'openssl engine' to load all engines on demand.
7053 Changed the engine header files to avoid the duplication of some
7054 declarations (they differed!).
7057 *) 'openssl engine' can now list capabilities.
7060 *) Better error reporting in 'openssl engine'.
7063 *) Never call load_dh_param(NULL) in s_server.
7066 *) Add engine application. It can currently list engines by name and
7067 identity, and test if they are actually available.
7070 *) Improve RPM specification file by forcing symbolic linking and making
7071 sure the installed documentation is also owned by root.root.
7072 [Damien Miller <djm@mindrot.org>]
7074 *) Give the OpenSSL applications more possibilities to make use of
7075 keys (public as well as private) handled by engines.
7078 *) Add OCSP code that comes from CertCo.
7081 *) Add VMS support for the Rijndael code.
7084 *) Added untested support for Nuron crypto accelerator.
7087 *) Add support for external cryptographic devices. This code was
7088 previously distributed separately as the "engine" branch.
7089 [Geoff Thorpe, Richard Levitte]
7091 *) Rework the filename-translation in the DSO code. It is now possible to
7092 have far greater control over how a "name" is turned into a filename
7093 depending on the operating environment and any oddities about the
7094 different shared library filenames on each system.
7097 *) Support threads on FreeBSD-elf in Configure.
7100 *) Fix for SHA1 assembly problem with MASM: it produces
7101 warnings about corrupt line number information when assembling
7102 with debugging information. This is caused by the overlapping
7104 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7107 NCONF_get_number() has no error checking at all. As a replacement,
7108 NCONF_get_number_e() is defined (_e for "error checking") and is
7109 promoted strongly. The old NCONF_get_number is kept around for
7110 binary backward compatibility.
7111 Make it possible for methods to load from something other than a BIO,
7112 by providing a function pointer that is given a name instead of a BIO.
7113 For example, this could be used to load configuration data from an
7117 *) Fix for non blocking accept BIOs. Added new I/O special reason
7118 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7119 with non blocking I/O was not possible because no retry code was
7120 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7124 *) Added the beginnings of Rijndael support.
7127 *) Fix for bug in DirectoryString mask setting. Add support for
7128 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7129 to allow certificate printing to more controllable, additional
7130 'certopt' option to 'x509' to allow new printing options to be
7134 *) Clean old EAY MD5 hack from e_os.h.
7137 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7139 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7140 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7141 [Joe Orton, Steve Henson]
7143 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7145 *) Fix additional bug revealed by the NISCC test suite:
7147 Stop bug triggering large recursion when presented with
7148 certain ASN.1 tags (CVE-2003-0851)
7151 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7153 *) Fix various bugs revealed by running the NISCC test suite:
7155 Stop out of bounds reads in the ASN1 code when presented with
7156 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7158 If verify callback ignores invalid public key errors don't try to check
7159 certificate signature with the NULL public key.
7163 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7164 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7168 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7169 extra data after the compression methods not only for TLS 1.0
7170 but also for SSL 3.0 (as required by the specification).
7171 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7173 *) Change X509_certificate_type() to mark the key as exported/exportable
7174 when it's 512 *bits* long, not 512 bytes.
7177 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7179 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7180 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7181 a protocol version number mismatch like a decryption error
7182 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7185 *) Turn on RSA blinding by default in the default implementation
7186 to avoid a timing attack. Applications that don't want it can call
7187 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7188 They would be ill-advised to do so in most cases.
7189 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7191 *) Change RSA blinding code so that it works when the PRNG is not
7192 seeded (in this case, the secret RSA exponent is abused as
7193 an unpredictable seed -- if it is not unpredictable, there
7194 is no point in blinding anyway). Make RSA blinding thread-safe
7195 by remembering the creator's thread ID in rsa->blinding and
7196 having all other threads use local one-time blinding factors
7197 (this requires more computation than sharing rsa->blinding, but
7198 avoids excessive locking; and if an RSA object is not shared
7199 between threads, blinding will still be very fast).
7202 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7204 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7205 via timing by performing a MAC computation even if incorrrect
7206 block cipher padding has been found. This is a countermeasure
7207 against active attacks where the attacker has to distinguish
7208 between bad padding and a MAC verification error. (CVE-2003-0078)
7210 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7211 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7212 Martin Vuagnoux (EPFL, Ilion)]
7214 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7216 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7217 memory from it's contents. This is done with a counter that will
7218 place alternating values in each byte. This can be used to solve
7219 two issues: 1) the removal of calls to memset() by highly optimizing
7220 compilers, and 2) cleansing with other values than 0, since those can
7221 be read through on certain media, for example a swap space on disk.
7224 *) Bugfix: client side session caching did not work with external caching,
7225 because the session->cipher setting was not restored when reloading
7226 from the external cache. This problem was masked, when
7227 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7228 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7231 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7232 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7233 [Zeev Lieber <zeev-l@yahoo.com>]
7235 *) Undo an undocumented change introduced in 0.9.6e which caused
7236 repeated calls to OpenSSL_add_all_ciphers() and
7237 OpenSSL_add_all_digests() to be ignored, even after calling
7241 *) Change the default configuration reader to deal with last line not
7242 being properly terminated.
7245 *) Change X509_NAME_cmp() so it applies the special rules on handling
7246 DN values that are of type PrintableString, as well as RDNs of type
7247 emailAddress where the value has the type ia5String.
7248 [stefank@valicert.com via Richard Levitte]
7250 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7251 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7252 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7253 the bitwise-OR of the two for use by the majority of applications
7254 wanting this behaviour, and update the docs. The documented
7255 behaviour and actual behaviour were inconsistent and had been
7256 changing anyway, so this is more a bug-fix than a behavioural
7258 [Geoff Thorpe, diagnosed by Nadav Har'El]
7260 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7261 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7264 *) Fix initialization code race conditions in
7265 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7266 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7267 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7268 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7269 ssl2_get_cipher_by_char(),
7270 ssl3_get_cipher_by_char().
7271 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7273 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7274 the cached sessions are flushed, as the remove_cb() might use ex_data
7275 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7276 (see [openssl.org #212]).
7277 [Geoff Thorpe, Lutz Jaenicke]
7279 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7280 length, instead of the encoding length to d2i_ASN1_OBJECT.
7283 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7285 *) [In 0.9.6g-engine release:]
7286 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7287 [Lynn Gazis <lgazis@rainbow.com>]
7289 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7291 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7292 and get fix the header length calculation.
7293 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7294 Alon Kantor <alonk@checkpoint.com> (and others),
7297 *) Use proper error handling instead of 'assertions' in buffer
7298 overflow checks added in 0.9.6e. This prevents DoS (the
7299 assertions could call abort()).
7300 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7302 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7304 *) Add various sanity checks to asn1_get_length() to reject
7305 the ASN1 length bytes if they exceed sizeof(long), will appear
7306 negative or the content length exceeds the length of the
7308 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7310 *) Fix cipher selection routines: ciphers without encryption had no flags
7311 for the cipher strength set and where therefore not handled correctly
7312 by the selection routines (PR #130).
7315 *) Fix EVP_dsa_sha macro.
7319 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7320 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7321 that was added in OpenSSL 0.9.6d.
7323 As the countermeasure turned out to be incompatible with some
7324 broken SSL implementations, the new option is part of SSL_OP_ALL.
7325 SSL_OP_ALL is usually employed when compatibility with weird SSL
7326 implementations is desired (e.g. '-bugs' option to 's_client' and
7327 's_server'), so the new option is automatically set in many
7331 *) Changes in security patch:
7333 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7334 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7335 Air Force Materiel Command, USAF, under agreement number
7338 *) Add various sanity checks to asn1_get_length() to reject
7339 the ASN1 length bytes if they exceed sizeof(long), will appear
7340 negative or the content length exceeds the length of the
7341 supplied buffer. (CVE-2002-0659)
7342 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7344 *) Assertions for various potential buffer overflows, not known to
7346 [Ben Laurie (CHATS)]
7348 *) Various temporary buffers to hold ASCII versions of integers were
7349 too small for 64 bit platforms. (CVE-2002-0655)
7350 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7352 *) Remote buffer overflow in SSL3 protocol - an attacker could
7353 supply an oversized session ID to a client. (CVE-2002-0656)
7354 [Ben Laurie (CHATS)]
7356 *) Remote buffer overflow in SSL2 protocol - an attacker could
7357 supply an oversized client master key. (CVE-2002-0656)
7358 [Ben Laurie (CHATS)]
7360 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7362 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7363 encoded as NULL) with id-dsa-with-sha1.
7364 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7366 *) Check various X509_...() return values in apps/req.c.
7367 [Nils Larsch <nla@trustcenter.de>]
7369 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7370 an end-of-file condition would erronously be flagged, when the CRLF
7371 was just at the end of a processed block. The bug was discovered when
7372 processing data through a buffering memory BIO handing the data to a
7373 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7374 <ptsekov@syntrex.com> and Nedelcho Stanev.
7377 *) Implement a countermeasure against a vulnerability recently found
7378 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7379 before application data chunks to avoid the use of known IVs
7380 with data potentially chosen by the attacker.
7383 *) Fix length checks in ssl3_get_client_hello().
7386 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7387 to prevent ssl3_read_internal() from incorrectly assuming that
7388 ssl3_read_bytes() found application data while handshake
7389 processing was enabled when in fact s->s3->in_read_app_data was
7390 merely automatically cleared during the initial handshake.
7391 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7393 *) Fix object definitions for Private and Enterprise: they were not
7394 recognized in their shortname (=lowercase) representation. Extend
7395 obj_dat.pl to issue an error when using undefined keywords instead
7396 of silently ignoring the problem (Svenning Sorensen
7397 <sss@sss.dnsalias.net>).
7400 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7401 generators, i.e. generators other than 2 and 5. (Previously, the
7402 code did not properly initialise the 'add' and 'rem' values to
7403 BN_generate_prime().)
7405 In the new general case, we do not insist that 'generator' is
7406 actually a primitive root: This requirement is rather pointless;
7407 a generator of the order-q subgroup is just as good, if not
7411 *) Map new X509 verification errors to alerts. Discovered and submitted by
7412 Tom Wu <tom@arcot.com>.
7415 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7416 returning non-zero before the data has been completely received
7417 when using non-blocking I/O.
7418 [Bodo Moeller; problem pointed out by John Hughes]
7420 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7421 [Ben Laurie, Lutz Jaenicke]
7423 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7424 Yoram Zahavi <YoramZ@gilian.com>).
7427 *) Add information about CygWin 1.3 and on, and preserve proper
7428 configuration for the versions before that.
7429 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7431 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7432 check whether we deal with a copy of a session and do not delete from
7433 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7434 <izhar@checkpoint.com>.
7437 *) Do not store session data into the internal session cache, if it
7438 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7439 flag is set). Proposed by Aslam <aslam@funk.com>.
7442 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7446 *) [In 0.9.6d-engine release:]
7447 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7448 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7450 *) Add the configuration target linux-s390x.
7451 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7453 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7454 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7455 variable as an indication that a ClientHello message has been
7456 received. As the flag value will be lost between multiple
7457 invocations of ssl3_accept when using non-blocking I/O, the
7458 function may not be aware that a handshake has actually taken
7459 place, thus preventing a new session from being added to the
7462 To avoid this problem, we now set s->new_session to 2 instead of
7463 using a local variable.
7464 [Lutz Jaenicke, Bodo Moeller]
7466 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7467 if the SSL_R_LENGTH_MISMATCH error is detected.
7468 [Geoff Thorpe, Bodo Moeller]
7470 *) New 'shared_ldflag' column in Configure platform table.
7473 *) Fix EVP_CIPHER_mode macro.
7474 ["Dan S. Camper" <dan@bti.net>]
7476 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7477 type, we must throw them away by setting rr->length to 0.
7478 [D P Chang <dpc@qualys.com>]
7480 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7482 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7483 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7484 worked incorrectly for those cases where range = 10..._2 and
7485 3*range is two bits longer than range.)
7488 *) Only add signing time to PKCS7 structures if it is not already
7492 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7493 OBJ_ld_ce should be OBJ_id_ce.
7494 Also some ip-pda OIDs in crypto/objects/objects.txt were
7495 incorrect (cf. RFC 3039).
7496 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7498 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7499 returns early because it has nothing to do.
7500 [Andy Schneider <andy.schneider@bjss.co.uk>]
7502 *) [In 0.9.6c-engine release:]
7503 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7504 [Andy Schneider <andy.schneider@bjss.co.uk>]
7506 *) [In 0.9.6c-engine release:]
7507 Add support for Cryptographic Appliance's keyserver technology.
7508 (Use engine 'keyclient')
7509 [Cryptographic Appliances and Geoff Thorpe]
7511 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7512 is called via tools/c89.sh because arguments have to be
7513 rearranged (all '-L' options must appear before the first object
7515 [Richard Shapiro <rshapiro@abinitio.com>]
7517 *) [In 0.9.6c-engine release:]
7518 Add support for Broadcom crypto accelerator cards, backported
7520 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7522 *) [In 0.9.6c-engine release:]
7523 Add support for SureWare crypto accelerator cards from
7524 Baltimore Technologies. (Use engine 'sureware')
7525 [Baltimore Technologies and Mark Cox]
7527 *) [In 0.9.6c-engine release:]
7528 Add support for crypto accelerator cards from Accelerated
7529 Encryption Processing, www.aep.ie. (Use engine 'aep')
7530 [AEP Inc. and Mark Cox]
7532 *) Add a configuration entry for gcc on UnixWare.
7533 [Gary Benson <gbenson@redhat.com>]
7535 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7536 messages are stored in a single piece (fixed-length part and
7537 variable-length part combined) and fix various bugs found on the way.
7540 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7541 instead. BIO_gethostbyname() does not know what timeouts are
7542 appropriate, so entries would stay in cache even when they have
7544 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7546 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7547 faced with a pathologically small ClientHello fragment that does
7548 not contain client_version: Instead of aborting with an error,
7549 simply choose the highest available protocol version (i.e.,
7550 TLS 1.0 unless it is disabled). In practice, ClientHello
7551 messages are never sent like this, but this change gives us
7552 strictly correct behaviour at least for TLS.
7555 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7556 never resets s->method to s->ctx->method when called from within
7557 one of the SSL handshake functions.
7558 [Bodo Moeller; problem pointed out by Niko Baric]
7560 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7561 (sent using the client's version number) if client_version is
7562 smaller than the protocol version in use. Also change
7563 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7564 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7565 the client will at least see that alert.
7568 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7572 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7573 client receives HelloRequest while in a handshake.
7574 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7576 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7577 should end in 'break', not 'goto end' which circuments various
7578 cleanups done in state SSL_ST_OK. But session related stuff
7579 must be disabled for SSL_ST_OK in the case that we just sent a
7582 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7583 before just sending a HelloRequest.
7584 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7586 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7587 reveal whether illegal block cipher padding was found or a MAC
7588 verification error occured. (Neither SSLerr() codes nor alerts
7589 are directly visible to potential attackers, but the information
7590 may leak via logfiles.)
7592 Similar changes are not required for the SSL 2.0 implementation
7593 because the number of padding bytes is sent in clear for SSL 2.0,
7594 and the extra bytes are just ignored. However ssl/s2_pkt.c
7595 failed to verify that the purported number of padding bytes is in
7599 *) Add OpenUNIX-8 support including shared libraries
7600 (Boyd Lynn Gerber <gerberb@zenez.com>).
7603 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7604 'wristwatch attack' using huge encoding parameters (cf.
7605 James H. Manger's CRYPTO 2001 paper). Note that the
7606 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7607 encoding parameters and hence was not vulnerable.
7610 *) BN_sqr() bug fix.
7611 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7613 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7614 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7615 followed by modular reduction.
7616 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7618 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7619 equivalent based on BN_pseudo_rand() instead of BN_rand().
7622 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7623 This function was broken, as the check for a new client hello message
7624 to handle SGC did not allow these large messages.
7625 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7628 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7631 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7632 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7635 *) Rework the configuration and shared library support for Tru64 Unix.
7636 The configuration part makes use of modern compiler features and
7637 still retains old compiler behavior for those that run older versions
7638 of the OS. The shared library support part includes a variant that
7639 uses the RPATH feature, and is available through the special
7640 configuration target "alpha-cc-rpath", which will never be selected
7642 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7644 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7645 with the same message size as in ssl3_get_certificate_request().
7646 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7647 messages might inadvertently be reject as too long.
7648 [Petr Lampa <lampa@fee.vutbr.cz>]
7650 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7653 *) Modified SSL library such that the verify_callback that has been set
7654 specificly for an SSL object with SSL_set_verify() is actually being
7655 used. Before the change, a verify_callback set with this function was
7656 ignored and the verify_callback() set in the SSL_CTX at the time of
7657 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7658 to allow the necessary settings.
7661 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7662 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7663 done automatically (in contradiction to the requirements of the C
7664 standard). This made problems when used from OpenSSH.
7667 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7668 dh->length and always used
7670 BN_rand_range(priv_key, dh->p).
7672 BN_rand_range() is not necessary for Diffie-Hellman, and this
7673 specific range makes Diffie-Hellman unnecessarily inefficient if
7674 dh->length (recommended exponent length) is much smaller than the
7675 length of dh->p. We could use BN_rand_range() if the order of
7676 the subgroup was stored in the DH structure, but we only have
7681 BN_rand(priv_key, l, ...)
7683 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7689 RSA_eay_public_encrypt
7690 RSA_eay_private_decrypt
7691 RSA_eay_private_encrypt (signing)
7692 RSA_eay_public_decrypt (signature verification)
7694 (default implementations for RSA_public_encrypt,
7695 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7696 always reject numbers >= n.
7699 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7700 to synchronize access to 'locking_thread'. This is necessary on
7701 systems where access to 'locking_thread' (an 'unsigned long'
7702 variable) is not atomic.
7705 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7706 *before* setting the 'crypto_lock_rand' flag. The previous code had
7707 a race condition if 0 is a valid thread ID.
7708 [Travis Vitek <vitek@roguewave.com>]
7710 *) Add support for shared libraries under Irix.
7711 [Albert Chin-A-Young <china@thewrittenword.com>]
7713 *) Add configuration option to build on Linux on both big-endian and
7715 [Ralf Baechle <ralf@uni-koblenz.de>]
7717 *) Add the possibility to create shared libraries on HP-UX.
7720 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7722 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7723 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7724 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7725 PRNG state recovery was possible based on the output of
7726 one PRNG request appropriately sized to gain knowledge on
7727 'md' followed by enough consecutive 1-byte PRNG requests
7728 to traverse all of 'state'.
7730 1. When updating 'md_local' (the current thread's copy of 'md')
7731 during PRNG output generation, hash all of the previous
7732 'md_local' value, not just the half used for PRNG output.
7734 2. Make the number of bytes from 'state' included into the hash
7735 independent from the number of PRNG bytes requested.
7737 The first measure alone would be sufficient to avoid
7738 Markku-Juhani's attack. (Actually it had never occurred
7739 to me that the half of 'md_local' used for chaining was the
7740 half from which PRNG output bytes were taken -- I had always
7741 assumed that the secret half would be used.) The second
7742 measure makes sure that additional data from 'state' is never
7743 mixed into 'md_local' in small portions; this heuristically
7744 further strengthens the PRNG.
7747 *) Fix crypto/bn/asm/mips3.s.
7750 *) When only the key is given to "enc", the IV is undefined. Print out
7751 an error message in this case.
7754 *) Handle special case when X509_NAME is empty in X509 printing routines.
7757 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7758 positive and less than q.
7761 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7762 used: it isn't thread safe and the add_lock_callback should handle
7764 [Paul Rose <Paul.Rose@bridge.com>]
7766 *) Verify that incoming data obeys the block size in
7767 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7771 [Ulf Möller, Bodo Möller]
7773 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7774 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7775 when fixing the server behaviour for backwards-compatible 'client
7776 hello' messages. (Note that the attack is impractical against
7777 SSL 3.0 and TLS 1.0 anyway because length and version checking
7778 means that the probability of guessing a valid ciphertext is
7779 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7782 Before 0.9.5, the countermeasure (hide the error by generating a
7783 random 'decryption result') did not work properly because
7784 ERR_clear_error() was missing, meaning that SSL_get_error() would
7785 detect the supposedly ignored error.
7787 Both problems are now fixed.
7790 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7791 (previously it was 1024).
7794 *) Fix for compatibility mode trust settings: ignore trust settings
7795 unless some valid trust or reject settings are present.
7798 *) Fix for blowfish EVP: its a variable length cipher.
7801 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7802 parameters in DSA public key structures and return an error in the
7803 DSA routines if parameters are absent.
7806 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7807 in the current directory if neither $RANDFILE nor $HOME was set.
7808 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7809 caused some confusion to Windows users who haven't defined $HOME.
7810 Thus RAND_file_name() is changed again: e_os.h can define a
7811 DEFAULT_HOME, which will be used if $HOME is not set.
7812 For Windows, we use "C:"; on other platforms, we still require
7813 environment variables.
7815 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7816 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7817 having multiple threads call RAND_poll() concurrently.
7820 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7821 combination of a flag and a thread ID variable.
7822 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7823 flag), *other* threads can enter ssleay_add_bytes without obeying
7824 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7825 that they do not hold after the first thread unsets add_do_not_lock).
7828 *) Change bctest again: '-x' expressions are not available in all
7832 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7834 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7835 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7837 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7838 the default extension for executables, if any. Also, make the perl
7839 scripts that use symlink() to test if it really exists and use "cp"
7840 if it doesn't. All this made OpenSSL compilable and installable in
7844 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7845 If SEQUENCE is length is indefinite just set c->slen to the total
7846 amount of data available.
7847 [Steve Henson, reported by shige@FreeBSD.org]
7848 [This change does not apply to 0.9.7.]
7850 *) Change bctest to avoid here-documents inside command substitution
7851 (workaround for FreeBSD /bin/sh bug).
7852 For compatibility with Ultrix, avoid shell functions (introduced
7853 in the bctest version that searches along $PATH).
7856 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7857 with des_encrypt() defined on some operating systems, like Solaris
7861 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7862 On the Importance of Eliminating Errors in Cryptographic
7863 Computations, J. Cryptology 14 (2001) 2, 101-119,
7864 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7867 *) MIPS assembler BIGNUM division bug fix.
7870 *) Disabled incorrect Alpha assembler code.
7873 *) Fix PKCS#7 decode routines so they correctly update the length
7874 after reading an EOC for the EXPLICIT tag.
7876 [This change does not apply to 0.9.7.]
7878 *) Fix bug in PKCS#12 key generation routines. This was triggered
7879 if a 3DES key was generated with a 0 initial byte. Include
7880 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7881 (but broken) behaviour.
7884 *) Enhance bctest to search for a working bc along $PATH and print
7886 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7888 *) Fix memory leaks in err.c: free err_data string if necessary;
7889 don't write to the wrong index in ERR_set_error_data.
7892 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7896 *) Replace rdtsc with _emit statements for VC++ version 5.
7897 [Jeremy Cooper <jeremy@baymoo.org>]
7899 *) Make it possible to reuse SSLv2 sessions.
7902 *) In copy_email() check for >= 0 as a return value for
7903 X509_NAME_get_index_by_NID() since 0 is a valid index.
7904 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7906 *) Avoid coredump with unsupported or invalid public keys by checking if
7907 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7908 PKCS7_verify() fails with non detached data.
7911 *) Don't use getenv in library functions when run as setuid/setgid.
7912 New function OPENSSL_issetugid().
7915 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7916 due to incorrect handling of multi-threading:
7918 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7920 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7922 3. Count how many times MemCheck_off() has been called so that
7923 nested use can be treated correctly. This also avoids
7924 inband-signalling in the previous code (which relied on the
7925 assumption that thread ID 0 is impossible).
7928 *) Add "-rand" option also to s_client and s_server.
7931 *) Fix CPU detection on Irix 6.x.
7932 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7933 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7935 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7938 [This change does not apply to 0.9.7.]
7940 *) Use the cached encoding of an X509_NAME structure rather than
7941 copying it. This is apparently the reason for the libsafe "errors"
7942 but the code is actually correct.
7945 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7946 Bleichenbacher's DSA attack.
7947 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7948 to be set and top=0 forces the highest bit to be set; top=-1 is new
7949 and leaves the highest bit random.
7950 [Ulf Moeller, Bodo Moeller]
7952 *) In the NCONF_...-based implementations for CONF_... queries
7953 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7954 a temporary CONF structure with the data component set to NULL
7955 (which gives segmentation faults in lh_retrieve).
7956 Instead, use NULL for the CONF pointer in CONF_get_string and
7957 CONF_get_number (which may use environment variables) and directly
7958 return NULL from CONF_get_section.
7961 *) Fix potential buffer overrun for EBCDIC.
7964 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7965 keyUsage if basicConstraints absent for a CA.
7968 *) Make SMIME_write_PKCS7() write mail header values with a format that
7969 is more generally accepted (no spaces before the semicolon), since
7970 some programs can't parse those values properly otherwise. Also make
7971 sure BIO's that break lines after each write do not create invalid
7975 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7976 macros previously used would not encode an empty SEQUENCE OF
7977 and break the signature.
7979 [This change does not apply to 0.9.7.]
7981 *) Zero the premaster secret after deriving the master secret in
7985 *) Add some EVP_add_digest_alias registrations (as found in
7986 OpenSSL_add_all_digests()) to SSL_library_init()
7987 aka OpenSSL_add_ssl_algorithms(). This provides improved
7988 compatibility with peers using X.509 certificates
7989 with unconventional AlgorithmIdentifier OIDs.
7992 *) Fix for Irix with NO_ASM.
7993 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7995 *) ./config script fixes.
7996 [Ulf Moeller, Richard Levitte]
7998 *) Fix 'openssl passwd -1'.
8001 *) Change PKCS12_key_gen_asc() so it can cope with non null
8002 terminated strings whose length is passed in the passlen
8003 parameter, for example from PEM callbacks. This was done
8004 by adding an extra length parameter to asc2uni().
8005 [Steve Henson, reported by <oddissey@samsung.co.kr>]
8007 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
8008 call failed, free the DSA structure.
8011 *) Fix to uni2asc() to cope with zero length Unicode strings.
8012 These are present in some PKCS#12 files.
8015 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
8016 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
8017 when writing a 32767 byte record.
8018 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
8020 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
8021 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
8023 (RSA objects have a reference count access to which is protected
8024 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
8025 so they are meant to be shared between threads.)
8026 [Bodo Moeller, Geoff Thorpe; original patch submitted by
8027 "Reddie, Steven" <Steven.Reddie@ca.com>]
8029 *) Fix a deadlock in CRYPTO_mem_leaks().
8032 *) Use better test patterns in bntest.
8035 *) rand_win.c fix for Borland C.
8038 *) BN_rshift bugfix for n == 0.
8041 *) Add a 'bctest' script that checks for some known 'bc' bugs
8042 so that 'make test' does not abort just because 'bc' is broken.
8045 *) Store verify_result within SSL_SESSION also for client side to
8046 avoid potential security hole. (Re-used sessions on the client side
8047 always resulted in verify_result==X509_V_OK, not using the original
8048 result of the server certificate verification.)
8051 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
8052 SSL3_RT_APPLICATION_DATA, return 0.
8053 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
8057 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
8058 releases, have been re-implemented by renaming the previous
8059 implementations of ssl2_read and ssl3_read to ssl2_read_internal
8060 and ssl3_read_internal, respectively, and adding 'peek' parameters
8061 to them. The new ssl[23]_{read,peek} functions are calls to
8062 ssl[23]_read_internal with the 'peek' flag set appropriately.
8063 A 'peek' parameter has also been added to ssl3_read_bytes, which
8064 does the actual work for ssl3_read_internal.
8067 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
8068 the method-specific "init()" handler. Also clean up ex_data after
8069 calling the method-specific "finish()" handler. Previously, this was
8070 happening the other way round.
8073 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
8074 The previous value, 12, was not always sufficient for BN_mod_exp().
8077 *) Make sure that shared libraries get the internal name engine with
8078 the full version number and not just 0. This should mark the
8079 shared libraries as not backward compatible. Of course, this should
8080 be changed again when we can guarantee backward binary compatibility.
8083 *) Fix typo in get_cert_by_subject() in by_dir.c
8084 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8086 *) Rework the system to generate shared libraries:
8088 - Make note of the expected extension for the shared libraries and
8089 if there is a need for symbolic links from for example libcrypto.so.0
8090 to libcrypto.so.0.9.7. There is extended info in Configure for
8093 - Make as few rebuilds of the shared libraries as possible.
8095 - Still avoid linking the OpenSSL programs with the shared libraries.
8097 - When installing, install the shared libraries separately from the
8101 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8103 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8104 and not in SSL_clear because the latter is also used by the
8105 accept/connect functions; previously, the settings made by
8106 SSL_set_read_ahead would be lost during the handshake.
8107 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8109 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8110 Previously, it would create entries for disableed algorithms no
8114 *) Added several new manual pages for SSL_* function.
8117 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8119 *) In ssl23_get_client_hello, generate an error message when faced
8120 with an initial SSL 3.0/TLS record that is too small to contain the
8121 first two bytes of the ClientHello message, i.e. client_version.
8122 (Note that this is a pathologic case that probably has never happened
8123 in real life.) The previous approach was to use the version number
8124 from the record header as a substitute; but our protocol choice
8125 should not depend on that one because it is not authenticated
8126 by the Finished messages.
8129 *) More robust randomness gathering functions for Windows.
8130 [Jeffrey Altman <jaltman@columbia.edu>]
8132 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8133 not set then we don't setup the error code for issuer check errors
8134 to avoid possibly overwriting other errors which the callback does
8135 handle. If an application does set the flag then we assume it knows
8136 what it is doing and can handle the new informational codes
8140 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8141 a general "ANY" type, as such it should be able to decode anything
8142 including tagged types. However it didn't check the class so it would
8143 wrongly interpret tagged types in the same way as their universal
8144 counterpart and unknown types were just rejected. Changed so that the
8145 tagged and unknown types are handled in the same way as a SEQUENCE:
8146 that is the encoding is stored intact. There is also a new type
8147 "V_ASN1_OTHER" which is used when the class is not universal, in this
8148 case we have no idea what the actual type is so we just lump them all
8152 *) On VMS, stdout may very well lead to a file that is written to
8153 in a record-oriented fashion. That means that every write() will
8154 write a separate record, which will be read separately by the
8155 programs trying to read from it. This can be very confusing.
8157 The solution is to put a BIO filter in the way that will buffer
8158 text until a linefeed is reached, and then write everything a
8159 line at a time, so every record written will be an actual line,
8160 not chunks of lines and not (usually doesn't happen, but I've
8161 seen it once) several lines in one record. BIO_f_linebuffer() is
8164 Currently, it's a VMS-only method, because that's where it has
8165 been tested well enough.
8168 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8169 it can return incorrect results.
8170 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8171 but it was in 0.9.6-beta[12].)
8174 *) Disable the check for content being present when verifying detached
8175 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8176 include zero length content when signing messages.
8179 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8180 BIO_ctrl (for BIO pairs).
8183 *) Add DSO method for VMS.
8186 *) Bug fix: Montgomery multiplication could produce results with the
8190 *) Add RPM specification openssl.spec and modify it to build three
8191 packages. The default package contains applications, application
8192 documentation and run-time libraries. The devel package contains
8193 include files, static libraries and function documentation. The
8194 doc package contains the contents of the doc directory. The original
8195 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8198 *) Add a large number of documentation files for many SSL routines.
8199 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8201 *) Add a configuration entry for Sony News 4.
8202 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8204 *) Don't set the two most significant bits to one when generating a
8205 random number < q in the DSA library.
8208 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8209 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8210 the underlying transport is blocking) if a handshake took place.
8211 (The default behaviour is needed by applications such as s_client
8212 and s_server that use select() to determine when to use SSL_read;
8213 but for applications that know in advance when to expect data, it
8214 just makes things more complicated.)
8217 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8221 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8222 work better on such systems.
8223 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8225 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8226 Update PKCS12_parse() so it copies the friendlyName and the
8227 keyid to the certificates aux info.
8230 *) Fix bug in PKCS7_verify() which caused an infinite loop
8231 if there was more than one signature.
8232 [Sven Uszpelkat <su@celocom.de>]
8234 *) Major change in util/mkdef.pl to include extra information
8235 about each symbol, as well as presentig variables as well
8236 as functions. This change means that there's n more need
8237 to rebuild the .num files when some algorithms are excluded.
8240 *) Allow the verify time to be set by an application,
8241 rather than always using the current time.
8244 *) Phase 2 verify code reorganisation. The certificate
8245 verify code now looks up an issuer certificate by a
8246 number of criteria: subject name, authority key id
8247 and key usage. It also verifies self signed certificates
8248 by the same criteria. The main comparison function is
8249 X509_check_issued() which performs these checks.
8251 Lot of changes were necessary in order to support this
8252 without completely rewriting the lookup code.
8254 Authority and subject key identifier are now cached.
8256 The LHASH 'certs' is X509_STORE has now been replaced
8257 by a STACK_OF(X509_OBJECT). This is mainly because an
8258 LHASH can't store or retrieve multiple objects with
8259 the same hash value.
8261 As a result various functions (which were all internal
8262 use only) have changed to handle the new X509_STORE
8263 structure. This will break anything that messed round
8264 with X509_STORE internally.
8266 The functions X509_STORE_add_cert() now checks for an
8267 exact match, rather than just subject name.
8269 The X509_STORE API doesn't directly support the retrieval
8270 of multiple certificates matching a given criteria, however
8271 this can be worked round by performing a lookup first
8272 (which will fill the cache with candidate certificates)
8273 and then examining the cache for matches. This is probably
8274 the best we can do without throwing out X509_LOOKUP
8275 entirely (maybe later...).
8277 The X509_VERIFY_CTX structure has been enhanced considerably.
8279 All certificate lookup operations now go via a get_issuer()
8280 callback. Although this currently uses an X509_STORE it
8281 can be replaced by custom lookups. This is a simple way
8282 to bypass the X509_STORE hackery necessary to make this
8283 work and makes it possible to use more efficient techniques
8284 in future. A very simple version which uses a simple
8285 STACK for its trusted certificate store is also provided
8286 using X509_STORE_CTX_trusted_stack().
8288 The verify_cb() and verify() callbacks now have equivalents
8289 in the X509_STORE_CTX structure.
8291 X509_STORE_CTX also has a 'flags' field which can be used
8292 to customise the verify behaviour.
8295 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8296 excludes S/MIME capabilities.
8299 *) When a certificate request is read in keep a copy of the
8300 original encoding of the signed data and use it when outputing
8301 again. Signatures then use the original encoding rather than
8302 a decoded, encoded version which may cause problems if the
8303 request is improperly encoded.
8306 *) For consistency with other BIO_puts implementations, call
8307 buffer_write(b, ...) directly in buffer_puts instead of calling
8310 In BIO_puts, increment b->num_write as in BIO_write.
8311 [Peter.Sylvester@EdelWeb.fr]
8313 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8314 BN_zero, we may not return a BIGNUM with an array consisting of
8318 *) Avoid calling abort() from within the library when problems are
8319 detected, except if preprocessor symbols have been defined
8320 (such as REF_CHECK, BN_DEBUG etc.).
8323 *) New openssl application 'rsautl'. This utility can be
8324 used for low level RSA operations. DER public key
8325 BIO/fp routines also added.
8328 *) New Configure entry and patches for compiling on QNX 4.
8329 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8331 *) A demo state-machine implementation was sponsored by
8332 Nuron (http://www.nuron.com/) and is now available in
8333 demos/state_machine.
8336 *) New options added to the 'dgst' utility for signature
8337 generation and verification.
8340 *) Unrecognized PKCS#7 content types are now handled via a
8341 catch all ASN1_TYPE structure. This allows unsupported
8342 types to be stored as a "blob" and an application can
8343 encode and decode it manually.
8346 *) Fix various signed/unsigned issues to make a_strex.c
8348 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8350 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8351 length if passed a buffer. ASN1_INTEGER_to_BN failed
8352 if passed a NULL BN and its argument was negative.
8353 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8355 *) Modification to PKCS#7 encoding routines to output definite
8356 length encoding. Since currently the whole structures are in
8357 memory there's not real point in using indefinite length
8358 constructed encoding. However if OpenSSL is compiled with
8359 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8362 *) Added BIO_vprintf() and BIO_vsnprintf().
8365 *) Added more prefixes to parse for in the the strings written
8366 through a logging bio, to cover all the levels that are available
8367 through syslog. The prefixes are now:
8369 PANIC, EMERG, EMR => LOG_EMERG
8370 ALERT, ALR => LOG_ALERT
8371 CRIT, CRI => LOG_CRIT
8372 ERROR, ERR => LOG_ERR
8373 WARNING, WARN, WAR => LOG_WARNING
8374 NOTICE, NOTE, NOT => LOG_NOTICE
8375 INFO, INF => LOG_INFO
8376 DEBUG, DBG => LOG_DEBUG
8378 and as before, if none of those prefixes are present at the
8379 beginning of the string, LOG_ERR is chosen.
8381 On Win32, the LOG_* levels are mapped according to this:
8383 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8384 LOG_WARNING => EVENTLOG_WARNING_TYPE
8385 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8389 *) Made it possible to reconfigure with just the configuration
8390 argument "reconf" or "reconfigure". The command line arguments
8391 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8392 and are retrieved from there when reconfiguring.
8396 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8398 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8401 *) The obj_dat.pl script was messing up the sorting of object
8402 names. The reason was that it compared the quoted version
8403 of strings as a result "OCSP" > "OCSP Signing" because
8404 " > SPACE. Changed script to store unquoted versions of
8405 names and add quotes on output. It was also omitting some
8406 names from the lookup table if they were given a default
8407 value (that is if SN is missing it is given the same
8408 value as LN and vice versa), these are now added on the
8409 grounds that if an object has a name we should be able to
8410 look it up. Finally added warning output when duplicate
8411 short or long names are found.
8414 *) Changes needed for Tandem NSK.
8415 [Scott Uroff <scott@xypro.com>]
8417 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8418 RSA_padding_check_SSLv23(), special padding was never detected
8419 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8420 version rollback attacks was not effective.
8422 In s23_clnt.c, don't use special rollback-attack detection padding
8423 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8424 client; similarly, in s23_srvr.c, don't do the rollback check if
8425 SSL 2.0 is the only protocol enabled in the server.
8428 *) Make it possible to get hexdumps of unprintable data with 'openssl
8429 asn1parse'. By implication, the functions ASN1_parse_dump() and
8430 BIO_dump_indent() are added.
8433 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8434 these print out strings and name structures based on various
8435 flags including RFC2253 support and proper handling of
8436 multibyte characters. Added options to the 'x509' utility
8437 to allow the various flags to be set.
8440 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8441 Also change the functions X509_cmp_current_time() and
8442 X509_gmtime_adj() work with an ASN1_TIME structure,
8443 this will enable certificates using GeneralizedTime in validity
8444 dates to be checked.
8447 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8448 negative public key encodings) on by default,
8449 NO_NEG_PUBKEY_BUG can be set to disable it.
8452 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8453 content octets. An i2c_ASN1_OBJECT is unnecessary because
8454 the encoding can be trivially obtained from the structure.
8457 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8458 not read locks (CRYPTO_r_[un]lock).
8461 *) A first attempt at creating official support for shared
8462 libraries through configuration. I've kept it so the
8463 default is static libraries only, and the OpenSSL programs
8464 are always statically linked for now, but there are
8465 preparations for dynamic linking in place.
8466 This has been tested on Linux and Tru64.
8469 *) Randomness polling function for Win9x, as described in:
8470 Peter Gutmann, Software Generation of Practically Strong
8474 *) Fix so PRNG is seeded in req if using an already existing
8478 *) New options to smime application. -inform and -outform
8479 allow alternative formats for the S/MIME message including
8480 PEM and DER. The -content option allows the content to be
8481 specified separately. This should allow things like Netscape
8482 form signing output easier to verify.
8485 *) Fix the ASN1 encoding of tags using the 'long form'.
8488 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8489 STRING types. These convert content octets to and from the
8490 underlying type. The actual tag and length octets are
8491 already assumed to have been read in and checked. These
8492 are needed because all other string types have virtually
8493 identical handling apart from the tag. By having versions
8494 of the ASN1 functions that just operate on content octets
8495 IMPLICIT tagging can be handled properly. It also allows
8496 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8497 and ASN1_INTEGER are identical apart from the tag.
8500 *) Change the handling of OID objects as follows:
8502 - New object identifiers are inserted in objects.txt, following
8503 the syntax given in objects.README.
8504 - objects.pl is used to process obj_mac.num and create a new
8506 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8509 This is currently kind of a hack, and the perl code in objects.pl
8510 isn't very elegant, but it works as I intended. The simplest way
8511 to check that it worked correctly is to look in obj_dat.h and
8512 check the array nid_objs and make sure the objects haven't moved
8513 around (this is important!). Additions are OK, as well as
8514 consistent name changes.
8517 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8520 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8521 The given file adds to whatever has already been seeded into the
8522 random pool through the RANDFILE configuration file option or
8523 environment variable, or the default random state file.
8526 *) mkstack.pl now sorts each macro group into lexical order.
8527 Previously the output order depended on the order the files
8528 appeared in the directory, resulting in needless rewriting
8532 *) Patches to make OpenSSL compile under Win32 again. Mostly
8533 work arounds for the VC++ problem that it treats func() as
8534 func(void). Also stripped out the parts of mkdef.pl that
8535 added extra typesafe functions: these no longer exist.
8538 *) Reorganisation of the stack code. The macros are now all
8539 collected in safestack.h . Each macro is defined in terms of
8540 a "stack macro" of the form SKM_<name>(type, a, b). The
8541 DEBUG_SAFESTACK is now handled in terms of function casts,
8542 this has the advantage of retaining type safety without the
8543 use of additional functions. If DEBUG_SAFESTACK is not defined
8544 then the non typesafe macros are used instead. Also modified the
8545 mkstack.pl script to handle the new form. Needs testing to see
8546 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8547 the default if no major problems. Similar behaviour for ASN1_SET_OF
8548 and PKCS12_STACK_OF.
8551 *) When some versions of IIS use the 'NET' form of private key the
8552 key derivation algorithm is different. Normally MD5(password) is
8553 used as a 128 bit RC4 key. In the modified case
8554 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
8555 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8556 as the old Netscape_RSA functions except they have an additional
8557 'sgckey' parameter which uses the modified algorithm. Also added
8558 an -sgckey command line option to the rsa utility. Thanks to
8559 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8560 algorithm to openssl-dev.
8563 *) The evp_local.h macros were using 'c.##kname' which resulted in
8564 invalid expansion on some systems (SCO 5.0.5 for example).
8565 Corrected to 'c.kname'.
8566 [Phillip Porch <root@theporch.com>]
8568 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8569 a STACK of email addresses from a certificate or request, these look
8570 in the subject name and the subject alternative name extensions and
8571 omit any duplicate addresses.
8574 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8575 This makes DSA verification about 2 % faster.
8578 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8579 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8580 plus overhead for 1024 bit moduli).
8581 This makes exponentiations about 0.5 % faster for 1024 bit
8582 exponents (as measured by "openssl speed rsa2048").
8585 *) Rename memory handling macros to avoid conflicts with other
8587 Malloc => OPENSSL_malloc
8588 Malloc_locked => OPENSSL_malloc_locked
8589 Realloc => OPENSSL_realloc
8590 Free => OPENSSL_free
8593 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8594 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8597 *) CygWin32 support.
8598 [John Jarvie <jjarvie@newsguy.com>]
8600 *) The type-safe stack code has been rejigged. It is now only compiled
8601 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8602 by default all type-specific stack functions are "#define"d back to
8603 standard stack functions. This results in more streamlined output
8604 but retains the type-safety checking possibilities of the original
8608 *) The STACK code has been cleaned up, and certain type declarations
8609 that didn't make a lot of sense have been brought in line. This has
8610 also involved a cleanup of sorts in safestack.h to more correctly
8611 map type-safe stack functions onto their plain stack counterparts.
8612 This work has also resulted in a variety of "const"ifications of
8613 lots of the code, especially "_cmp" operations which should normally
8614 be prototyped with "const" parameters anyway.
8617 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8618 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8619 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8620 where all of 'md' is used each time the PRNG is used, but 'state'
8621 is used only indexed by a cyclic counter. As entropy may not be
8622 well distributed from the beginning, 'md' is important as a
8623 chaining variable. However, the output function chains only half
8624 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8625 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8626 in all of 'state' being rewritten, with the new values depending
8627 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8630 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8631 the handshake is continued after ssl_verify_cert_chain();
8632 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8633 can lead to 'unexplainable' connection aborts later.
8634 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8636 *) Major EVP API cipher revision.
8637 Add hooks for extra EVP features. This allows various cipher
8638 parameters to be set in the EVP interface. Support added for variable
8639 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8640 setting of RC2 and RC5 parameters.
8642 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8645 Remove lots of duplicated code from the EVP library. For example *every*
8646 cipher init() function handles the 'iv' in the same way according to the
8647 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8648 for CFB and OFB modes they zero ctx->num.
8650 New functionality allows removal of S/MIME code RC2 hack.
8652 Most of the routines have the same form and so can be declared in terms
8655 By shifting this to the top level EVP_CipherInit() it can be removed from
8656 all individual ciphers. If the cipher wants to handle IVs or keys
8657 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8660 Change lots of functions like EVP_EncryptUpdate() to now return a
8661 value: although software versions of the algorithms cannot fail
8662 any installed hardware versions can.
8665 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8666 this option is set, tolerate broken clients that send the negotiated
8667 protocol version number instead of the requested protocol version
8671 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8672 i.e. non-zero for export ciphersuites, zero otherwise.
8673 Previous versions had this flag inverted, inconsistent with
8674 rsa_tmp_cb (..._TMP_RSA_CB).
8675 [Bodo Moeller; problem reported by Amit Chopra]
8677 *) Add missing DSA library text string. Work around for some IIS
8678 key files with invalid SEQUENCE encoding.
8681 *) Add a document (doc/standards.txt) that list all kinds of standards
8682 and so on that are implemented in OpenSSL.
8685 *) Enhance c_rehash script. Old version would mishandle certificates
8686 with the same subject name hash and wouldn't handle CRLs at all.
8687 Added -fingerprint option to crl utility, to support new c_rehash
8691 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8694 *) Fix for SSL server purpose checking. Server checking was
8695 rejecting certificates which had extended key usage present
8696 but no ssl client purpose.
8697 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8699 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8700 is a little unclear about how a blank password is handled.
8701 Since the password in encoded as a BMPString with terminating
8702 double NULL a zero length password would end up as just the
8703 double NULL. However no password at all is different and is
8704 handled differently in the PKCS#12 key generation code. NS
8705 treats a blank password as zero length. MSIE treats it as no
8706 password on export: but it will try both on import. We now do
8707 the same: PKCS12_parse() tries zero length and no password if
8708 the password is set to "" or NULL (NULL is now a valid password:
8709 it wasn't before) as does the pkcs12 application.
8712 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8713 perror when PEM_read_bio_X509_REQ fails, the error message must
8714 be obtained from the error queue.
8717 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8718 it in ERR_remove_state if appropriate, and change ERR_get_state
8719 accordingly to avoid race conditions (this is necessary because
8720 thread_hash is no longer constant once set).
8723 *) Bugfix for linux-elf makefile.one.
8726 *) RSA_get_default_method() will now cause a default
8727 RSA_METHOD to be chosen if one doesn't exist already.
8728 Previously this was only set during a call to RSA_new()
8729 or RSA_new_method(NULL) meaning it was possible for
8730 RSA_get_default_method() to return NULL.
8733 *) Added native name translation to the existing DSO code
8734 that will convert (if the flag to do so is set) filenames
8735 that are sufficiently small and have no path information
8736 into a canonical native form. Eg. "blah" converted to
8737 "libblah.so" or "blah.dll" etc.
8740 *) New function ERR_error_string_n(e, buf, len) which is like
8741 ERR_error_string(e, buf), but writes at most 'len' bytes
8742 including the 0 terminator. For ERR_error_string_n, 'buf'
8744 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8746 *) CONF library reworked to become more general. A new CONF
8747 configuration file reader "class" is implemented as well as a
8748 new functions (NCONF_*, for "New CONF") to handle it. The now
8749 old CONF_* functions are still there, but are reimplemented to
8750 work in terms of the new functions. Also, a set of functions
8751 to handle the internal storage of the configuration data is
8752 provided to make it easier to write new configuration file
8753 reader "classes" (I can definitely see something reading a
8754 configuration file in XML format, for example), called _CONF_*,
8755 or "the configuration storage API"...
8757 The new configuration file reading functions are:
8759 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8760 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8762 NCONF_default, NCONF_WIN32
8764 NCONF_dump_fp, NCONF_dump_bio
8766 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8767 NCONF_new creates a new CONF object. This works in the same way
8768 as other interfaces in OpenSSL, like the BIO interface.
8769 NCONF_dump_* dump the internal storage of the configuration file,
8770 which is useful for debugging. All other functions take the same
8771 arguments as the old CONF_* functions wth the exception of the
8772 first that must be a `CONF *' instead of a `LHASH *'.
8774 To make it easer to use the new classes with the old CONF_* functions,
8775 the function CONF_set_default_method is provided.
8778 *) Add '-tls1' option to 'openssl ciphers', which was already
8779 mentioned in the documentation but had not been implemented.
8780 (This option is not yet really useful because even the additional
8781 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8784 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8785 OpenSSL-based applications) load shared libraries and bind to
8786 them in a portable way.
8787 [Geoff Thorpe, with contributions from Richard Levitte]
8789 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8791 *) Make sure _lrotl and _lrotr are only used with MSVC.
8793 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8794 (the default implementation of RAND_status).
8796 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8797 to '-clrext' (= clear extensions), as intended and documented.
8798 [Bodo Moeller; inconsistency pointed out by Michael Attili
8801 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8802 was larger than the MD block size.
8803 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8805 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8806 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8807 using the passed key: if the passed key was a private key the result
8808 of X509_print(), for example, would be to print out all the private key
8812 *) des_quad_cksum() byte order bug fix.
8813 [Ulf Möller, using the problem description in krb4-0.9.7, where
8814 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8816 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8818 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8820 *) For easily testing in shell scripts whether some command
8821 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8822 returns with exit code 0 iff no command of the given name is available.
8823 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8824 the output goes to stdout and nothing is printed to stderr.
8825 Additional arguments are always ignored.
8827 Since for each cipher there is a command of the same name,
8828 the 'no-cipher' compilation switches can be tested this way.
8830 ('openssl no-XXX' is not able to detect pseudo-commands such
8831 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8834 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8837 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8838 is set; it will be thrown away anyway because each handshake creates
8840 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8841 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8842 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
8843 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8846 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8847 'Q' and 'R' lose their special meanings (quit/renegotiate).
8848 This is part of what -quiet does; unlike -quiet, -ign_eof
8849 does not suppress any output.
8852 *) Add compatibility options to the purpose and trust code. The
8853 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8854 accepts a certificate or CA, this was the previous behaviour,
8855 with all the associated security issues.
8857 X509_TRUST_COMPAT is the old trust behaviour: only and
8858 automatically trust self signed roots in certificate store. A
8859 new trust setting X509_TRUST_DEFAULT is used to specify that
8860 a purpose has no associated trust setting and it should instead
8861 use the value in the default purpose.
8864 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8865 and fix a memory leak.
8868 *) In util/mkerr.pl (which implements 'make errors'), preserve
8869 reason strings from the previous version of the .c file, as
8870 the default to have only downcase letters (and digits) in
8871 automatically generated reasons codes is not always appropriate.
8874 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8875 using strerror. Previously, ERR_reason_error_string() returned
8876 library names as reason strings for SYSerr; but SYSerr is a special
8877 case where small numbers are errno values, not library numbers.
8880 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8881 converts DSA parameters into DH parameters. (When creating parameters,
8882 DSA_generate_parameters is used.)
8885 *) Include 'length' (recommended exponent length) in C code generated
8886 by 'openssl dhparam -C'.
8889 *) The second argument to set_label in perlasm was already being used
8890 so couldn't be used as a "file scope" flag. Moved to third argument
8894 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8895 instead of RAND_bytes for encryption IVs and salts.
8898 *) Include RAND_status() into RAND_METHOD instead of implementing
8899 it only for md_rand.c Otherwise replacing the PRNG by calling
8900 RAND_set_rand_method would be impossible.
8903 *) Don't let DSA_generate_key() enter an infinite loop if the random
8904 number generation fails.
8907 *) New 'rand' application for creating pseudo-random output.
8910 *) Added configuration support for Linux/IA64
8911 [Rolf Haberrecker <rolf@suse.de>]
8913 *) Assembler module support for Mingw32.
8916 *) Shared library support for HPUX (in shlib/).
8917 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8919 *) Shared library support for Solaris gcc.
8920 [Lutz Behnke <behnke@trustcenter.de>]
8922 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8924 *) PKCS7_encrypt() was adding text MIME headers twice because they
8925 were added manually and by SMIME_crlf_copy().
8928 *) In bntest.c don't call BN_rand with zero bits argument.
8929 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8931 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8932 case was implemented. This caused BN_div_recp() to fail occasionally.
8935 *) Add an optional second argument to the set_label() in the perl
8936 assembly language builder. If this argument exists and is set
8937 to 1 it signals that the assembler should use a symbol whose
8938 scope is the entire file, not just the current function. This
8939 is needed with MASM which uses the format label:: for this scope.
8940 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8942 *) Change the ASN1 types so they are typedefs by default. Before
8943 almost all types were #define'd to ASN1_STRING which was causing
8944 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8948 *) Change names of new functions to the new get1/get0 naming
8949 convention: After 'get1', the caller owns a reference count
8950 and has to call ..._free; 'get0' returns a pointer to some
8951 data structure without incrementing reference counters.
8952 (Some of the existing 'get' functions increment a reference
8953 counter, some don't.)
8954 Similarly, 'set1' and 'add1' functions increase reference
8955 counters or duplicate objects.
8958 *) Allow for the possibility of temp RSA key generation failure:
8959 the code used to assume it always worked and crashed on failure.
8962 *) Fix potential buffer overrun problem in BIO_printf().
8963 [Ulf Möller, using public domain code by Patrick Powell; problem
8964 pointed out by David Sacerdote <das33@cornell.edu>]
8966 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8967 RAND_egd() and RAND_status(). In the command line application,
8968 the EGD socket can be specified like a seed file using RANDFILE
8972 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8973 Some CAs (e.g. Verisign) distribute certificates in this form.
8976 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8977 list to exclude them. This means that no special compilation option
8978 is needed to use anonymous DH: it just needs to be included in the
8982 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8983 EVP_MD_type. The old functionality is available in a new macro called
8984 EVP_MD_md(). Change code that uses it and update docs.
8987 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8988 where the 'void *' argument is replaced by a function pointer argument.
8989 Previously 'void *' was abused to point to functions, which works on
8990 many platforms, but is not correct. As these functions are usually
8991 called by macros defined in OpenSSL header files, most source code
8992 should work without changes.
8995 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8996 sections with information on -D... compiler switches used for
8997 compiling the library so that applications can see them. To enable
8998 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8999 must be defined. E.g.,
9000 #define OPENSSL_ALGORITHM_DEFINES
9001 #include <openssl/opensslconf.h>
9002 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
9003 [Richard Levitte, Ulf and Bodo Möller]
9005 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
9009 *) Change the 'other' type in certificate aux info to a STACK_OF
9010 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
9011 the required ASN1 format: arbitrary types determined by an OID.
9014 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
9015 argument to 'req'. This is not because the function is newer or
9016 better than others it just uses the work 'NEW' in the certificate
9017 request header lines. Some software needs this.
9020 *) Reorganise password command line arguments: now passwords can be
9021 obtained from various sources. Delete the PEM_cb function and make
9022 it the default behaviour: i.e. if the callback is NULL and the
9023 usrdata argument is not NULL interpret it as a null terminated pass
9024 phrase. If usrdata and the callback are NULL then the pass phrase
9025 is prompted for as usual.
9028 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
9029 the support is automatically enabled. The resulting binaries will
9030 autodetect the card and use it if present.
9031 [Ben Laurie and Compaq Inc.]
9033 *) Work around for Netscape hang bug. This sends certificate request
9034 and server done in one record. Since this is perfectly legal in the
9035 SSL/TLS protocol it isn't a "bug" option and is on by default. See
9036 the bugs/SSLv3 entry for more info.
9039 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
9042 *) Add -rand argument to smime and pkcs12 applications and read/write
9046 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
9049 *) Add command line password options to the remaining applications.
9052 *) Bug fix for BN_div_recp() for numerators with an even number of
9056 *) More tests in bntest.c, and changed test_bn output.
9059 *) ./config recognizes MacOS X now.
9062 *) Bug fix for BN_div() when the first words of num and divsor are
9063 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
9066 *) Add support for various broken PKCS#8 formats, and command line
9067 options to produce them.
9070 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
9071 get temporary BIGNUMs from a BN_CTX.
9074 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
9078 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
9079 include a #define from the old name to the new. The original intent
9080 was that statically linked binaries could for example just call
9081 SSLeay_add_all_ciphers() to just add ciphers to the table and not
9082 link with digests. This never worked becayse SSLeay_add_all_digests()
9083 and SSLeay_add_all_ciphers() were in the same source file so calling
9084 one would link with the other. They are now in separate source files.
9087 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9090 *) Use a less unusual form of the Miller-Rabin primality test (it used
9091 a binary algorithm for exponentiation integrated into the Miller-Rabin
9092 loop, our standard modexp algorithms are faster).
9095 *) Support for the EBCDIC character set completed.
9096 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9098 *) Source code cleanups: use const where appropriate, eliminate casts,
9099 use void * instead of char * in lhash.
9102 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9103 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9104 this the server could overwrite ephemeral keys that the client
9108 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9109 using 50 iterations of the Rabin-Miller test.
9111 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9112 iterations of the Rabin-Miller test as required by the appendix
9113 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9114 As BN_is_prime_fasttest includes trial division, DSA parameter
9115 generation becomes much faster.
9117 This implies a change for the callback functions in DSA_is_prime
9118 and DSA_generate_parameters: The callback function is called once
9119 for each positive witness in the Rabin-Miller test, not just
9120 occasionally in the inner loop; and the parameters to the
9121 callback function now provide an iteration count for the outer
9122 loop rather than for the current invocation of the inner loop.
9123 DSA_generate_parameters additionally can call the callback
9124 function with an 'iteration count' of -1, meaning that a
9125 candidate has passed the trial division test (when q is generated
9126 from an application-provided seed, trial division is skipped).
9129 *) New function BN_is_prime_fasttest that optionally does trial
9130 division before starting the Rabin-Miller test and has
9131 an additional BN_CTX * argument (whereas BN_is_prime always
9132 has to allocate at least one BN_CTX).
9133 'callback(1, -1, cb_arg)' is called when a number has passed the
9134 trial division stage.
9137 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9141 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9144 *) New function BN_pseudo_rand().
9147 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9148 bignum version of BN_from_montgomery() with the working code from
9149 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9153 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9154 made it impossible to use the same SSL_SESSION data structure in
9155 SSL2 clients in multiple threads.
9158 *) The return value of RAND_load_file() no longer counts bytes obtained
9159 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9160 to seed the PRNG (previously an explicit byte count was required).
9161 [Ulf Möller, Bodo Möller]
9163 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9164 used (char *) instead of (void *) and had casts all over the place.
9167 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9170 *) Retain source code compatibility for BN_prime_checks macro:
9171 BN_is_prime(..., BN_prime_checks, ...) now uses
9172 BN_prime_checks_for_size to determine the appropriate number of
9173 Rabin-Miller iterations.
9176 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9177 DH_CHECK_P_NOT_SAFE_PRIME.
9178 (Check if this is true? OpenPGP calls them "strong".)
9181 *) Merge the functionality of "dh" and "gendh" programs into a new program
9182 "dhparam". The old programs are retained for now but will handle DH keys
9183 (instead of parameters) in future.
9186 *) Make the ciphers, s_server and s_client programs check the return values
9187 when a new cipher list is set.
9190 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9191 ciphers. Before when the 56bit ciphers were enabled the sorting was
9194 The syntax for the cipher sorting has been extended to support sorting by
9195 cipher-strength (using the strength_bits hard coded in the tables).
9196 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9198 Fix a bug in the cipher-command parser: when supplying a cipher command
9199 string with an "undefined" symbol (neither command nor alphanumeric
9200 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9201 an error is flagged.
9203 Due to the strength-sorting extension, the code of the
9204 ssl_create_cipher_list() function was completely rearranged. I hope that
9205 the readability was also increased :-)
9206 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9208 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9209 for the first serial number and places 2 in the serial number file. This
9210 avoids problems when the root CA is created with serial number zero and
9211 the first user certificate has the same issuer name and serial number
9215 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9216 the new code. Add documentation for this stuff.
9219 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9220 X509_*() to X509at_*() on the grounds that they don't handle X509
9221 structures and behave in an analagous way to the X509v3 functions:
9222 they shouldn't be called directly but wrapper functions should be used
9225 So we also now have some wrapper functions that call the X509at functions
9226 when passed certificate requests. (TO DO: similar things can be done with
9227 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9228 things. Some of these need some d2i or i2d and print functionality
9229 because they handle more complex structures.)
9232 *) Add missing #ifndefs that caused missing symbols when building libssl
9233 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9234 NO_RSA in ssl/s2*.c.
9235 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9237 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9238 has a return value which indicates the quality of the random data
9239 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9240 error queue. New function RAND_pseudo_bytes() generates output that is
9241 guaranteed to be unique but not unpredictable. RAND_add is like
9242 RAND_seed, but takes an extra argument for an entropy estimate
9243 (RAND_seed always assumes full entropy).
9246 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9247 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9248 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9249 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9250 false-positive rate of at most 2^-80 for random input.
9253 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9256 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9257 in the 0.9.5 release), this returns the chain
9258 from an X509_CTX structure with a dup of the stack and all
9259 the X509 reference counts upped: so the stack will exist
9260 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9263 Also make SSL_SESSION_print() print out the verify return
9267 *) Add manpage for the pkcs12 command. Also change the default
9268 behaviour so MAC iteration counts are used unless the new
9269 -nomaciter option is used. This improves file security and
9270 only older versions of MSIE (4.0 for example) need it.
9273 *) Honor the no-xxx Configure options when creating .DEF files.
9276 *) Add PKCS#10 attributes to field table: challengePassword,
9277 unstructuredName and unstructuredAddress. These are taken from
9278 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9279 international characters are used.
9281 More changes to X509_ATTRIBUTE code: allow the setting of types
9282 based on strings. Remove the 'loc' parameter when adding
9283 attributes because these will be a SET OF encoding which is sorted
9287 *) Initial changes to the 'req' utility to allow request generation
9288 automation. This will allow an application to just generate a template
9289 file containing all the field values and have req construct the
9292 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9293 used all over the place including certificate requests and PKCS#7
9294 structures. They are currently handled manually where necessary with
9295 some primitive wrappers for PKCS#7. The new functions behave in a
9296 manner analogous to the X509 extension functions: they allow
9297 attributes to be looked up by NID and added.
9299 Later something similar to the X509V3 code would be desirable to
9300 automatically handle the encoding, decoding and printing of the
9301 more complex types. The string types like challengePassword can
9302 be handled by the string table functions.
9304 Also modified the multi byte string table handling. Now there is
9305 a 'global mask' which masks out certain types. The table itself
9306 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9307 is useful when for example there is only one permissible type
9308 (as in countryName) and using the mask might result in no valid
9312 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9313 SSL_get_peer_finished to allow applications to obtain the latest
9314 Finished messages sent to the peer or expected from the peer,
9315 respectively. (SSL_get_peer_finished is usually the Finished message
9316 actually received from the peer, otherwise the protocol will be aborted.)
9318 As the Finished message are message digests of the complete handshake
9319 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9320 be used for external authentication procedures when the authentication
9321 provided by SSL/TLS is not desired or is not enough.
9324 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9325 the host supports BWX extension and if Compaq C is present on the
9326 $PATH. Just exploiting of the BWX extension results in 20-30%
9327 performance kick for some algorithms, e.g. DES and RC4 to mention
9328 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9332 *) Add support for MS "fast SGC". This is arguably a violation of the
9333 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9334 weak crypto and after checking the certificate is SGC a second one
9335 with strong crypto. MS SGC stops the first handshake after receiving
9336 the server certificate message and sends a second client hello. Since
9337 a server will typically do all the time consuming operations before
9338 expecting any further messages from the client (server key exchange
9339 is the most expensive) there is little difference between the two.
9341 To get OpenSSL to support MS SGC we have to permit a second client
9342 hello message after we have sent server done. In addition we have to
9343 reset the MAC if we do get this second client hello.
9346 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9347 if a DER encoded private key is RSA or DSA traditional format. Changed
9348 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9349 format DER encoded private key. Newer code should use PKCS#8 format which
9350 has the key type encoded in the ASN1 structure. Added DER private key
9351 support to pkcs8 application.
9354 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9355 ciphersuites has been selected (as required by the SSL 3/TLS 1
9356 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9357 is set, we interpret this as a request to violate the specification
9358 (the worst that can happen is a handshake failure, and 'correct'
9359 behaviour would result in a handshake failure anyway).
9362 *) In SSL_CTX_add_session, take into account that there might be multiple
9363 SSL_SESSION structures with the same session ID (e.g. when two threads
9364 concurrently obtain them from an external cache).
9365 The internal cache can handle only one SSL_SESSION with a given ID,
9366 so if there's a conflict, we now throw out the old one to achieve
9370 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9371 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9372 some routines that use cipher OIDs: some ciphers do not have OIDs
9373 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9377 *) Simplify the trust setting structure and code. Now we just have
9378 two sequences of OIDs for trusted and rejected settings. These will
9379 typically have values the same as the extended key usage extension
9380 and any application specific purposes.
9382 The trust checking code now has a default behaviour: it will just
9383 check for an object with the same NID as the passed id. Functions can
9384 be provided to override either the default behaviour or the behaviour
9385 for a given id. SSL client, server and email already have functions
9386 in place for compatibility: they check the NID and also return "trusted"
9387 if the certificate is self signed.
9390 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9391 traditional format into an EVP_PKEY structure.
9394 *) Add a password callback function PEM_cb() which either prompts for
9395 a password if usr_data is NULL or otherwise assumes it is a null
9396 terminated password. Allow passwords to be passed on command line
9397 environment or config files in a few more utilities.
9400 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9401 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9402 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9403 Update documentation.
9406 *) Support for ASN1 "NULL" type. This could be handled before by using
9407 ASN1_TYPE but there wasn't any function that would try to read a NULL
9408 and produce an error if it couldn't. For compatibility we also have
9409 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9410 don't allocate anything because they don't need to.
9413 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9415 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9417 *) Rebuild of the memory allocation routines used by OpenSSL code and
9418 possibly others as well. The purpose is to make an interface that
9419 provide hooks so anyone can build a separate set of allocation and
9420 deallocation routines to be used by OpenSSL, for example memory
9421 pool implementations, or something else, which was previously hard
9422 since Malloc(), Realloc() and Free() were defined as macros having
9423 the values malloc, realloc and free, respectively (except for Win32
9424 compilations). The same is provided for memory debugging code.
9425 OpenSSL already comes with functionality to find memory leaks, but
9426 this gives people a chance to debug other memory problems.
9428 With these changes, a new set of functions and macros have appeared:
9430 CRYPTO_set_mem_debug_functions() [F]
9431 CRYPTO_get_mem_debug_functions() [F]
9432 CRYPTO_dbg_set_options() [F]
9433 CRYPTO_dbg_get_options() [F]
9434 CRYPTO_malloc_debug_init() [M]
9436 The memory debug functions are NULL by default, unless the library
9437 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9438 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9439 gives the standard debugging functions that come with OpenSSL) or
9440 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9441 provided by the library user) must be used. When the standard
9442 debugging functions are used, CRYPTO_dbg_set_options can be used to
9443 request additional information:
9444 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9445 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9447 Also, things like CRYPTO_set_mem_functions will always give the
9448 expected result (the new set of functions is used for allocation
9449 and deallocation) at all times, regardless of platform and compiler
9452 To finish it up, some functions that were never use in any other
9453 way than through macros have a new API and new semantic:
9456 CRYPTO_dbg_realloc()
9459 All macros of value have retained their old syntax.
9460 [Richard Levitte and Bodo Moeller]
9462 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9463 ordering of SMIMECapabilities wasn't in "strength order" and there
9464 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9468 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9469 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9470 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9472 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9473 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9474 functionality to handle multipart/signed properly) and a utility
9475 called 'smime' to call all this stuff. This is based on code I
9476 originally wrote for Celo who have kindly allowed it to be
9477 included in OpenSSL.
9480 *) Add variants des_set_key_checked and des_set_key_unchecked of
9481 des_set_key (aka des_key_sched). Global variable des_check_key
9482 decides which of these is called by des_set_key; this way
9483 des_check_key behaves as it always did, but applications and
9484 the library itself, which was buggy for des_check_key == 1,
9485 have a cleaner way to pick the version they need.
9488 *) New function PKCS12_newpass() which changes the password of a
9492 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9493 dynamic mix. In both cases the ids can be used as an index into the
9494 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9495 functions so they accept a list of the field values and the
9496 application doesn't need to directly manipulate the X509_TRUST
9500 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9504 *) Modify the way the V3 extension code looks up extensions. This now
9505 works in a similar way to the object code: we have some "standard"
9506 extensions in a static table which is searched with OBJ_bsearch()
9507 and the application can add dynamic ones if needed. The file
9508 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9509 updated whenever a new extension is added to the core code and kept
9510 in ext_nid order. There is a simple program 'tabtest.c' which checks
9511 this. New extensions are not added too often so this file can readily
9512 be maintained manually.
9514 There are two big advantages in doing things this way. The extensions
9515 can be looked up immediately and no longer need to be "added" using
9516 X509V3_add_standard_extensions(): this function now does nothing.
9517 [Side note: I get *lots* of email saying the extension code doesn't
9518 work because people forget to call this function]
9519 Also no dynamic allocation is done unless new extensions are added:
9520 so if we don't add custom extensions there is no need to call
9521 X509V3_EXT_cleanup().
9524 *) Modify enc utility's salting as follows: make salting the default. Add a
9525 magic header, so unsalted files fail gracefully instead of just decrypting
9526 to garbage. This is because not salting is a big security hole, so people
9527 should be discouraged from doing it.
9530 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9531 digest to be passed on the command line but it only used this
9532 parameter when signing a certificate. Modified so all relevant
9533 operations are affected by the digest parameter including the
9534 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9535 DSA key was used because it didn't fix the digest.
9538 *) Initial certificate chain verify code. Currently tests the untrusted
9539 certificates for consistency with the verify purpose (which is set
9540 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9542 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9543 this is because it will reject chains with invalid extensions whereas
9544 every previous version of OpenSSL and SSLeay made no checks at all.
9546 Trust code: checks the root CA for the relevant trust settings. Trust
9547 settings have an initial value consistent with the verify purpose: e.g.
9548 if the verify purpose is for SSL client use it expects the CA to be
9549 trusted for SSL client use. However the default value can be changed to
9550 permit custom trust settings: one example of this would be to only trust
9551 certificates from a specific "secure" set of CAs.
9553 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9554 which should be used for version portability: especially since the
9555 verify structure is likely to change more often now.
9557 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9558 to set them. If not set then assume SSL clients will verify SSL servers
9561 Two new options to the verify program: -untrusted allows a set of
9562 untrusted certificates to be passed in and -purpose which sets the
9563 intended purpose of the certificate. If a purpose is set then the
9564 new chain verify code is used to check extension consistency.
9567 *) Support for the authority information access extension.
9570 *) Modify RSA and DSA PEM read routines to transparently handle
9571 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9572 public keys in a format compatible with certificate
9573 SubjectPublicKeyInfo structures. Unfortunately there were already
9574 functions called *_PublicKey_* which used various odd formats so
9575 these are retained for compatibility: however the DSA variants were
9576 never in a public release so they have been deleted. Changed dsa/rsa
9577 utilities to handle the new format: note no releases ever handled public
9578 keys so we should be OK.
9580 The primary motivation for this change is to avoid the same fiasco
9581 that dogs private keys: there are several incompatible private key
9582 formats some of which are standard and some OpenSSL specific and
9583 require various evil hacks to allow partial transparent handling and
9584 even then it doesn't work with DER formats. Given the option anything
9585 other than PKCS#8 should be dumped: but the other formats have to
9586 stay in the name of compatibility.
9588 With public keys and the benefit of hindsight one standard format
9589 is used which works with EVP_PKEY, RSA or DSA structures: though
9590 it clearly returns an error if you try to read the wrong kind of key.
9592 Added a -pubkey option to the 'x509' utility to output the public key.
9593 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9594 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9595 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9596 that do the same as the EVP_PKEY_assign_*() except they up the
9597 reference count of the added key (they don't "swallow" the
9601 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9602 CRLs would fail if the file contained no certificates or no CRLs:
9603 added a new function to read in both types and return the number
9604 read: this means that if none are read it will be an error. The
9605 DER versions of the certificate and CRL reader would always fail
9606 because it isn't possible to mix certificates and CRLs in DER format
9607 without choking one or the other routine. Changed this to just read
9608 a certificate: this is the best we can do. Also modified the code
9609 in apps/verify.c to take notice of return codes: it was previously
9610 attempting to read in certificates from NULL pointers and ignoring
9611 any errors: this is one reason why the cert and CRL reader seemed
9612 to work. It doesn't check return codes from the default certificate
9613 routines: these may well fail if the certificates aren't installed.
9616 *) Code to support otherName option in GeneralName.
9619 *) First update to verify code. Change the verify utility
9620 so it warns if it is passed a self signed certificate:
9621 for consistency with the normal behaviour. X509_verify
9622 has been modified to it will now verify a self signed
9623 certificate if *exactly* the same certificate appears
9624 in the store: it was previously impossible to trust a
9625 single self signed certificate. This means that:
9626 openssl verify ss.pem
9627 now gives a warning about a self signed certificate but
9628 openssl verify -CAfile ss.pem ss.pem
9632 *) For servers, store verify_result in SSL_SESSION data structure
9633 (and add it to external session representation).
9634 This is needed when client certificate verifications fails,
9635 but an application-provided verification callback (set by
9636 SSL_CTX_set_cert_verify_callback) allows accepting the session
9637 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9638 but returns 1): When the session is reused, we have to set
9639 ssl->verify_result to the appropriate error code to avoid
9641 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9643 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9644 case in PKCS7_dataInit() where the signed PKCS7 structure
9645 didn't contain any existing data because it was being created.
9646 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9648 *) Add a salt to the key derivation routines in enc.c. This
9649 forms the first 8 bytes of the encrypted file. Also add a
9650 -S option to allow a salt to be input on the command line.
9653 *) New function X509_cmp(). Oddly enough there wasn't a function
9654 to compare two certificates. We do this by working out the SHA1
9655 hash and comparing that. X509_cmp() will be needed by the trust
9659 *) SSL_get1_session() is like SSL_get_session(), but increments
9660 the reference count in the SSL_SESSION returned.
9661 [Geoff Thorpe <geoff@eu.c2.net>]
9663 *) Fix for 'req': it was adding a null to request attributes.
9664 Also change the X509_LOOKUP and X509_INFO code to handle
9665 certificate auxiliary information.
9668 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9672 *) Add the possibility to add extra information to the memory leak
9673 detecting output, to form tracebacks, showing from where each
9674 allocation was originated: CRYPTO_push_info("constant string") adds
9675 the string plus current file name and line number to a per-thread
9676 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9677 is like calling CYRPTO_pop_info() until the stack is empty.
9678 Also updated memory leak detection code to be multi-thread-safe.
9681 *) Add options -text and -noout to pkcs7 utility and delete the
9682 encryption options which never did anything. Update docs.
9685 *) Add options to some of the utilities to allow the pass phrase
9686 to be included on either the command line (not recommended on
9687 OSes like Unix) or read from the environment. Update the
9688 manpages and fix a few bugs.
9691 *) Add a few manpages for some of the openssl commands.
9694 *) Fix the -revoke option in ca. It was freeing up memory twice,
9695 leaking and not finding already revoked certificates.
9698 *) Extensive changes to support certificate auxiliary information.
9699 This involves the use of X509_CERT_AUX structure and X509_AUX
9700 functions. An X509_AUX function such as PEM_read_X509_AUX()
9701 can still read in a certificate file in the usual way but it
9702 will also read in any additional "auxiliary information". By
9703 doing things this way a fair degree of compatibility can be
9704 retained: existing certificates can have this information added
9705 using the new 'x509' options.
9707 Current auxiliary information includes an "alias" and some trust
9708 settings. The trust settings will ultimately be used in enhanced
9709 certificate chain verification routines: currently a certificate
9710 can only be trusted if it is self signed and then it is trusted
9714 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9715 The problem was that one of the replacement routines had not been working
9716 since SSLeay releases. For now the offending routine has been replaced
9717 with non-optimised assembler. Even so, this now gives around 95%
9718 performance improvement for 1024 bit RSA signs.
9721 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9722 handling. Most clients have the effective key size in bits equal to
9723 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9724 A few however don't do this and instead use the size of the decrypted key
9725 to determine the RC2 key length and the AlgorithmIdentifier to determine
9726 the effective key length. In this case the effective key length can still
9727 be 40 bits but the key length can be 168 bits for example. This is fixed
9728 by manually forcing an RC2 key into the EVP_PKEY structure because the
9729 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9730 the key length and effective key length are equal.
9733 *) Add a bunch of functions that should simplify the creation of
9734 X509_NAME structures. Now you should be able to do:
9735 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9736 and have it automatically work out the correct field type and fill in
9737 the structures. The more adventurous can try:
9738 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9739 and it will (hopefully) work out the correct multibyte encoding.
9742 *) Change the 'req' utility to use the new field handling and multibyte
9743 copy routines. Before the DN field creation was handled in an ad hoc
9744 way in req, ca, and x509 which was rather broken and didn't support
9745 BMPStrings or UTF8Strings. Since some software doesn't implement
9746 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9747 using the dirstring_type option. See the new comment in the default
9748 openssl.cnf for more info.
9751 *) Make crypto/rand/md_rand.c more robust:
9752 - Assure unique random numbers after fork().
9753 - Make sure that concurrent threads access the global counter and
9754 md serializably so that we never lose entropy in them
9755 or use exactly the same state in multiple threads.
9756 Access to the large state is not always serializable because
9757 the additional locking could be a performance killer, and
9758 md should be large enough anyway.
9761 *) New file apps/app_rand.c with commonly needed functionality
9762 for handling the random seed file.
9764 Use the random seed file in some applications that previously did not:
9766 dsaparam -genkey (which also ignored its '-rand' option),
9769 x509 (when signing).
9770 Except on systems with /dev/urandom, it is crucial to have a random
9771 seed file at least for key creation, DSA signing, and for DH exchanges;
9772 for RSA signatures we could do without one.
9774 gendh and gendsa (unlike genrsa) used to read only the first byte
9775 of each file listed in the '-rand' option. The function as previously
9776 found in genrsa is now in app_rand.c and is used by all programs
9777 that support '-rand'.
9780 *) In RAND_write_file, use mode 0600 for creating files;
9781 don't just chmod when it may be too late.
9784 *) Report an error from X509_STORE_load_locations
9785 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9788 *) New function ASN1_mbstring_copy() this copies a string in either
9789 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9790 into an ASN1_STRING type. A mask of permissible types is passed
9791 and it chooses the "minimal" type to use or an error if not type
9795 *) Add function equivalents to the various macros in asn1.h. The old
9796 macros are retained with an M_ prefix. Code inside the library can
9797 use the M_ macros. External code (including the openssl utility)
9798 should *NOT* in order to be "shared library friendly".
9801 *) Add various functions that can check a certificate's extensions
9802 to see if it usable for various purposes such as SSL client,
9803 server or S/MIME and CAs of these types. This is currently
9804 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9805 verification. Also added a -purpose flag to x509 utility to
9806 print out all the purposes.
9809 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9813 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9814 for, obtain and decode and extension and obtain its critical flag.
9815 This allows all the necessary extension code to be handled in a
9816 single function call.
9819 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9820 platforms. See crypto/rc4/rc4_enc.c for further details.
9823 *) New -noout option to asn1parse. This causes no output to be produced
9824 its main use is when combined with -strparse and -out to extract data
9825 from a file (which may not be in ASN.1 format).
9828 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9829 when producing the local key id.
9830 [Richard Levitte <levitte@stacken.kth.se>]
9832 *) New option -dhparam in s_server. This allows a DH parameter file to be
9833 stated explicitly. If it is not stated then it tries the first server
9834 certificate file. The previous behaviour hard coded the filename
9838 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9839 a public key to be input or output. For example:
9840 openssl rsa -in key.pem -pubout -out pubkey.pem
9841 Also added necessary DSA public key functions to handle this.
9844 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9845 in the message. This was handled by allowing
9846 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9847 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9849 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9850 to the end of the strings whereas this didn't. This would cause problems
9851 if strings read with d2i_ASN1_bytes() were later modified.
9852 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9854 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9855 data and it contains EOF it will end up returning an error. This is
9856 caused by input 46 bytes long. The cause is due to the way base64
9857 BIOs find the start of base64 encoded data. They do this by trying a
9858 trial decode on each line until they find one that works. When they
9859 do a flag is set and it starts again knowing it can pass all the
9860 data directly through the decoder. Unfortunately it doesn't reset
9861 the context it uses. This means that if EOF is reached an attempt
9862 is made to pass two EOFs through the context and this causes the
9863 resulting error. This can also cause other problems as well. As is
9864 usual with these problems it takes *ages* to find and the fix is
9865 trivial: move one line.
9866 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9868 *) Ugly workaround to get s_client and s_server working under Windows. The
9869 old code wouldn't work because it needed to select() on sockets and the
9870 tty (for keypresses and to see if data could be written). Win32 only
9871 supports select() on sockets so we select() with a 1s timeout on the
9872 sockets and then see if any characters are waiting to be read, if none
9873 are present then we retry, we also assume we can always write data to
9874 the tty. This isn't nice because the code then blocks until we've
9875 received a complete line of data and it is effectively polling the
9876 keyboard at 1s intervals: however it's quite a bit better than not
9877 working at all :-) A dedicated Windows application might handle this
9878 with an event loop for example.
9881 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9882 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9883 will be called when RSA_sign() and RSA_verify() are used. This is useful
9884 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9885 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9886 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9887 This necessitated the support of an extra signature type NID_md5_sha1
9888 for SSL signatures and modifications to the SSL library to use it instead
9889 of calling RSA_public_decrypt() and RSA_private_encrypt().
9892 *) Add new -verify -CAfile and -CApath options to the crl program, these
9893 will lookup a CRL issuers certificate and verify the signature in a
9894 similar way to the verify program. Tidy up the crl program so it
9895 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9896 less strict. It will now permit CRL extensions even if it is not
9897 a V2 CRL: this will allow it to tolerate some broken CRLs.
9900 *) Initialize all non-automatic variables each time one of the openssl
9901 sub-programs is started (this is necessary as they may be started
9902 multiple times from the "OpenSSL>" prompt).
9903 [Lennart Bang, Bodo Moeller]
9905 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9906 removing all other RSA functionality (this is what NO_RSA does). This
9907 is so (for example) those in the US can disable those operations covered
9908 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9912 *) Non-copying interface to BIO pairs.
9913 (still largely untested)
9916 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9917 ASCII string. This was handled independently in various places before.
9920 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9921 UTF8 strings a character at a time.
9924 *) Use client_version from client hello to select the protocol
9925 (s23_srvr.c) and for RSA client key exchange verification
9926 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9929 *) Add various utility functions to handle SPKACs, these were previously
9930 handled by poking round in the structure internals. Added new function
9931 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9932 print, verify and generate SPKACs. Based on an original idea from
9933 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9936 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9939 *) Allow the config file extension section to be overwritten on the
9940 command line. Based on an original idea from Massimiliano Pala
9941 <madwolf@comune.modena.it>. The new option is called -extensions
9942 and can be applied to ca, req and x509. Also -reqexts to override
9943 the request extensions in req and -crlexts to override the crl extensions
9947 *) Add new feature to the SPKAC handling in ca. Now you can include
9948 the same field multiple times by preceding it by "XXXX." for example:
9951 this is the same syntax as used in the req config file.
9954 *) Allow certificate extensions to be added to certificate requests. These
9955 are specified in a 'req_extensions' option of the req section of the
9956 config file. They can be printed out with the -text option to req but
9957 are otherwise ignored at present.
9960 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9961 data read consists of only the final block it would not decrypted because
9962 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9963 A misplaced 'break' also meant the decrypted final block might not be
9964 copied until the next read.
9967 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9968 a few extra parameters to the DH structure: these will be useful if
9969 for example we want the value of 'q' or implement X9.42 DH.
9972 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9973 provides hooks that allow the default DSA functions or functions on a
9974 "per key" basis to be replaced. This allows hardware acceleration and
9975 hardware key storage to be handled without major modification to the
9976 library. Also added low level modexp hooks and CRYPTO_EX structure and
9977 associated functions.
9980 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9981 as "read only": it can't be written to and the buffer it points to will
9982 not be freed. Reading from a read only BIO is much more efficient than
9983 a normal memory BIO. This was added because there are several times when
9984 an area of memory needs to be read from a BIO. The previous method was
9985 to create a memory BIO and write the data to it, this results in two
9986 copies of the data and an O(n^2) reading algorithm. There is a new
9987 function BIO_new_mem_buf() which creates a read only memory BIO from
9988 an area of memory. Also modified the PKCS#7 routines to use read only
9992 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9993 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9994 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9995 but a retry condition occured while trying to read the rest.
9998 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9999 NID_pkcs7_encrypted by default: this was wrong since this should almost
10000 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
10001 the encrypted data type: this is a more sensible place to put it and it
10002 allows the PKCS#12 code to be tidied up that duplicated this
10006 *) Changed obj_dat.pl script so it takes its input and output files on
10007 the command line. This should avoid shell escape redirection problems
10011 *) Initial support for certificate extension requests, these are included
10012 in things like Xenroll certificate requests. Included functions to allow
10013 extensions to be obtained and added.
10016 *) -crlf option to s_client and s_server for sending newlines as
10017 CRLF (as required by many protocols).
10020 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
10022 *) Install libRSAglue.a when OpenSSL is built with RSAref.
10023 [Ralf S. Engelschall]
10025 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
10026 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
10028 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
10032 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
10033 DH parameters/keys (q is lost during that conversion, but the resulting
10034 DH parameters contain its length).
10036 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
10037 much faster than DH_generate_parameters (which creates parameters
10038 where p = 2*q + 1), and also the smaller q makes DH computations
10039 much more efficient (160-bit exponentiation instead of 1024-bit
10040 exponentiation); so this provides a convenient way to support DHE
10041 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
10042 utter importance to use
10043 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10045 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10046 when such DH parameters are used, because otherwise small subgroup
10047 attacks may become possible!
10050 *) Avoid memory leak in i2d_DHparams.
10053 *) Allow the -k option to be used more than once in the enc program:
10054 this allows the same encrypted message to be read by multiple recipients.
10057 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
10058 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
10059 it will always use the numerical form of the OID, even if it has a short
10063 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
10064 method only got called if p,q,dmp1,dmq1,iqmp components were present,
10065 otherwise bn_mod_exp was called. In the case of hardware keys for example
10066 no private key components need be present and it might store extra data
10067 in the RSA structure, which cannot be accessed from bn_mod_exp.
10068 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
10069 private key operations.
10072 *) Added support for SPARC Linux.
10075 *) pem_password_cb function type incompatibly changed from
10076 typedef int pem_password_cb(char *buf, int size, int rwflag);
10078 ....(char *buf, int size, int rwflag, void *userdata);
10079 so that applications can pass data to their callbacks:
10080 The PEM[_ASN1]_{read,write}... functions and macros now take an
10081 additional void * argument, which is just handed through whenever
10082 the password callback is called.
10083 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10085 New function SSL_CTX_set_default_passwd_cb_userdata.
10087 Compatibility note: As many C implementations push function arguments
10088 onto the stack in reverse order, the new library version is likely to
10089 interoperate with programs that have been compiled with the old
10090 pem_password_cb definition (PEM_whatever takes some data that
10091 happens to be on the stack as its last argument, and the callback
10092 just ignores this garbage); but there is no guarantee whatsoever that
10095 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10096 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10097 problems not only on Windows, but also on some Unix platforms.
10098 To avoid problematic command lines, these definitions are now in an
10099 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10100 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10103 *) MIPS III/IV assembler module is reimplemented.
10106 *) More DES library cleanups: remove references to srand/rand and
10107 delete an unused file.
10110 *) Add support for the the free Netwide assembler (NASM) under Win32,
10111 since not many people have MASM (ml) and it can be hard to obtain.
10112 This is currently experimental but it seems to work OK and pass all
10113 the tests. Check out INSTALL.W32 for info.
10116 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10117 without temporary keys kept an extra copy of the server key,
10118 and connections with temporary keys did not free everything in case
10122 *) New function RSA_check_key and new openssl rsa option -check
10123 for verifying the consistency of RSA keys.
10124 [Ulf Moeller, Bodo Moeller]
10126 *) Various changes to make Win32 compile work:
10127 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10128 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10129 comparison" warnings.
10130 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10133 *) Add a debugging option to PKCS#5 v2 key generation function: when
10134 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10135 derived keys are printed to stderr.
10138 *) Copy the flags in ASN1_STRING_dup().
10139 [Roman E. Pavlov <pre@mo.msk.ru>]
10141 *) The x509 application mishandled signing requests containing DSA
10142 keys when the signing key was also DSA and the parameters didn't match.
10144 It was supposed to omit the parameters when they matched the signing key:
10145 the verifying software was then supposed to automatically use the CA's
10146 parameters if they were absent from the end user certificate.
10148 Omitting parameters is no longer recommended. The test was also
10149 the wrong way round! This was probably due to unusual behaviour in
10150 EVP_cmp_parameters() which returns 1 if the parameters match.
10151 This meant that parameters were omitted when they *didn't* match and
10152 the certificate was useless. Certificates signed with 'ca' didn't have
10154 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10156 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10157 The interface is as follows:
10158 Applications can use
10159 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10160 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10161 "off" is now the default.
10162 The library internally uses
10163 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10164 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10165 to disable memory-checking temporarily.
10167 Some inconsistent states that previously were possible (and were
10168 even the default) are now avoided.
10170 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10171 with each memory chunk allocated; this is occasionally more helpful
10172 than just having a counter.
10174 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10176 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10180 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10181 which largely parallels "options", but is for changing API behaviour,
10182 whereas "options" are about protocol behaviour.
10183 Initial "mode" flags are:
10185 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10186 a single record has been written.
10187 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10188 retries use the same buffer location.
10189 (But all of the contents must be
10193 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10196 *) Fix problems with no-hmac etc.
10197 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10199 *) New functions RSA_get_default_method(), RSA_set_method() and
10200 RSA_get_method(). These allows replacement of RSA_METHODs without having
10201 to mess around with the internals of an RSA structure.
10204 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10205 Also really enable memory leak checks in openssl.c and in some
10207 [Chad C. Mulligan, Bodo Moeller]
10209 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10210 up the length of negative integers. This has now been simplified to just
10211 store the length when it is first determined and use it later, rather
10212 than trying to keep track of where data is copied and updating it to
10214 [Steve Henson, reported by Brien Wheeler
10215 <bwheeler@authentica-security.com>]
10217 *) Add a new function PKCS7_signatureVerify. This allows the verification
10218 of a PKCS#7 signature but with the signing certificate passed to the
10219 function itself. This contrasts with PKCS7_dataVerify which assumes the
10220 certificate is present in the PKCS#7 structure. This isn't always the
10221 case: certificates can be omitted from a PKCS#7 structure and be
10222 distributed by "out of band" means (such as a certificate database).
10225 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10226 function prototypes in pem.h, also change util/mkdef.pl to add the
10227 necessary function names.
10230 *) mk1mf.pl (used by Windows builds) did not properly read the
10231 options set by Configure in the top level Makefile, and Configure
10232 was not even able to write more than one option correctly.
10233 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10236 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10237 file to be loaded from a BIO or FILE pointer. The BIO version will
10238 for example allow memory BIOs to contain config info.
10241 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10242 Whoever hopes to achieve shared-library compatibility across versions
10243 must use this, not the compile-time macro.
10244 (Exercise 0.9.4: Which is the minimum library version required by
10246 Note: All this applies only to multi-threaded programs, others don't
10250 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10251 through a BIO pair triggered the default case, i.e.
10252 SSLerr(...,SSL_R_UNKNOWN_STATE).
10255 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10256 can use the SSL library even if none of the specific BIOs is
10260 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10261 for the encoded length.
10262 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10264 *) Add initial documentation of the X509V3 functions.
10267 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10268 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10269 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10270 secure PKCS#8 private key format with a high iteration count.
10273 *) Fix determination of Perl interpreter: A perl or perl5
10274 _directory_ in $PATH was also accepted as the interpreter.
10275 [Ralf S. Engelschall]
10277 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10278 wrong with it but it was very old and did things like calling
10279 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10280 unusual formatting.
10283 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10284 to use the new extension code.
10287 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10288 with macros. This should make it easier to change their form, add extra
10289 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10293 *) Add to configuration table a new entry that can specify an alternative
10294 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10295 according to Mark Crispin <MRC@Panda.COM>.
10299 *) DES CBC did not update the IV. Weird.
10302 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10303 Changing the behaviour of the former might break existing programs --
10304 where IV updating is needed, des_ncbc_encrypt can be used.
10307 *) When bntest is run from "make test" it drives bc to check its
10308 calculations, as well as internally checking them. If an internal check
10309 fails, it needs to cause bc to give a non-zero result or make test carries
10310 on without noticing the failure. Fixed.
10313 *) DES library cleanups.
10316 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10317 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10318 ciphers. NOTE: although the key derivation function has been verified
10319 against some published test vectors it has not been extensively tested
10320 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10324 *) Instead of "mkdir -p", which is not fully portable, use new
10325 Perl script "util/mkdir-p.pl".
10328 *) Rewrite the way password based encryption (PBE) is handled. It used to
10329 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10330 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10331 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10332 the 'parameter' field of the AlgorithmIdentifier is passed to the
10333 underlying key generation function so it must do its own ASN1 parsing.
10334 This has also changed the EVP_PBE_CipherInit() function which now has a
10335 'parameter' argument instead of literal salt and iteration count values
10336 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10339 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10340 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10341 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10342 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10343 value was just used as a "magic string" and not used directly its
10344 value doesn't matter.
10347 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10351 *) "linux-sparc64" configuration (ultrapenguin).
10352 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10353 "linux-sparc" configuration.
10354 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10356 *) config now generates no-xxx options for missing ciphers.
10359 *) Support the EBCDIC character set (work in progress).
10360 File ebcdic.c not yet included because it has a different license.
10361 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10363 *) Support BS2000/OSD-POSIX.
10364 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10366 *) Make callbacks for key generation use void * instead of char *.
10369 *) Make S/MIME samples compile (not yet tested).
10372 *) Additional typesafe stacks.
10375 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10379 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10381 *) New configuration variant "sco5-gcc".
10383 *) Updated some demos.
10384 [Sean O Riordain, Wade Scholine]
10386 *) Add missing BIO_free at exit of pkcs12 application.
10389 *) Fix memory leak in conf.c.
10392 *) Updates for Win32 to assembler version of MD5.
10395 *) Set #! path to perl in apps/der_chop to where we found it
10396 instead of using a fixed path.
10399 *) SHA library changes for irix64-mips4-cc.
10402 *) Improvements for VMS support.
10406 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10408 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10409 This also avoids the problems with SC4.2 and unpatched SC5.
10410 [Andy Polyakov <appro@fy.chalmers.se>]
10412 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10413 These are required because of the typesafe stack would otherwise break
10414 existing code. If old code used a structure member which used to be STACK
10415 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10416 sk_num or sk_value it would produce an error because the num, data members
10417 are not present in STACK_OF. Now it just produces a warning. sk_set
10418 replaces the old method of assigning a value to sk_value
10419 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10420 that does this will no longer work (and should use sk_set instead) but
10421 this could be regarded as a "questionable" behaviour anyway.
10424 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10425 correctly handle encrypted S/MIME data.
10428 *) Change type of various DES function arguments from des_cblock
10429 (which means, in function argument declarations, pointer to char)
10430 to des_cblock * (meaning pointer to array with 8 char elements),
10431 which allows the compiler to do more typechecking; it was like
10432 that back in SSLeay, but with lots of ugly casts.
10434 Introduce new type const_des_cblock.
10437 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10438 problems: find RecipientInfo structure that matches recipient certificate
10439 and initialise the ASN1 structures properly based on passed cipher.
10442 *) Belatedly make the BN tests actually check the results.
10445 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10446 to and from BNs: it was completely broken. New compilation option
10447 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10448 key elements as negative integers.
10451 *) Reorganize and speed up MD5.
10452 [Andy Polyakov <appro@fy.chalmers.se>]
10455 [Richard Levitte <richard@levitte.org>]
10457 *) New option -out to asn1parse to allow the parsed structure to be
10458 output to a file. This is most useful when combined with the -strparse
10459 option to examine the output of things like OCTET STRINGS.
10462 *) Make SSL library a little more fool-proof by not requiring any longer
10463 that SSL_set_{accept,connect}_state be called before
10464 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10465 in many applications because usually everything *appeared* to work as
10466 intended anyway -- now it really works as intended).
10469 *) Move openssl.cnf out of lib/.
10472 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10473 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10474 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10475 [Ralf S. Engelschall]
10477 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10478 handle PKCS#7 enveloped data properly.
10479 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10481 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10482 copying pointers. The cert_st handling is changed by this in
10483 various ways (and thus what used to be known as ctx->default_cert
10484 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10485 any longer when s->cert does not give us what we need).
10486 ssl_cert_instantiate becomes obsolete by this change.
10487 As soon as we've got the new code right (possibly it already is?),
10488 we have solved a couple of bugs of the earlier code where s->cert
10489 was used as if it could not have been shared with other SSL structures.
10491 Note that using the SSL API in certain dirty ways now will result
10492 in different behaviour than observed with earlier library versions:
10493 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10494 does not influence s as it used to.
10496 In order to clean up things more thoroughly, inside SSL_SESSION
10497 we don't use CERT any longer, but a new structure SESS_CERT
10498 that holds per-session data (if available); currently, this is
10499 the peer's certificate chain and, for clients, the server's certificate
10500 and temporary key. CERT holds only those values that can have
10501 meaningful defaults in an SSL_CTX.
10504 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10505 from the internal representation. Various PKCS#7 fixes: remove some
10506 evil casts and set the enc_dig_alg field properly based on the signing
10510 *) Allow PKCS#12 password to be set from the command line or the
10511 environment. Let 'ca' get its config file name from the environment
10512 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10516 *) Allow certificate policies extension to use an IA5STRING for the
10517 organization field. This is contrary to the PKIX definition but
10518 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10522 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10523 without disallowing inline assembler and the like for non-pedantic builds.
10526 *) Support Borland C++ builder.
10527 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10529 *) Support Mingw32.
10532 *) SHA-1 cleanups and performance enhancements.
10533 [Andy Polyakov <appro@fy.chalmers.se>]
10535 *) Sparc v8plus assembler for the bignum library.
10536 [Andy Polyakov <appro@fy.chalmers.se>]
10538 *) Accept any -xxx and +xxx compiler options in Configure.
10541 *) Update HPUX configuration.
10544 *) Add missing sk_<type>_unshift() function to safestack.h
10545 [Ralf S. Engelschall]
10547 *) New function SSL_CTX_use_certificate_chain_file that sets the
10548 "extra_cert"s in addition to the certificate. (This makes sense
10549 only for "PEM" format files, as chains as a whole are not
10553 *) Support verify_depth from the SSL API.
10554 x509_vfy.c had what can be considered an off-by-one-error:
10555 Its depth (which was not part of the external interface)
10556 was actually counting the number of certificates in a chain;
10557 now it really counts the depth.
10560 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10561 instead of X509err, which often resulted in confusing error
10562 messages since the error codes are not globally unique
10563 (e.g. an alleged error in ssl3_accept when a certificate
10564 didn't match the private key).
10566 *) New function SSL_CTX_set_session_id_context that allows to set a default
10567 value (so that you don't need SSL_set_session_id_context for each
10568 connection using the SSL_CTX).
10571 *) OAEP decoding bug fix.
10574 *) Support INSTALL_PREFIX for package builders, as proposed by
10578 *) New Configure options "threads" and "no-threads". For systems
10579 where the proper compiler options are known (currently Solaris
10580 and Linux), "threads" is the default.
10583 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10586 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10587 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10588 such as /usr/local/bin.
10591 *) "make linux-shared" to build shared libraries.
10592 [Niels Poppe <niels@netbox.org>]
10594 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10597 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10598 extension adding in x509 utility.
10601 *) Remove NOPROTO sections and error code comments.
10604 *) Partial rewrite of the DEF file generator to now parse the ANSI
10608 *) New Configure options --prefix=DIR and --openssldir=DIR.
10611 *) Complete rewrite of the error code script(s). It is all now handled
10612 by one script at the top level which handles error code gathering,
10613 header rewriting and C source file generation. It should be much better
10614 than the old method: it now uses a modified version of Ulf's parser to
10615 read the ANSI prototypes in all header files (thus the old K&R definitions
10616 aren't needed for error creation any more) and do a better job of
10617 translating function codes into names. The old 'ASN1 error code imbedded
10618 in a comment' is no longer necessary and it doesn't use .err files which
10619 have now been deleted. Also the error code call doesn't have to appear all
10620 on one line (which resulted in some large lines...).
10623 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10626 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10627 0 (which usually indicates a closed connection), but continue reading.
10630 *) Fix some race conditions.
10633 *) Add support for CRL distribution points extension. Add Certificate
10634 Policies and CRL distribution points documentation.
10637 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10640 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10641 8 of keying material. Merlin has also confirmed interop with this fix
10642 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10643 [Merlin Hughes <merlin@baltimore.ie>]
10645 *) Fix lots of warnings.
10646 [Richard Levitte <levitte@stacken.kth.se>]
10648 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10649 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10650 [Richard Levitte <levitte@stacken.kth.se>]
10652 *) Fix problems with sizeof(long) == 8.
10653 [Andy Polyakov <appro@fy.chalmers.se>]
10655 *) Change functions to ANSI C.
10658 *) Fix typos in error codes.
10659 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10661 *) Remove defunct assembler files from Configure.
10664 *) SPARC v8 assembler BIGNUM implementation.
10665 [Andy Polyakov <appro@fy.chalmers.se>]
10667 *) Support for Certificate Policies extension: both print and set.
10668 Various additions to support the r2i method this uses.
10671 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10672 return a const string when you are expecting an allocated buffer.
10675 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10676 types DirectoryString and DisplayText.
10679 *) Add code to allow r2i extensions to access the configuration database,
10680 add an LHASH database driver and add several ctx helper functions.
10683 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10684 fail when they extended the size of a BIGNUM.
10687 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10688 support typesafe stack.
10691 *) Fix typo in SSL_[gs]et_options().
10692 [Nils Frostberg <nils@medcom.se>]
10694 *) Delete various functions and files that belonged to the (now obsolete)
10695 old X509V3 handling code.
10698 *) New Configure option "rsaref".
10701 *) Don't auto-generate pem.h.
10704 *) Introduce type-safe ASN.1 SETs.
10707 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10708 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10710 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10711 that links with OpenSSL (well at least cause lots of warnings), but fear
10712 not: the conversion is trivial, and it eliminates loads of evil casts. A
10713 few STACKed things have been converted already. Feel free to convert more.
10714 In the fullness of time, I'll do away with the STACK type altogether.
10717 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10718 specified in <certfile> by updating the entry in the index.txt file.
10719 This way one no longer has to edit the index.txt file manually for
10720 revoking a certificate. The -revoke option does the gory details now.
10721 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10723 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10724 `-text' option at all and this way the `-noout -text' combination was
10725 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10726 [Ralf S. Engelschall]
10728 *) Make sure a corresponding plain text error message exists for the
10729 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10730 verify callback function determined that a certificate was revoked.
10731 [Ralf S. Engelschall]
10733 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10734 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10735 all available cipers including rc5, which was forgotten until now.
10736 In order to let the testing shell script know which algorithms
10737 are available, a new (up to now undocumented) command
10738 "openssl list-cipher-commands" is used.
10741 *) Bugfix: s_client occasionally would sleep in select() when
10742 it should have checked SSL_pending() first.
10745 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10746 the raw DSA values prior to ASN.1 encoding.
10749 *) Tweaks to Configure
10750 [Niels Poppe <niels@netbox.org>]
10752 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10756 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10759 *) New config option to avoid instructions that are illegal on the 80386.
10760 The default code is faster, but requires at least a 486.
10763 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10764 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10765 same as SSL2_VERSION anyway.
10768 *) New "-showcerts" option for s_client.
10771 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10772 application. Various cleanups and fixes.
10775 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10776 modify error routines to work internally. Add error codes and PBE init
10777 to library startup routines.
10780 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10781 packing functions to asn1 and evp. Changed function names and error
10782 codes along the way.
10785 *) PKCS12 integration: and so it begins... First of several patches to
10786 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10787 objects to objects.h
10790 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10791 and display support for Thawte strong extranet extension.
10794 *) Add LinuxPPC support.
10795 [Jeff Dubrule <igor@pobox.org>]
10797 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10798 bn_div_words in alpha.s.
10799 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10801 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10802 OAEP isn't supported when OpenSSL is built with RSAref.
10803 [Ulf Moeller <ulf@fitug.de>]
10805 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10806 so they no longer are missing under -DNOPROTO.
10807 [Soren S. Jorvang <soren@t.dk>]
10810 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10812 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10813 doesn't work when the session is reused. Coming soon!
10816 *) Fix a security hole, that allows sessions to be reused in the wrong
10817 context thus bypassing client cert protection! All software that uses
10818 client certs and session caches in multiple contexts NEEDS PATCHING to
10819 allow session reuse! A fuller solution is in the works.
10820 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10822 *) Some more source tree cleanups (removed obsolete files
10823 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10824 permission on "config" script to be executable) and a fix for the INSTALL
10826 [Ulf Moeller <ulf@fitug.de>]
10828 *) Remove some legacy and erroneous uses of malloc, free instead of
10830 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10832 *) Make rsa_oaep_test return non-zero on error.
10833 [Ulf Moeller <ulf@fitug.de>]
10835 *) Add support for native Solaris shared libraries. Configure
10836 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10837 if someone would make that last step automatic.
10838 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10840 *) ctx_size was not built with the right compiler during "make links". Fixed.
10843 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10844 except NULL ciphers". This means the default cipher list will no longer
10845 enable NULL ciphers. They need to be specifically enabled e.g. with
10846 the string "DEFAULT:eNULL".
10849 *) Fix to RSA private encryption routines: if p < q then it would
10850 occasionally produce an invalid result. This will only happen with
10851 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10854 *) Be less restrictive and allow also `perl util/perlpath.pl
10855 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10856 because this way one can also use an interpreter named `perl5' (which is
10857 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10858 installed as `perl').
10859 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10861 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10862 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10864 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10865 advapi32.lib to Win32 build and change the pem test comparision
10866 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10867 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10868 and crypto/des/ede_cbcm_enc.c.
10871 *) DES quad checksum was broken on big-endian architectures. Fixed.
10874 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10875 Win32 test batch file so it (might) work again. The Win32 test batch file
10876 is horrible: I feel ill....
10879 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10880 in e_os.h. Audit of header files to check ANSI and non ANSI
10881 sections: 10 functions were absent from non ANSI section and not exported
10882 from Windows DLLs. Fixed up libeay.num for new functions.
10885 *) Make `openssl version' output lines consistent.
10886 [Ralf S. Engelschall]
10888 *) Fix Win32 symbol export lists for BIO functions: Added
10889 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10890 to ms/libeay{16,32}.def.
10891 [Ralf S. Engelschall]
10893 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10894 fine under Unix and passes some trivial tests I've now added. But the
10895 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10896 added to make sure no one expects that this stuff really works in the
10897 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10898 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10900 [Ralf S. Engelschall]
10902 *) Fix the generation of two part addresses in perl.
10903 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10905 *) Add config entry for Linux on MIPS.
10906 [John Tobey <jtobey@channel1.com>]
10908 *) Make links whenever Configure is run, unless we are on Windoze.
10911 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10912 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10916 *) Add a useful kludge to allow package maintainers to specify compiler and
10917 other platforms details on the command line without having to patch the
10918 Configure script everytime: One now can use ``perl Configure
10919 <id>:<details>'', i.e. platform ids are allowed to have details appended
10920 to them (seperated by colons). This is treated as there would be a static
10921 pre-configured entry in Configure's %table under key <id> with value
10922 <details> and ``perl Configure <id>'' is called. So, when you want to
10923 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10924 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10925 now, which overrides the FreeBSD-elf entry on-the-fly.
10926 [Ralf S. Engelschall]
10928 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10931 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10932 on the `perl Configure ...' command line. This way one can compile
10933 OpenSSL libraries with Position Independent Code (PIC) which is needed
10934 for linking it into DSOs.
10935 [Ralf S. Engelschall]
10937 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10941 *) Cleaned up the LICENSE document: The official contact for any license
10942 questions now is the OpenSSL core team under openssl-core@openssl.org.
10943 And add a paragraph about the dual-license situation to make sure people
10944 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10945 to the OpenSSL toolkit.
10946 [Ralf S. Engelschall]
10948 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10949 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10950 Additonally cleaned up the `make links' target: Remove unnecessary
10951 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10952 to speed processing and no longer clutter the display with confusing
10953 stuff. Instead only the actually done links are displayed.
10954 [Ralf S. Engelschall]
10956 *) Permit null encryption ciphersuites, used for authentication only. It used
10957 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10958 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10962 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10963 signed attributes when verifying signatures (this would break them),
10964 the detached data encoding was wrong and public keys obtained using
10965 X509_get_pubkey() weren't freed.
10968 *) Add text documentation for the BUFFER functions. Also added a work around
10969 to a Win95 console bug. This was triggered by the password read stuff: the
10970 last character typed gets carried over to the next fread(). If you were
10971 generating a new cert request using 'req' for example then the last
10972 character of the passphrase would be CR which would then enter the first
10976 *) Added the new `Includes OpenSSL Cryptography Software' button as
10977 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10978 button and can be used by applications based on OpenSSL to show the
10979 relationship to the OpenSSL project.
10980 [Ralf S. Engelschall]
10982 *) Remove confusing variables in function signatures in files
10983 ssl/ssl_lib.c and ssl/ssl.h.
10984 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10986 *) Don't install bss_file.c under PREFIX/include/
10987 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10989 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10990 functions that return function pointers and has support for NT specific
10991 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10992 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10993 unsigned to signed types: this was killing the Win32 compile.
10996 *) Add new certificate file to stack functions,
10997 SSL_add_dir_cert_subjects_to_stack() and
10998 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10999 SSL_load_client_CA_file(), and can be used to add multiple certs easily
11000 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
11001 This means that Apache-SSL and similar packages don't have to mess around
11002 to add as many CAs as they want to the preferred list.
11005 *) Experiment with doxygen documentation. Currently only partially applied to
11007 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
11008 openssl.doxy as the configuration file.
11011 *) Get rid of remaining C++-style comments which strict C compilers hate.
11012 [Ralf S. Engelschall, pointed out by Carlos Amengual]
11014 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
11015 compiled in by default: it has problems with large keys.
11018 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
11019 DH private keys and/or callback functions which directly correspond to
11020 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
11021 is needed for applications which have to configure certificates on a
11022 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
11024 For the RSA certificate situation is makes no difference, but
11025 for the DSA certificate situation this fixes the "no shared cipher"
11026 problem where the OpenSSL cipher selection procedure failed because the
11027 temporary keys were not overtaken from the context and the API provided
11028 no way to reconfigure them.
11029 The new functions now let applications reconfigure the stuff and they
11030 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
11031 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
11032 non-public-API function ssl_cert_instantiate() is used as a helper
11033 function and also to reduce code redundancy inside ssl_rsa.c.
11034 [Ralf S. Engelschall]
11036 *) Move s_server -dcert and -dkey options out of the undocumented feature
11037 area because they are useful for the DSA situation and should be
11038 recognized by the users.
11039 [Ralf S. Engelschall]
11041 *) Fix the cipher decision scheme for export ciphers: the export bits are
11042 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
11043 SSL_EXP_MASK. So, the original variable has to be used instead of the
11044 already masked variable.
11045 [Richard Levitte <levitte@stacken.kth.se>]
11047 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
11048 [Richard Levitte <levitte@stacken.kth.se>]
11050 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
11051 from `int' to `unsigned int' because it's a length and initialized by
11052 EVP_DigestFinal() which expects an `unsigned int *'.
11053 [Richard Levitte <levitte@stacken.kth.se>]
11055 *) Don't hard-code path to Perl interpreter on shebang line of Configure
11056 script. Instead use the usual Shell->Perl transition trick.
11057 [Ralf S. Engelschall]
11059 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
11060 (in addition to RSA certificates) to match the behaviour of `openssl dsa
11061 -noout -modulus' as it's already the case for `openssl rsa -noout
11062 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
11063 currently the public key is printed (a decision which was already done by
11064 `openssl dsa -modulus' in the past) which serves a similar purpose.
11065 Additionally the NO_RSA no longer completely removes the whole -modulus
11066 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
11068 [Ralf S. Engelschall]
11070 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
11071 BIO. See the source (crypto/evp/bio_ok.c) for more info.
11072 [Arne Ansper <arne@ats.cyber.ee>]
11074 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
11075 to be added. Now both 'req' and 'ca' can use new objects defined in the
11079 *) Add cool BIO that does syslog (or event log on NT).
11080 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
11082 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
11083 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11084 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11085 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11088 *) Add preliminary config info for new extension code.
11091 *) Make RSA_NO_PADDING really use no padding.
11092 [Ulf Moeller <ulf@fitug.de>]
11094 *) Generate errors when private/public key check is done.
11097 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11098 for some CRL extensions and new objects added.
11101 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11102 key usage extension and fuller support for authority key id.
11105 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11106 padding method for RSA, which is recommended for new applications in PKCS
11107 #1 v2.0 (RFC 2437, October 1998).
11108 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11109 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11110 against Bleichbacher's attack on RSA.
11111 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11114 *) Updates to the new SSL compression code
11115 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11117 *) Fix so that the version number in the master secret, when passed
11118 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11119 (because the server will not accept higher), that the version number
11120 is 0x03,0x01, not 0x03,0x00
11121 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11123 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11124 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11125 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11128 *) Support for RAW extensions where an arbitrary extension can be
11129 created by including its DER encoding. See apps/openssl.cnf for
11133 *) Make sure latest Perl versions don't interpret some generated C array
11134 code as Perl array code in the crypto/err/err_genc.pl script.
11135 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11137 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11138 not many people have the assembler. Various Win32 compilation fixes and
11139 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11140 build instructions.
11143 *) Modify configure script 'Configure' to automatically create crypto/date.h
11144 file under Win32 and also build pem.h from pem.org. New script
11145 util/mkfiles.pl to create the MINFO file on environments that can't do a
11146 'make files': perl util/mkfiles.pl >MINFO should work.
11149 *) Major rework of DES function declarations, in the pursuit of correctness
11150 and purity. As a result, many evil casts evaporated, and some weirdness,
11151 too. You may find this causes warnings in your code. Zapping your evil
11152 casts will probably fix them. Mostly.
11155 *) Fix for a typo in asn1.h. Bug fix to object creation script
11156 obj_dat.pl. It considered a zero in an object definition to mean
11157 "end of object": none of the objects in objects.h have any zeros
11158 so it wasn't spotted.
11159 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11161 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11162 Masking (CBCM). In the absence of test vectors, the best I have been able
11163 to do is check that the decrypt undoes the encrypt, so far. Send me test
11164 vectors if you have them.
11167 *) Correct calculation of key length for export ciphers (too much space was
11168 allocated for null ciphers). This has not been tested!
11171 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11172 message is now correct (it understands "crypto" and "ssl" on its
11173 command line). There is also now an "update" option. This will update
11174 the util/ssleay.num and util/libeay.num files with any new functions.
11176 perl util/mkdef.pl crypto ssl update
11177 it will update them.
11180 *) Overhauled the Perl interface (perl/*):
11181 - ported BN stuff to OpenSSL's different BN library
11182 - made the perl/ source tree CVS-aware
11183 - renamed the package from SSLeay to OpenSSL (the files still contain
11184 their history because I've copied them in the repository)
11185 - removed obsolete files (the test scripts will be replaced
11186 by better Test::Harness variants in the future)
11187 [Ralf S. Engelschall]
11189 *) First cut for a very conservative source tree cleanup:
11190 1. merge various obsolete readme texts into doc/ssleay.txt
11191 where we collect the old documents and readme texts.
11192 2. remove the first part of files where I'm already sure that we no
11193 longer need them because of three reasons: either they are just temporary
11194 files which were left by Eric or they are preserved original files where
11195 I've verified that the diff is also available in the CVS via "cvs diff
11196 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11197 the crypto/md/ stuff).
11198 [Ralf S. Engelschall]
11200 *) More extension code. Incomplete support for subject and issuer alt
11201 name, issuer and authority key id. Change the i2v function parameters
11202 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11203 what that's for :-) Fix to ASN1 macro which messed up
11204 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11207 *) Preliminary support for ENUMERATED type. This is largely copied from the
11211 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11212 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11214 *) Make sure `make rehash' target really finds the `openssl' program.
11215 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11217 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11218 like to hear about it if this slows down other processors.
11221 *) Add CygWin32 platform information to Configure script.
11222 [Alan Batie <batie@aahz.jf.intel.com>]
11224 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11225 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11227 *) New program nseq to manipulate netscape certificate sequences
11230 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11234 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11235 but the BN code had some problems that would cause failures when
11236 doing certificate verification and some other functions.
11237 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11239 *) Add ASN1 and PEM code to support netscape certificate sequences.
11242 *) Add ASN1 and PEM code to support netscape certificate sequences.
11245 *) Add several PKIX and private extended key usage OIDs.
11248 *) Modify the 'ca' program to handle the new extension code. Modify
11249 openssl.cnf for new extension format, add comments.
11252 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11253 and add a sample to openssl.cnf so req -x509 now adds appropriate
11257 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11258 error code, add initial support to X509_print() and x509 application.
11261 *) Takes a deep breath and start addding X509 V3 extension support code. Add
11262 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11263 stuff is currently isolated and isn't even compiled yet.
11266 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11267 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11268 Removed the versions check from X509 routines when loading extensions:
11269 this allows certain broken certificates that don't set the version
11270 properly to be processed.
11273 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11274 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11275 can still be regenerated with "make depend".
11278 *) Spelling mistake in C version of CAST-128.
11279 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11281 *) Changes to the error generation code. The perl script err-code.pl
11282 now reads in the old error codes and retains the old numbers, only
11283 adding new ones if necessary. It also only changes the .err files if new
11284 codes are added. The makefiles have been modified to only insert errors
11285 when needed (to avoid needlessly modifying header files). This is done
11286 by only inserting errors if the .err file is newer than the auto generated
11287 C file. To rebuild all the error codes from scratch (the old behaviour)
11288 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11289 or delete all the .err files.
11292 *) CAST-128 was incorrectly implemented for short keys. The C version has
11293 been fixed, but is untested. The assembler versions are also fixed, but
11294 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11295 to regenerate it if needed.
11296 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11297 Hagino <itojun@kame.net>]
11299 *) File was opened incorrectly in randfile.c.
11300 [Ulf Möller <ulf@fitug.de>]
11302 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11303 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11304 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11305 al: it's just almost always a UTCTime. Note this patch adds new error
11306 codes so do a "make errors" if there are problems.
11309 *) Correct Linux 1 recognition in config.
11310 [Ulf Möller <ulf@fitug.de>]
11312 *) Remove pointless MD5 hash when using DSA keys in ca.
11313 [Anonymous <nobody@replay.com>]
11315 *) Generate an error if given an empty string as a cert directory. Also
11316 generate an error if handed NULL (previously returned 0 to indicate an
11317 error, but didn't set one).
11318 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11320 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11323 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11324 parameters. This was causing a warning which killed off the Win32 compile.
11327 *) Remove C++ style comments from crypto/bn/bn_local.h.
11328 [Neil Costigan <neil.costigan@celocom.com>]
11330 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11331 based on a text string, looking up short and long names and finally
11332 "dot" format. The "dot" format stuff didn't work. Added new function
11333 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11334 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11335 OID is not part of the table.
11338 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11339 X509_LOOKUP_by_alias().
11342 *) Sort openssl functions by name.
11345 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11346 encryption from sample DSA keys (in case anyone is interested the password
11350 *) Make _all_ *_free functions accept a NULL pointer.
11351 [Frans Heymans <fheymans@isaserver.be>]
11353 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11355 [Anonymous <nobody@replay.com>]
11357 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11358 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11360 *) Don't blow it for numeric -newkey arguments to apps/req.
11361 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11363 *) Temp key "for export" tests were wrong in s3_srvr.c.
11364 [Anonymous <nobody@replay.com>]
11366 *) Add prototype for temp key callback functions
11367 SSL_CTX_set_tmp_{rsa,dh}_callback().
11370 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11371 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11374 *) X509_name_add_entry() freed the wrong thing after an error.
11375 [Arne Ansper <arne@ats.cyber.ee>]
11377 *) rsa_eay.c would attempt to free a NULL context.
11378 [Arne Ansper <arne@ats.cyber.ee>]
11380 *) BIO_s_socket() had a broken should_retry() on Windoze.
11381 [Arne Ansper <arne@ats.cyber.ee>]
11383 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11384 [Arne Ansper <arne@ats.cyber.ee>]
11386 *) Make sure the already existing X509_STORE->depth variable is initialized
11387 in X509_STORE_new(), but document the fact that this variable is still
11388 unused in the certificate verification process.
11389 [Ralf S. Engelschall]
11391 *) Fix the various library and apps files to free up pkeys obtained from
11392 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11395 *) Fix reference counting in X509_PUBKEY_get(). This makes
11396 demos/maurice/example2.c work, amongst others, probably.
11397 [Steve Henson and Ben Laurie]
11399 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11400 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11401 are no longer created. This way we have a single and consistent command
11402 line interface `openssl <command>', similar to `cvs <command>'.
11403 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11405 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11406 BIT STRING wrapper always have zero unused bits.
11409 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11412 *) Make the top-level INSTALL documentation easier to understand.
11415 *) Makefiles updated to exit if an error occurs in a sub-directory
11416 make (including if user presses ^C) [Paul Sutton]
11418 *) Make Montgomery context stuff explicit in RSA data structure.
11421 *) Fix build order of pem and err to allow for generated pem.h.
11424 *) Fix renumbering bug in X509_NAME_delete_entry().
11427 *) Enhanced the err-ins.pl script so it makes the error library number
11428 global and can add a library name. This is needed for external ASN1 and
11429 other error libraries.
11432 *) Fixed sk_insert which never worked properly.
11435 *) Fix ASN1 macros so they can handle indefinite length construted
11436 EXPLICIT tags. Some non standard certificates use these: they can now
11440 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11441 into a single doc/ssleay.txt bundle. This way the information is still
11442 preserved but no longer messes up this directory. Now it's new room for
11443 the new set of documenation files.
11444 [Ralf S. Engelschall]
11446 *) SETs were incorrectly DER encoded. This was a major pain, because they
11447 shared code with SEQUENCEs, which aren't coded the same. This means that
11448 almost everything to do with SETs or SEQUENCEs has either changed name or
11449 number of arguments.
11450 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11452 *) Fix test data to work with the above.
11455 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11456 was already fixed by Eric for 0.9.1 it seems.
11457 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11459 *) Autodetect FreeBSD3.
11462 *) Fix various bugs in Configure. This affects the following platforms:
11466 unixware-2.0-pentium
11470 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11471 before they are needed.
11474 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11478 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11480 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11481 changed SSLeay to OpenSSL in version strings.
11482 [Ralf S. Engelschall]
11484 *) Some fixups to the top-level documents.
11487 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11488 because the symlink to include/ was missing.
11489 [Ralf S. Engelschall]
11491 *) Incorporated the popular no-RSA/DSA-only patches
11492 which allow to compile a RSA-free SSLeay.
11493 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11495 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11496 when "ssleay" is still not found.
11497 [Ralf S. Engelschall]
11499 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11500 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11502 *) Updated the README file.
11503 [Ralf S. Engelschall]
11505 *) Added various .cvsignore files in the CVS repository subdirs
11506 to make a "cvs update" really silent.
11507 [Ralf S. Engelschall]
11509 *) Recompiled the error-definition header files and added
11510 missing symbols to the Win32 linker tables.
11511 [Ralf S. Engelschall]
11513 *) Cleaned up the top-level documents;
11514 o new files: CHANGES and LICENSE
11515 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11516 o merged COPYRIGHT into LICENSE
11517 o removed obsolete TODO file
11518 o renamed MICROSOFT to INSTALL.W32
11519 [Ralf S. Engelschall]
11521 *) Removed dummy files from the 0.9.1b source tree:
11522 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11523 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11524 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11525 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11526 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11527 [Ralf S. Engelschall]
11529 *) Added various platform portability fixes.
11532 *) The Genesis of the OpenSSL rpject:
11533 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11534 Young and Tim J. Hudson created while they were working for C2Net until
11536 [The OpenSSL Project]
11539 Changes between 0.9.0b and 0.9.1b [not released]
11541 *) Updated a few CA certificates under certs/
11544 *) Changed some BIGNUM api stuff.
11547 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11548 DGUX x86, Linux Alpha, etc.
11551 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11552 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11556 *) Add -strparse option to asn1pars program which parses nested
11558 [Dr Stephen Henson <shenson@bigfoot.com>]
11560 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11563 *) DSA fix for "ca" program.
11566 *) Added "-genkey" option to "dsaparam" program.
11569 *) Added RIPE MD160 (rmd160) message digest.
11572 *) Added -a (all) option to "ssleay version" command.
11575 *) Added PLATFORM define which is the id given to Configure.
11578 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11581 *) Extended the ASN.1 parser routines.
11584 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11587 *) Added a BN_CTX to the BN library.
11590 *) Fixed the weak key values in DES library
11593 *) Changed API in EVP library for cipher aliases.
11596 *) Added support for RC2/64bit cipher.
11599 *) Converted the lhash library to the crypto/mem.c functions.
11602 *) Added more recognized ASN.1 object ids.
11605 *) Added more RSA padding checks for SSL/TLS.
11608 *) Added BIO proxy/filter functionality.
11611 *) Added extra_certs to SSL_CTX which can be used
11612 send extra CA certificates to the client in the CA cert chain sending
11613 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11616 *) Now Fortezza is denied in the authentication phase because
11617 this is key exchange mechanism is not supported by SSLeay at all.
11620 *) Additional PKCS1 checks.
11623 *) Support the string "TLSv1" for all TLS v1 ciphers.
11626 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11627 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11630 *) Fixed a few memory leaks.
11633 *) Fixed various code and comment typos.
11636 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11637 bytes sent in the client random.
11638 [Edward Bishop <ebishop@spyglass.com>]