5 Changes between 1.0.1u and 1.0.1v [xx XXX xxxx]
9 Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
11 *) OCSP Status Request extension unbounded memory growth
13 A malicious client can send an excessively large OCSP Status Request
14 extension. If that client continually requests renegotiation, sending a
15 large OCSP Status Request extension each time, then there will be unbounded
16 memory growth on the server. This will eventually lead to a Denial Of
17 Service attack through memory exhaustion. Servers with a default
18 configuration are vulnerable even if they do not support OCSP. Builds using
19 the "no-ocsp" build time option are not affected.
21 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
25 *) In order to mitigate the SWEET32 attack, the DES ciphers were moved from
28 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
33 *) OOB write in MDC2_Update()
35 An overflow can occur in MDC2_Update() either if called directly or
36 through the EVP_DigestUpdate() function using MDC2. If an attacker
37 is able to supply very large amounts of input data after a previous
38 call to EVP_EncryptUpdate() with a partial block then a length check
39 can overflow resulting in a heap corruption.
41 The amount of data needed is comparable to SIZE_MAX which is impractical
44 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
48 *) Malformed SHA512 ticket DoS
50 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
51 DoS attack where a malformed ticket will result in an OOB read which will
54 The use of SHA512 in TLS session tickets is comparatively rare as it requires
55 a custom server callback and ticket lookup mechanism.
57 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
61 *) OOB write in BN_bn2dec()
63 The function BN_bn2dec() does not check the return value of BN_div_word().
64 This can cause an OOB write if an application uses this function with an
65 overly large BIGNUM. This could be a problem if an overly large certificate
66 or CRL is printed out from an untrusted source. TLS is not affected because
67 record limits will reject an oversized certificate before it is parsed.
69 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
73 *) OOB read in TS_OBJ_print_bio()
75 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
76 the total length the OID text representation would use and not the amount
77 of data written. This will result in OOB reads when large OIDs are
80 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
84 *) Pointer arithmetic undefined behaviour
86 Avoid some undefined pointer arithmetic
88 A common idiom in the codebase is to check limits in the following manner:
91 Where "p" points to some malloc'd data of SIZE bytes and
94 "len" here could be from some externally supplied data (e.g. from a TLS
97 The rules of C pointer arithmetic are such that "p + len" is only well
98 defined where len <= SIZE. Therefore the above idiom is actually
101 For example this could cause problems if some malloc implementation
102 provides an address for "p" such that "p + len" actually overflows for
103 values of len that are too big and therefore p + len < limit.
105 This issue was reported to OpenSSL by Guido Vranken
109 *) Constant time flag not preserved in DSA signing
111 Operations in the DSA signing algorithm should run in constant time in
112 order to avoid side channel attacks. A flaw in the OpenSSL DSA
113 implementation means that a non-constant time codepath is followed for
114 certain operations. This has been demonstrated through a cache-timing
115 attack to be sufficient for an attacker to recover the private DSA key.
117 This issue was reported by César Pereida (Aalto University), Billy Brumley
118 (Tampere University of Technology), and Yuval Yarom (The University of
123 *) DTLS buffered message DoS
125 In a DTLS connection where handshake messages are delivered out-of-order
126 those messages that OpenSSL is not yet ready to process will be buffered
127 for later use. Under certain circumstances, a flaw in the logic means that
128 those messages do not get removed from the buffer even though the handshake
129 has been completed. An attacker could force up to approx. 15 messages to
130 remain in the buffer when they are no longer required. These messages will
131 be cleared when the DTLS connection is closed. The default maximum size for
132 a message is 100k. Therefore the attacker could force an additional 1500k
133 to be consumed per connection. By opening many simulataneous connections an
134 attacker could cause a DoS attack through memory exhaustion.
136 This issue was reported to OpenSSL by Quan Luo.
140 *) DTLS replay protection DoS
142 A flaw in the DTLS replay attack protection mechanism means that records
143 that arrive for future epochs update the replay protection "window" before
144 the MAC for the record has been validated. This could be exploited by an
145 attacker by sending a record for the next epoch (which does not have to
146 decrypt or have a valid MAC), with a very large sequence number. This means
147 that all subsequent legitimate packets are dropped causing a denial of
148 service for a specific DTLS connection.
150 This issue was reported to OpenSSL by the OCAP audit team.
154 *) Certificate message OOB reads
156 In OpenSSL 1.0.2 and earlier some missing message length checks can result
157 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
158 theoretical DoS risk but this has not been observed in practice on common
161 The messages affected are client certificate, client certificate request
162 and server certificate. As a result the attack can only be performed
163 against a client or a server which enables client authentication.
165 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
169 Changes between 1.0.1s and 1.0.1t [3 May 2016]
171 *) Prevent padding oracle in AES-NI CBC MAC check
173 A MITM attacker can use a padding oracle attack to decrypt traffic
174 when the connection uses an AES CBC cipher and the server support
177 This issue was introduced as part of the fix for Lucky 13 padding
178 attack (CVE-2013-0169). The padding check was rewritten to be in
179 constant time by making sure that always the same bytes are read and
180 compared against either the MAC or padding bytes. But it no longer
181 checked that there was enough data to have both the MAC and padding
184 This issue was reported by Juraj Somorovsky using TLS-Attacker.
188 *) Fix EVP_EncodeUpdate overflow
190 An overflow can occur in the EVP_EncodeUpdate() function which is used for
191 Base64 encoding of binary data. If an attacker is able to supply very large
192 amounts of input data then a length check can overflow resulting in a heap
195 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
196 the PEM_write_bio* family of functions. These are mainly used within the
197 OpenSSL command line applications, so any application which processes data
198 from an untrusted source and outputs it as a PEM file should be considered
199 vulnerable to this issue. User applications that call these APIs directly
200 with large amounts of untrusted data may also be vulnerable.
202 This issue was reported by Guido Vranken.
206 *) Fix EVP_EncryptUpdate overflow
208 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
209 is able to supply very large amounts of input data after a previous call to
210 EVP_EncryptUpdate() with a partial block then a length check can overflow
211 resulting in a heap corruption. Following an analysis of all OpenSSL
212 internal usage of the EVP_EncryptUpdate() function all usage is one of two
213 forms. The first form is where the EVP_EncryptUpdate() call is known to be
214 the first called function after an EVP_EncryptInit(), and therefore that
215 specific call must be safe. The second form is where the length passed to
216 EVP_EncryptUpdate() can be seen from the code to be some small value and
217 therefore there is no possibility of an overflow. Since all instances are
218 one of these two forms, it is believed that there can be no overflows in
219 internal code due to this problem. It should be noted that
220 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
221 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
222 of these calls have also been analysed too and it is believed there are no
223 instances in internal usage where an overflow could occur.
225 This issue was reported by Guido Vranken.
229 *) Prevent ASN.1 BIO excessive memory allocation
231 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
232 a short invalid encoding can casuse allocation of large amounts of memory
233 potentially consuming excessive resources or exhausting memory.
235 Any application parsing untrusted data through d2i BIO functions is
236 affected. The memory based functions such as d2i_X509() are *not* affected.
237 Since the memory based functions are used by the TLS library, TLS
238 applications are not affected.
240 This issue was reported by Brian Carpenter.
246 ASN1 Strings that are over 1024 bytes can cause an overread in applications
247 using the X509_NAME_oneline() function on EBCDIC systems. This could result
248 in arbitrary stack data being returned in the buffer.
250 This issue was reported by Guido Vranken.
254 *) Modify behavior of ALPN to invoke callback after SNI/servername
255 callback, such that updates to the SSL_CTX affect ALPN.
258 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
262 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
263 methods are enabled and ssl2 is disabled the methods return NULL.
266 Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
268 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
269 Builds that are not configured with "enable-weak-ssl-ciphers" will not
270 provide any "EXPORT" or "LOW" strength ciphers.
273 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
274 is by default disabled at build-time. Builds that are not configured with
275 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
276 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
277 will need to explicitly call either of:
279 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
281 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
283 as appropriate. Even if either of those is used, or the application
284 explicitly uses the version-specific SSLv2_method() or its client and
285 server variants, SSLv2 ciphers vulnerable to exhaustive search key
286 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
287 ciphers, and SSLv2 56-bit DES are no longer available.
291 *) Fix a double-free in DSA code
293 A double free bug was discovered when OpenSSL parses malformed DSA private
294 keys and could lead to a DoS attack or memory corruption for applications
295 that receive DSA private keys from untrusted sources. This scenario is
298 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
303 *) Disable SRP fake user seed to address a server memory leak.
305 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
307 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
308 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
309 was changed to ignore the "fake user" SRP seed, even if the seed
312 Users should use SRP_VBASE_get1_by_user instead. Note that in
313 SRP_VBASE_get1_by_user, caller must free the returned value. Note
314 also that even though configuring the SRP seed attempts to hide
315 invalid usernames by continuing the handshake with fake
316 credentials, this behaviour is not constant time and no strong
317 guarantees are made that the handshake is indistinguishable from
318 that of a valid user.
322 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
324 In the BN_hex2bn function the number of hex digits is calculated using an
325 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
326 large values of |i| this can result in |bn_expand| not allocating any
327 memory because |i * 4| is negative. This can leave the internal BIGNUM data
328 field as NULL leading to a subsequent NULL ptr deref. For very large values
329 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
330 In this case memory is allocated to the internal BIGNUM data field, but it
331 is insufficiently sized leading to heap corruption. A similar issue exists
332 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
333 is ever called by user applications with very large untrusted hex/dec data.
334 This is anticipated to be a rare occurrence.
336 All OpenSSL internal usage of these functions use data that is not expected
337 to be untrusted, e.g. config file data or application command line
338 arguments. If user developed applications generate config file data based
339 on untrusted data then it is possible that this could also lead to security
340 consequences. This is also anticipated to be rare.
342 This issue was reported to OpenSSL by Guido Vranken.
346 *) Fix memory issues in BIO_*printf functions
348 The internal |fmtstr| function used in processing a "%s" format string in
349 the BIO_*printf functions could overflow while calculating the length of a
350 string and cause an OOB read when printing very long strings.
352 Additionally the internal |doapr_outch| function can attempt to write to an
353 OOB memory location (at an offset from the NULL pointer) in the event of a
354 memory allocation failure. In 1.0.2 and below this could be caused where
355 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
356 could be in processing a very long "%s" format string. Memory leaks can
359 The first issue may mask the second issue dependent on compiler behaviour.
360 These problems could enable attacks where large amounts of untrusted data
361 is passed to the BIO_*printf functions. If applications use these functions
362 in this way then they could be vulnerable. OpenSSL itself uses these
363 functions when printing out human-readable dumps of ASN.1 data. Therefore
364 applications that print this data could be vulnerable if the data is from
365 untrusted sources. OpenSSL command line applications could also be
366 vulnerable where they print out ASN.1 data, or if untrusted data is passed
367 as command line arguments.
369 Libssl is not considered directly vulnerable. Additionally certificates etc
370 received via remote connections via libssl are also unlikely to be able to
371 trigger these issues because of message size limits enforced within libssl.
373 This issue was reported to OpenSSL Guido Vranken.
377 *) Side channel attack on modular exponentiation
379 A side-channel attack was found which makes use of cache-bank conflicts on
380 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
381 of RSA keys. The ability to exploit this issue is limited as it relies on
382 an attacker who has control of code in a thread running on the same
383 hyper-threaded core as the victim thread which is performing decryptions.
385 This issue was reported to OpenSSL by Yuval Yarom, The University of
386 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
387 Nadia Heninger, University of Pennsylvania with more information at
388 http://cachebleed.info.
392 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
393 if no keysize is specified with default_bits. This fixes an
394 omission in an earlier change that changed all RSA/DSA key generation
395 apps to use 2048 bits by default.
398 Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
400 *) Protection for DH small subgroup attacks
402 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
403 switched on by default and cannot be disabled. This could have some
407 *) SSLv2 doesn't block disabled ciphers
409 A malicious client can negotiate SSLv2 ciphers that have been disabled on
410 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
411 been disabled, provided that the SSLv2 protocol was not also disabled via
414 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
415 and Sebastian Schinzel.
419 *) Reject DH handshakes with parameters shorter than 1024 bits.
422 Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
424 *) Certificate verify crash with missing PSS parameter
426 The signature verification routines will crash with a NULL pointer
427 dereference if presented with an ASN.1 signature using the RSA PSS
428 algorithm and absent mask generation function parameter. Since these
429 routines are used to verify certificate signature algorithms this can be
430 used to crash any certificate verification operation and exploited in a
431 DoS attack. Any application which performs certificate verification is
432 vulnerable including OpenSSL clients and servers which enable client
435 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
439 *) X509_ATTRIBUTE memory leak
441 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
442 memory. This structure is used by the PKCS#7 and CMS routines so any
443 application which reads PKCS#7 or CMS data from untrusted sources is
444 affected. SSL/TLS is not affected.
446 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
451 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
452 This changes the decoding behaviour for some invalid messages,
453 though the change is mostly in the more lenient direction, and
454 legacy behaviour is preserved as much as possible.
457 *) In DSA_generate_parameters_ex, if the provided seed is too short,
458 use a random seed, as already documented.
459 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
461 Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
463 *) Alternate chains certificate forgery
465 During certificate verfification, OpenSSL will attempt to find an
466 alternative certificate chain if the first attempt to build such a chain
467 fails. An error in the implementation of this logic can mean that an
468 attacker could cause certain checks on untrusted certificates to be
469 bypassed, such as the CA flag, enabling them to use a valid leaf
470 certificate to act as a CA and "issue" an invalid certificate.
472 This issue was reported to OpenSSL by Adam Langley/David Benjamin
477 *) Race condition handling PSK identify hint
479 If PSK identity hints are received by a multi-threaded client then
480 the values are wrongly updated in the parent SSL_CTX structure. This can
481 result in a race condition potentially leading to a double free of the
486 Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
487 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
488 incompatibility in the handling of HMAC. The previous ABI has now been
491 Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
493 *) Malformed ECParameters causes infinite loop
495 When processing an ECParameters structure OpenSSL enters an infinite loop
496 if the curve specified is over a specially malformed binary polynomial
499 This can be used to perform denial of service against any
500 system which processes public keys, certificate requests or
501 certificates. This includes TLS clients and TLS servers with
502 client authentication enabled.
504 This issue was reported to OpenSSL by Joseph Barr-Pixton.
508 *) Exploitable out-of-bounds read in X509_cmp_time
510 X509_cmp_time does not properly check the length of the ASN1_TIME
511 string and can read a few bytes out of bounds. In addition,
512 X509_cmp_time accepts an arbitrary number of fractional seconds in the
515 An attacker can use this to craft malformed certificates and CRLs of
516 various sizes and potentially cause a segmentation fault, resulting in
517 a DoS on applications that verify certificates or CRLs. TLS clients
518 that verify CRLs are affected. TLS clients and servers with client
519 authentication enabled may be affected if they use custom verification
522 This issue was reported to OpenSSL by Robert Swiecki (Google), and
523 independently by Hanno Böck.
527 *) PKCS7 crash with missing EnvelopedContent
529 The PKCS#7 parsing code does not handle missing inner EncryptedContent
530 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
531 with missing content and trigger a NULL pointer dereference on parsing.
533 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
534 structures from untrusted sources are affected. OpenSSL clients and
535 servers are not affected.
537 This issue was reported to OpenSSL by Michal Zalewski (Google).
541 *) CMS verify infinite loop with unknown hash function
543 When verifying a signedData message the CMS code can enter an infinite loop
544 if presented with an unknown hash function OID. This can be used to perform
545 denial of service against any system which verifies signedData messages using
547 This issue was reported to OpenSSL by Johannes Bauer.
551 *) Race condition handling NewSessionTicket
553 If a NewSessionTicket is received by a multi-threaded client when attempting to
554 reuse a previous ticket then a race condition can occur potentially leading to
555 a double free of the ticket data.
559 *) Reject DH handshakes with parameters shorter than 768 bits.
560 [Kurt Roeckx and Emilia Kasper]
562 *) dhparam: generate 2048-bit parameters by default.
563 [Kurt Roeckx and Emilia Kasper]
565 Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
567 *) Segmentation fault in ASN1_TYPE_cmp fix
569 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
570 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
571 certificate signature algorithm consistency this can be used to crash any
572 certificate verification operation and exploited in a DoS attack. Any
573 application which performs certificate verification is vulnerable including
574 OpenSSL clients and servers which enable client authentication.
578 *) ASN.1 structure reuse memory corruption fix
580 Reusing a structure in ASN.1 parsing may allow an attacker to cause
581 memory corruption via an invalid write. Such reuse is and has been
582 strongly discouraged and is believed to be rare.
584 Applications that parse structures containing CHOICE or ANY DEFINED BY
585 components may be affected. Certificate parsing (d2i_X509 and related
586 functions) are however not affected. OpenSSL clients and servers are
591 *) PKCS7 NULL pointer dereferences fix
593 The PKCS#7 parsing code does not handle missing outer ContentInfo
594 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
595 missing content and trigger a NULL pointer dereference on parsing.
597 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
598 otherwise parse PKCS#7 structures from untrusted sources are
599 affected. OpenSSL clients and servers are not affected.
601 This issue was reported to OpenSSL by Michal Zalewski (Google).
605 *) DoS via reachable assert in SSLv2 servers fix
607 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
608 servers that both support SSLv2 and enable export cipher suites by sending
609 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
611 This issue was discovered by Sean Burford (Google) and Emilia Käsper
612 (OpenSSL development team).
616 *) Use After Free following d2i_ECPrivatekey error fix
618 A malformed EC private key file consumed via the d2i_ECPrivateKey function
619 could cause a use after free condition. This, in turn, could cause a double
620 free in several private key parsing functions (such as d2i_PrivateKey
621 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
622 for applications that receive EC private keys from untrusted
623 sources. This scenario is considered rare.
625 This issue was discovered by the BoringSSL project and fixed in their
630 *) X509_to_X509_REQ NULL pointer deref fix
632 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
633 the certificate key is invalid. This function is rarely used in practice.
635 This issue was discovered by Brian Carpenter.
639 *) Removed the export ciphers from the DEFAULT ciphers
642 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
644 *) Build fixes for the Windows and OpenVMS platforms
645 [Matt Caswell and Richard Levitte]
647 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
649 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
650 message can cause a segmentation fault in OpenSSL due to a NULL pointer
651 dereference. This could lead to a Denial Of Service attack. Thanks to
652 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
656 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
657 dtls1_buffer_record function under certain conditions. In particular this
658 could occur if an attacker sent repeated DTLS records with the same
659 sequence number but for the next epoch. The memory leak could be exploited
660 by an attacker in a Denial of Service attack through memory exhaustion.
661 Thanks to Chris Mueller for reporting this issue.
665 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
666 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
667 method would be set to NULL which could later result in a NULL pointer
668 dereference. Thanks to Frank Schmirler for reporting this issue.
672 *) Abort handshake if server key exchange message is omitted for ephemeral
675 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
676 reporting this issue.
680 *) Remove non-export ephemeral RSA code on client and server. This code
681 violated the TLS standard by allowing the use of temporary RSA keys in
682 non-export ciphersuites and could be used by a server to effectively
683 downgrade the RSA key length used to a value smaller than the server
684 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
685 INRIA or reporting this issue.
689 *) Fixed issue where DH client certificates are accepted without verification.
690 An OpenSSL server will accept a DH certificate for client authentication
691 without the certificate verify message. This effectively allows a client to
692 authenticate without the use of a private key. This only affects servers
693 which trust a client certificate authority which issues certificates
694 containing DH keys: these are extremely rare and hardly ever encountered.
695 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
700 *) Ensure that the session ID context of an SSL is updated when its
701 SSL_CTX is updated via SSL_set_SSL_CTX.
703 The session ID context is typically set from the parent SSL_CTX,
704 and can vary with the CTX.
707 *) Fix various certificate fingerprint issues.
709 By using non-DER or invalid encodings outside the signed portion of a
710 certificate the fingerprint can be changed without breaking the signature.
711 Although no details of the signed portion of the certificate can be changed
712 this can cause problems with some applications: e.g. those using the
713 certificate fingerprint for blacklists.
715 1. Reject signatures with non zero unused bits.
717 If the BIT STRING containing the signature has non zero unused bits reject
718 the signature. All current signature algorithms require zero unused bits.
720 2. Check certificate algorithm consistency.
722 Check the AlgorithmIdentifier inside TBS matches the one in the
723 certificate signature. NB: this will result in signature failure
724 errors for some broken certificates.
726 Thanks to Konrad Kraszewski from Google for reporting this issue.
728 3. Check DSA/ECDSA signatures use DER.
730 Reencode DSA/ECDSA signatures and compare with the original received
731 signature. Return an error if there is a mismatch.
733 This will reject various cases including garbage after signature
734 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
735 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
736 (negative or with leading zeroes).
738 Further analysis was conducted and fixes were developed by Stephen Henson
739 of the OpenSSL core team.
744 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
745 results on some platforms, including x86_64. This bug occurs at random
746 with a very low probability, and is not known to be exploitable in any
747 way, though its exact impact is difficult to determine. Thanks to Pieter
748 Wuille (Blockstream) who reported this issue and also suggested an initial
749 fix. Further analysis was conducted by the OpenSSL development team and
750 Adam Langley of Google. The final fix was developed by Andy Polyakov of
751 the OpenSSL core team.
755 *) Do not resume sessions on the server if the negotiated protocol
756 version does not match the session's version. Resuming with a different
757 version, while not strictly forbidden by the RFC, is of questionable
758 sanity and breaks all known clients.
759 [David Benjamin, Emilia Käsper]
761 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
762 early CCS messages during renegotiation. (Note that because
763 renegotiation is encrypted, this early CCS was not exploitable.)
766 *) Tighten client-side session ticket handling during renegotiation:
767 ensure that the client only accepts a session ticket if the server sends
768 the extension anew in the ServerHello. Previously, a TLS client would
769 reuse the old extension state and thus accept a session ticket if one was
770 announced in the initial ServerHello.
772 Similarly, ensure that the client requires a session ticket if one
773 was advertised in the ServerHello. Previously, a TLS client would
774 ignore a missing NewSessionTicket message.
777 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
781 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
782 sends a carefully crafted handshake message, to cause OpenSSL to fail
783 to free up to 64k of memory causing a memory leak. This could be
784 exploited in a Denial Of Service attack. This issue affects OpenSSL
785 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
786 whether SRTP is used or configured. Implementations of OpenSSL that
787 have been compiled with OPENSSL_NO_SRTP defined are not affected.
789 The fix was developed by the OpenSSL team.
793 *) Session Ticket Memory Leak.
795 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
796 integrity of that ticket is first verified. In the event of a session
797 ticket integrity check failing, OpenSSL will fail to free memory
798 causing a memory leak. By sending a large number of invalid session
799 tickets an attacker could exploit this issue in a Denial Of Service
804 *) Build option no-ssl3 is incomplete.
806 When OpenSSL is configured with "no-ssl3" as a build option, servers
807 could accept and complete a SSL 3.0 handshake, and clients could be
808 configured to send them.
810 [Akamai and the OpenSSL team]
812 *) Add support for TLS_FALLBACK_SCSV.
813 Client applications doing fallback retries should call
814 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
816 [Adam Langley, Bodo Moeller]
818 *) Add additional DigestInfo checks.
820 Reencode DigestInto in DER and check against the original when
821 verifying RSA signature: this will reject any improperly encoded
822 DigestInfo structures.
824 Note: this is a precautionary measure and no attacks are currently known.
828 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
830 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
831 SRP code can be overrun an internal buffer. Add sanity check that
832 g, A, B < N to SRP code.
834 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
835 Group for discovering this issue.
839 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
840 TLS 1.0 instead of higher protocol versions when the ClientHello message
841 is badly fragmented. This allows a man-in-the-middle attacker to force a
842 downgrade to TLS 1.0 even if both the server and the client support a
843 higher protocol version, by modifying the client's TLS records.
845 Thanks to David Benjamin and Adam Langley (Google) for discovering and
846 researching this issue.
850 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
851 to a denial of service attack. A malicious server can crash the client
852 with a null pointer dereference (read) by specifying an anonymous (EC)DH
853 ciphersuite and sending carefully crafted handshake messages.
855 Thanks to Felix Gröbert (Google) for discovering and researching this
860 *) By sending carefully crafted DTLS packets an attacker could cause openssl
861 to leak memory. This can be exploited through a Denial of Service attack.
862 Thanks to Adam Langley for discovering and researching this issue.
866 *) An attacker can force openssl to consume large amounts of memory whilst
867 processing DTLS handshake messages. This can be exploited through a
868 Denial of Service attack.
869 Thanks to Adam Langley for discovering and researching this issue.
873 *) An attacker can force an error condition which causes openssl to crash
874 whilst processing DTLS packets due to memory being freed twice. This
875 can be exploited through a Denial of Service attack.
876 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
881 *) If a multithreaded client connects to a malicious server using a resumed
882 session and the server sends an ec point format extension it could write
883 up to 255 bytes to freed memory.
885 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
890 *) A malicious server can crash an OpenSSL client with a null pointer
891 dereference (read) by specifying an SRP ciphersuite even though it was not
892 properly negotiated with the client. This can be exploited through a
893 Denial of Service attack.
895 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
896 discovering and researching this issue.
900 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
901 X509_name_oneline, X509_name_print_ex et al. to leak some information
902 from the stack. Applications may be affected if they echo pretty printing
903 output to the attacker.
905 Thanks to Ivan Fratric (Google) for discovering this issue.
907 [Emilia Käsper, and Steve Henson]
909 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
910 for corner cases. (Certain input points at infinity could lead to
911 bogus results, with non-infinity inputs mapped to infinity too.)
914 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
916 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
917 handshake can force the use of weak keying material in OpenSSL
918 SSL/TLS clients and servers.
920 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
921 researching this issue. (CVE-2014-0224)
922 [KIKUCHI Masashi, Steve Henson]
924 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
925 OpenSSL DTLS client the code can be made to recurse eventually crashing
928 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
930 [Imre Rad, Steve Henson]
932 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
933 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
934 client or server. This is potentially exploitable to run arbitrary
935 code on a vulnerable client or server.
937 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
938 [Jüri Aedla, Steve Henson]
940 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
941 are subject to a denial of service attack.
943 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
944 this issue. (CVE-2014-3470)
945 [Felix Gröbert, Ivan Fratric, Steve Henson]
947 *) Harmonize version and its documentation. -f flag is used to display
949 [mancha <mancha1@zoho.com>]
951 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
953 [mancha <mancha1@zoho.com>]
955 *) Fix some double frees. These are not thought to be exploitable.
956 [mancha <mancha1@zoho.com>]
958 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
960 *) A missing bounds check in the handling of the TLS heartbeat extension
961 can be used to reveal up to 64k of memory to a connected client or
964 Thanks for Neel Mehta of Google Security for discovering this bug and to
965 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
966 preparing the fix (CVE-2014-0160)
967 [Adam Langley, Bodo Moeller]
969 *) Fix for the attack described in the paper "Recovering OpenSSL
970 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
971 by Yuval Yarom and Naomi Benger. Details can be obtained from:
972 http://eprint.iacr.org/2014/140
974 Thanks to Yuval Yarom and Naomi Benger for discovering this
975 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
976 [Yuval Yarom and Naomi Benger]
978 *) TLS pad extension: draft-agl-tls-padding-03
980 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
981 TLS client Hello record length value would otherwise be > 255 and
982 less that 512 pad with a dummy extension containing zeroes so it
983 is at least 512 bytes long.
985 [Adam Langley, Steve Henson]
987 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
989 *) Fix for TLS record tampering bug. A carefully crafted invalid
990 handshake could crash OpenSSL with a NULL pointer exception.
991 Thanks to Anton Johansson for reporting this issues.
994 *) Keep original DTLS digest and encryption contexts in retransmission
995 structures so we can use the previous session parameters if they need
996 to be resent. (CVE-2013-6450)
999 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1000 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1001 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1002 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1003 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1004 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1005 [Rob Stradling, Adam Langley]
1007 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1009 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1010 supporting platforms or when small records were transferred.
1011 [Andy Polyakov, Steve Henson]
1013 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1015 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1017 This addresses the flaw in CBC record processing discovered by
1018 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1019 at: http://www.isg.rhul.ac.uk/tls/
1021 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1022 Security Group at Royal Holloway, University of London
1023 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1024 Emilia Käsper for the initial patch.
1026 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1028 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1029 ciphersuites which can be exploited in a denial of service attack.
1030 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1031 and detecting this bug and to Wolfgang Ettlinger
1032 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1036 *) Return an error when checking OCSP signatures when key is NULL.
1037 This fixes a DoS attack. (CVE-2013-0166)
1040 *) Make openssl verify return errors.
1041 [Chris Palmer <palmer@google.com> and Ben Laurie]
1043 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1044 the right response is stapled. Also change SSL_get_certificate()
1045 so it returns the certificate actually sent.
1046 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1047 [Rob Stradling <rob.stradling@comodo.com>]
1049 *) Fix possible deadlock when decoding public keys.
1052 *) Don't use TLS 1.0 record version number in initial client hello
1056 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1058 *) Sanity check record length before skipping explicit IV in TLS
1059 1.2, 1.1 and DTLS to fix DoS attack.
1061 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1062 fuzzing as a service testing platform.
1066 *) Initialise tkeylen properly when encrypting CMS messages.
1067 Thanks to Solar Designer of Openwall for reporting this issue.
1070 *) In FIPS mode don't try to use composite ciphers as they are not
1074 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1076 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1077 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1078 mean any application compiled against OpenSSL 1.0.0 headers setting
1079 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1080 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1081 0x10000000L Any application which was previously compiled against
1082 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1083 will need to be recompiled as a result. Letting be results in
1084 inability to disable specifically TLS 1.1 and in client context,
1085 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1088 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1089 disable just protocol X, but all protocols above X *if* there are
1090 protocols *below* X still enabled. In more practical terms it means
1091 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1092 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1093 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1097 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1099 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1100 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1101 in CRYPTO_realloc_clean.
1103 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1104 issue and to Adam Langley <agl@chromium.org> for fixing it.
1106 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1108 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1111 *) Workarounds for some broken servers that "hang" if a client hello
1112 record length exceeds 255 bytes.
1114 1. Do not use record version number > TLS 1.0 in initial client
1115 hello: some (but not all) hanging servers will now work.
1116 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1117 the number of ciphers sent in the client hello. This should be
1118 set to an even number, such as 50, for example by passing:
1119 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1120 Most broken servers should now work.
1121 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1122 TLS 1.2 client support entirely.
1125 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1128 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1130 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1131 STRING form instead of a DigestInfo.
1134 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1135 and the RSA_sign/RSA_verify functions. This was made more apparent when
1136 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1137 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1138 the correct format in RSA_verify so both forms transparently work.
1141 *) Some servers which support TLS 1.0 can choke if we initially indicate
1142 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1143 encrypted premaster secret. As a workaround use the maximum pemitted
1144 client version in client hello, this should keep such servers happy
1145 and still work with previous versions of OpenSSL.
1148 *) Add support for TLS/DTLS heartbeats.
1149 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1151 *) Add support for SCTP.
1152 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1154 *) Improved PRNG seeding for VOS.
1155 [Paul Green <Paul.Green@stratus.com>]
1157 *) Extensive assembler packs updates, most notably:
1159 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1160 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1161 - x86_64: bit-sliced AES implementation;
1162 - ARM: NEON support, contemporary platforms optimizations;
1163 - s390x: z196 support;
1164 - *: GHASH and GF(2^m) multiplication implementations;
1168 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1169 (removal of unnecessary code)
1170 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1172 *) Add TLS key material exporter from RFC 5705.
1175 *) Add DTLS-SRTP negotiation from RFC 5764.
1178 *) Add Next Protocol Negotiation,
1179 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1180 disabled with a no-npn flag to config or Configure. Code donated
1182 [Adam Langley <agl@google.com> and Ben Laurie]
1184 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1185 NIST-P256, NIST-P521, with constant-time single point multiplication on
1186 typical inputs. Compiler support for the nonstandard type __uint128_t is
1187 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1188 Code made available under Apache License version 2.0.
1190 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1191 line to include this in your build of OpenSSL, and run "make depend" (or
1192 "make update"). This enables the following EC_METHODs:
1194 EC_GFp_nistp224_method()
1195 EC_GFp_nistp256_method()
1196 EC_GFp_nistp521_method()
1198 EC_GROUP_new_by_curve_name() will automatically use these (while
1199 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1201 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1203 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1204 all platforms. Move ssize_t definition from e_os.h to the public
1205 header file e_os2.h as it now appears in public header file cms.h
1208 *) New -sigopt option to the ca, req and x509 utilities. Additional
1209 signature parameters can be passed using this option and in
1213 *) Add RSA PSS signing function. This will generate and set the
1214 appropriate AlgorithmIdentifiers for PSS based on those in the
1215 corresponding EVP_MD_CTX structure. No application support yet.
1218 *) Support for companion algorithm specific ASN1 signing routines.
1219 New function ASN1_item_sign_ctx() signs a pre-initialised
1220 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1221 the appropriate parameters.
1224 *) Add new algorithm specific ASN1 verification initialisation function
1225 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1226 handling will be the same no matter what EVP_PKEY_METHOD is used.
1227 Add a PSS handler to support verification of PSS signatures: checked
1228 against a number of sample certificates.
1231 *) Add signature printing for PSS. Add PSS OIDs.
1232 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1234 *) Add algorithm specific signature printing. An individual ASN1 method
1235 can now print out signatures instead of the standard hex dump.
1237 More complex signatures (e.g. PSS) can print out more meaningful
1238 information. Include DSA version that prints out the signature
1242 *) Password based recipient info support for CMS library: implementing
1246 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1247 neatly separates the code into cipher and PBE sections and is required
1248 for some algorithms that split PBES2 into separate pieces (such as
1249 password based CMS).
1252 *) Session-handling fixes:
1253 - Fix handling of connections that are resuming with a session ID,
1254 but also support Session Tickets.
1255 - Fix a bug that suppressed issuing of a new ticket if the client
1256 presented a ticket with an expired session.
1257 - Try to set the ticket lifetime hint to something reasonable.
1258 - Make tickets shorter by excluding irrelevant information.
1259 - On the client side, don't ignore renewed tickets.
1260 [Adam Langley, Bodo Moeller (Google)]
1262 *) Fix PSK session representation.
1265 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1267 This work was sponsored by Intel.
1270 *) Add GCM support to TLS library. Some custom code is needed to split
1271 the IV between the fixed (from PRF) and explicit (from TLS record)
1272 portions. This adds all GCM ciphersuites supported by RFC5288 and
1273 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
1274 add a special AESGCM string for GCM only.
1277 *) Expand range of ctrls for AES GCM. Permit setting invocation
1278 field on decrypt and retrieval of invocation field only on encrypt.
1281 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1282 As required by RFC5289 these ciphersuites cannot be used if for
1283 versions of TLS earlier than 1.2.
1286 *) For FIPS capable OpenSSL interpret a NULL default public key method
1287 as unset and return the appopriate default but do *not* set the default.
1288 This means we can return the appopriate method in applications that
1289 swicth between FIPS and non-FIPS modes.
1292 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1293 ENGINE is used then we cannot handle that in the FIPS module so we
1294 keep original code iff non-FIPS operations are allowed.
1297 *) Add -attime option to openssl utilities.
1298 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1300 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1303 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1304 FIPS EC methods unconditionally for now.
1307 *) New build option no-ec2m to disable characteristic 2 code.
1310 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1311 all cases can be covered as some introduce binary incompatibilities.
1314 *) Redirect RSA operations to FIPS module including keygen,
1315 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1318 *) Add similar low level API blocking to ciphers.
1321 *) Low level digest APIs are not approved in FIPS mode: any attempt
1322 to use these will cause a fatal error. Applications that *really* want
1323 to use them can use the private_* version instead.
1326 *) Redirect cipher operations to FIPS module for FIPS builds.
1329 *) Redirect digest operations to FIPS module for FIPS builds.
1332 *) Update build system to add "fips" flag which will link in fipscanister.o
1333 for static and shared library builds embedding a signature if needed.
1336 *) Output TLS supported curves in preference order instead of numerical
1337 order. This is currently hardcoded for the highest order curves first.
1338 This should be configurable so applications can judge speed vs strength.
1341 *) Add TLS v1.2 server support for client authentication.
1344 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1348 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1349 FIPS modules versions.
1352 *) Add TLS v1.2 client side support for client authentication. Keep cache
1353 of handshake records longer as we don't know the hash algorithm to use
1354 until after the certificate request message is received.
1357 *) Initial TLS v1.2 client support. Add a default signature algorithms
1358 extension including all the algorithms we support. Parse new signature
1359 format in client key exchange. Relax some ECC signing restrictions for
1360 TLS v1.2 as indicated in RFC5246.
1363 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1364 to new signature format when needed using client digest preference.
1365 All server ciphersuites should now work correctly in TLS v1.2. No client
1366 support yet and no support for client certificates.
1369 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1370 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1371 ciphersuites. At present only RSA key exchange ciphersuites work with
1372 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1373 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1374 and version checking.
1377 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1378 with this defined it will not be affected by any changes to ssl internal
1379 structures. Add several utility functions to allow openssl application
1380 to work with OPENSSL_NO_SSL_INTERN defined.
1384 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1386 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1389 *) Permit abbreviated handshakes when renegotiating using the function
1390 SSL_renegotiate_abbreviated().
1391 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1393 *) Add call to ENGINE_register_all_complete() to
1394 ENGINE_load_builtin_engines(), so some implementations get used
1395 automatically instead of needing explicit application support.
1398 *) Add support for TLS key exporter as described in RFC5705.
1399 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1401 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1402 a few changes are required:
1404 Add SSL_OP_NO_TLSv1_1 flag.
1405 Add TLSv1_1 methods.
1406 Update version checking logic to handle version 1.1.
1407 Add explicit IV handling (ported from DTLS code).
1408 Add command line options to s_client/s_server.
1411 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1413 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1414 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1415 content decryption and always return the same error. Note: this attack
1416 needs on average 2^20 messages so it only affects automated senders. The
1417 old behaviour can be reenabled in the CMS code by setting the
1418 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1419 an MMA defence is not necessary.
1420 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1421 this issue. (CVE-2012-0884)
1424 *) Fix CVE-2011-4619: make sure we really are receiving a
1425 client hello before rejecting multiple SGC restarts. Thanks to
1426 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1429 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1431 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1432 Thanks to Antonio Martin, Enterprise Secure Access Research and
1433 Development, Cisco Systems, Inc. for discovering this bug and
1434 preparing a fix. (CVE-2012-0050)
1437 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1439 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1440 of the Vaudenay padding oracle attack on CBC mode encryption
1441 which enables an efficient plaintext recovery attack against
1442 the OpenSSL implementation of DTLS. Their attack exploits timing
1443 differences arising during decryption processing. A research
1444 paper describing this attack can be found at:
1445 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1446 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1447 Security Group at Royal Holloway, University of London
1448 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1449 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1450 for preparing the fix. (CVE-2011-4108)
1451 [Robin Seggelmann, Michael Tuexen]
1453 *) Clear bytes used for block padding of SSL 3.0 records.
1455 [Adam Langley (Google)]
1457 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1458 Kadianakis <desnacked@gmail.com> for discovering this issue and
1459 Adam Langley for preparing the fix. (CVE-2011-4619)
1460 [Adam Langley (Google)]
1462 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1463 [Andrey Kulikov <amdeich@gmail.com>]
1465 *) Prevent malformed RFC3779 data triggering an assertion failure.
1466 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1467 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1468 [Rob Austein <sra@hactrn.net>]
1470 *) Improved PRNG seeding for VOS.
1471 [Paul Green <Paul.Green@stratus.com>]
1473 *) Fix ssl_ciph.c set-up race.
1474 [Adam Langley (Google)]
1476 *) Fix spurious failures in ecdsatest.c.
1477 [Emilia Käsper (Google)]
1479 *) Fix the BIO_f_buffer() implementation (which was mixing different
1480 interpretations of the '..._len' fields).
1481 [Adam Langley (Google)]
1483 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1484 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1485 threads won't reuse the same blinding coefficients.
1487 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1488 lock to call BN_BLINDING_invert_ex, and avoids one use of
1489 BN_BLINDING_update for each BN_BLINDING structure (previously,
1490 the last update always remained unused).
1491 [Emilia Käsper (Google)]
1493 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1494 [Bob Buckholz (Google)]
1496 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1498 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1499 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1500 [Kaspar Brand <ossl@velox.ch>]
1502 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1503 for multi-threaded use of ECDH. (CVE-2011-3210)
1504 [Adam Langley (Google)]
1506 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1509 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1510 signature public key algorithm by using OID xref utilities instead.
1511 Before this you could only use some ECC ciphersuites with SHA1 only.
1514 *) Add protection against ECDSA timing attacks as mentioned in the paper
1515 by Billy Bob Brumley and Nicola Tuveri, see:
1517 http://eprint.iacr.org/2011/232.pdf
1519 [Billy Bob Brumley and Nicola Tuveri]
1521 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1523 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1524 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1526 *) Fix bug in string printing code: if *any* escaping is enabled we must
1527 escape the escape character (backslash) or the resulting string is
1531 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1533 *) Disable code workaround for ancient and obsolete Netscape browsers
1534 and servers: an attacker can use it in a ciphersuite downgrade attack.
1535 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1538 *) Fixed J-PAKE implementation error, originally discovered by
1539 Sebastien Martini, further info and confirmation from Stefan
1540 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1543 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1545 *) Fix extension code to avoid race conditions which can result in a buffer
1546 overrun vulnerability: resumed sessions must not be modified as they can
1547 be shared by multiple threads. CVE-2010-3864
1550 *) Fix WIN32 build system to correctly link an ENGINE directory into
1554 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1556 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1558 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1560 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1562 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1563 context. The operation can be customised via the ctrl mechanism in
1564 case ENGINEs want to include additional functionality.
1567 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1570 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1571 output hashes compatible with older versions of OpenSSL.
1572 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1574 *) Fix compression algorithm handling: if resuming a session use the
1575 compression algorithm of the resumed session instead of determining
1576 it from client hello again. Don't allow server to change algorithm.
1579 *) Add load_crls() function to apps tidying load_certs() too. Add option
1580 to verify utility to allow additional CRLs to be included.
1583 *) Update OCSP request code to permit adding custom headers to the request:
1584 some responders need this.
1587 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1589 [Julia Lawall <julia@diku.dk>]
1591 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1592 needlessly dereferenced structures, used obsolete functions and
1593 didn't handle all updated verify codes correctly.
1596 *) Disable MD2 in the default configuration.
1599 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1600 indicate the initial BIO being pushed or popped. This makes it possible
1601 to determine whether the BIO is the one explicitly called or as a result
1602 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1603 it handles reference counts correctly and doesn't zero out the I/O bio
1604 when it is not being explicitly popped. WARNING: applications which
1605 included workarounds for the old buggy behaviour will need to be modified
1606 or they could free up already freed BIOs.
1609 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1610 renaming to all platforms (within the 0.9.8 branch, this was
1611 done conditionally on Netware platforms to avoid a name clash).
1612 [Guenter <lists@gknw.net>]
1614 *) Add ECDHE and PSK support to DTLS.
1615 [Michael Tuexen <tuexen@fh-muenster.de>]
1617 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1621 *) Add "missing" function EVP_MD_flags() (without this the only way to
1622 retrieve a digest flags is by accessing the structure directly. Update
1623 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1624 or cipher is registered as in the "from" argument. Print out all
1625 registered digests in the dgst usage message instead of manually
1626 attempting to work them out.
1629 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1630 this allows the use of compression and extensions. Change default cipher
1631 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1632 by default unless an application cipher string requests it.
1635 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1636 key ids to find matching certificates and keys but some PKCS#12 files
1637 don't follow the (somewhat unwritten) rules and this strategy fails.
1638 Now just gather all certificates together and the first private key
1639 then look for the first certificate that matches the key.
1642 *) Support use of registered digest and cipher names for dgst and cipher
1643 commands instead of having to add each one as a special case. So now
1650 openssl dgst -sha256 foo
1652 and this works for ENGINE based algorithms too.
1656 *) Update Gost ENGINE to support parameter files.
1657 [Victor B. Wagner <vitus@cryptocom.ru>]
1659 *) Support GeneralizedTime in ca utility.
1660 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1662 *) Enhance the hash format used for certificate directory links. The new
1663 form uses the canonical encoding (meaning equivalent names will work
1664 even if they aren't identical) and uses SHA1 instead of MD5. This form
1665 is incompatible with the older format and as a result c_rehash should
1666 be used to rebuild symbolic links.
1669 *) Make PKCS#8 the default write format for private keys, replacing the
1670 traditional format. This form is standardised, more secure and doesn't
1671 include an implicit MD5 dependency.
1674 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1675 committed to OpenSSL should pass this lot as a minimum.
1678 *) Add session ticket override functionality for use by EAP-FAST.
1679 [Jouni Malinen <j@w1.fi>]
1681 *) Modify HMAC functions to return a value. Since these can be implemented
1682 in an ENGINE errors can occur.
1685 *) Type-checked OBJ_bsearch_ex.
1688 *) Type-checked OBJ_bsearch. Also some constification necessitated
1689 by type-checking. Still to come: TXT_DB, bsearch(?),
1690 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1694 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1695 seconds to a tm structure directly, instead of going through OS
1696 specific date routines. This avoids any issues with OS routines such
1697 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1698 and X509_time_adj_ex() to cover the extended range. The existing
1699 X509_time_adj() is still usable and will no longer have any date issues.
1702 *) Delta CRL support. New use deltas option which will attempt to locate
1703 and search any appropriate delta CRLs available.
1705 This work was sponsored by Google.
1708 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1709 code and add additional score elements. Validate alternate CRL paths
1710 as part of the CRL checking and indicate a new error "CRL path validation
1711 error" in this case. Applications wanting additional details can use
1712 the verify callback and check the new "parent" field. If this is not
1713 NULL CRL path validation is taking place. Existing applications wont
1714 see this because it requires extended CRL support which is off by
1717 This work was sponsored by Google.
1720 *) Support for freshest CRL extension.
1722 This work was sponsored by Google.
1725 *) Initial indirect CRL support. Currently only supported in the CRLs
1726 passed directly and not via lookup. Process certificate issuer
1727 CRL entry extension and lookup CRL entries by bother issuer name
1728 and serial number. Check and process CRL issuer entry in IDP extension.
1730 This work was sponsored by Google.
1733 *) Add support for distinct certificate and CRL paths. The CRL issuer
1734 certificate is validated separately in this case. Only enabled if
1735 an extended CRL support flag is set: this flag will enable additional
1736 CRL functionality in future.
1738 This work was sponsored by Google.
1741 *) Add support for policy mappings extension.
1743 This work was sponsored by Google.
1746 *) Fixes to pathlength constraint, self issued certificate handling,
1747 policy processing to align with RFC3280 and PKITS tests.
1749 This work was sponsored by Google.
1752 *) Support for name constraints certificate extension. DN, email, DNS
1753 and URI types are currently supported.
1755 This work was sponsored by Google.
1758 *) To cater for systems that provide a pointer-based thread ID rather
1759 than numeric, deprecate the current numeric thread ID mechanism and
1760 replace it with a structure and associated callback type. This
1761 mechanism allows a numeric "hash" to be extracted from a thread ID in
1762 either case, and on platforms where pointers are larger than 'long',
1763 mixing is done to help ensure the numeric 'hash' is usable even if it
1764 can't be guaranteed unique. The default mechanism is to use "&errno"
1765 as a pointer-based thread ID to distinguish between threads.
1767 Applications that want to provide their own thread IDs should now use
1768 CRYPTO_THREADID_set_callback() to register a callback that will call
1769 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1771 Note that ERR_remove_state() is now deprecated, because it is tied
1772 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1773 to free the current thread's error state should be replaced by
1774 ERR_remove_thread_state(NULL).
1776 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1777 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1778 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1779 application was previously providing a numeric thread callback that
1780 was inappropriate for distinguishing threads, then uniqueness might
1781 have been obtained with &errno that happened immediately in the
1782 intermediate development versions of OpenSSL; this is no longer the
1783 case, the numeric thread callback will now override the automatic use
1785 [Geoff Thorpe, with help from Bodo Moeller]
1787 *) Initial support for different CRL issuing certificates. This covers a
1788 simple case where the self issued certificates in the chain exist and
1789 the real CRL issuer is higher in the existing chain.
1791 This work was sponsored by Google.
1794 *) Removed effectively defunct crypto/store from the build.
1797 *) Revamp of STACK to provide stronger type-checking. Still to come:
1798 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1799 ASN1_STRING, CONF_VALUE.
1802 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1803 RAM on SSL connections. This option can save about 34k per idle SSL.
1806 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1807 STACK, TXT_DB, bsearch, qsort.
1810 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1811 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1812 support for data, signedData, compressedData, digestedData and
1813 encryptedData, envelopedData types included. Scripts to check against
1814 RFC4134 examples draft and interop and consistency checks of many
1815 content types and variants.
1818 *) Add options to enc utility to support use of zlib compression BIO.
1821 *) Extend mk1mf to support importing of options and assembly language
1822 files from Configure script, currently only included in VC-WIN32.
1823 The assembly language rules can now optionally generate the source
1824 files from the associated perl scripts.
1827 *) Implement remaining functionality needed to support GOST ciphersuites.
1828 Interop testing has been performed using CryptoPro implementations.
1829 [Victor B. Wagner <vitus@cryptocom.ru>]
1831 *) s390x assembler pack.
1834 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1838 *) Implement Opaque PRF Input TLS extension as specified in
1839 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1840 official specification yet and no extension type assignment by
1841 IANA exists, this extension (for now) will have to be explicitly
1842 enabled when building OpenSSL by providing the extension number
1843 to use. For example, specify an option
1845 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1847 to the "config" or "Configure" script to enable the extension,
1848 assuming extension number 0x9527 (which is a completely arbitrary
1849 and unofficial assignment based on the MD5 hash of the Internet
1850 Draft). Note that by doing so, you potentially lose
1851 interoperability with other TLS implementations since these might
1852 be using the same extension number for other purposes.
1854 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1855 opaque PRF input value to use in the handshake. This will create
1856 an interal copy of the length-'len' string at 'src', and will
1857 return non-zero for success.
1859 To get more control and flexibility, provide a callback function
1862 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1863 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1867 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1870 Callback function 'cb' will be called in handshakes, and is
1871 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1872 Argument 'arg' is for application purposes (the value as given to
1873 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1874 be provided to the callback function). The callback function
1875 has to return non-zero to report success: usually 1 to use opaque
1876 PRF input just if possible, or 2 to enforce use of the opaque PRF
1877 input. In the latter case, the library will abort the handshake
1878 if opaque PRF input is not successfully negotiated.
1880 Arguments 'peerinput' and 'len' given to the callback function
1881 will always be NULL and 0 in the case of a client. A server will
1882 see the client's opaque PRF input through these variables if
1883 available (NULL and 0 otherwise). Note that if the server
1884 provides an opaque PRF input, the length must be the same as the
1885 length of the client's opaque PRF input.
1887 Note that the callback function will only be called when creating
1888 a new session (session resumption can resume whatever was
1889 previously negotiated), and will not be called in SSL 2.0
1890 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1891 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1892 for applications that need to enforce opaque PRF input.
1896 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1899 [Victor B. Wagner <vitus@cryptocom.ru>]
1901 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1902 RFC4507bis. The encrypted ticket format is an encrypted encoded
1903 SSL_SESSION structure, that way new session features are automatically
1906 If a client application caches session in an SSL_SESSION structure
1907 support is transparent because tickets are now stored in the encoded
1910 The SSL_CTX structure automatically generates keys for ticket
1911 protection in servers so again support should be possible
1912 with no application modification.
1914 If a client or server wishes to disable RFC4507 support then the option
1915 SSL_OP_NO_TICKET can be set.
1917 Add a TLS extension debugging callback to allow the contents of any client
1918 or server extensions to be examined.
1920 This work was sponsored by Google.
1923 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1924 OpenSSL should now compile cleanly on gcc 4.2
1925 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1927 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1928 support including streaming MAC support: this is required for GOST
1929 ciphersuite support.
1930 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1932 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1933 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1934 to output in BER and PEM format.
1937 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1938 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1939 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1940 ENGINE support for HMAC keys which are unextractable. New -mac and
1941 -macopt options to dgst utility.
1944 *) New option -sigopt to dgst utility. Update dgst to use
1945 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1946 alternative signing paramaters such as X9.31 or PSS in the dgst
1950 *) Change ssl_cipher_apply_rule(), the internal function that does
1951 the work each time a ciphersuite string requests enabling
1952 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1953 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1954 the order of disabled ciphersuites such that those ciphersuites
1955 that most recently went from enabled to disabled not only stay
1956 in order with respect to each other, but also have higher priority
1957 than other disabled ciphersuites the next time ciphersuites are
1960 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1961 the same ciphersuites as with "HIGH" alone, but in a specific
1962 order where the PSK ciphersuites come first (since they are the
1963 most recently disabled ciphersuites when "HIGH" is parsed).
1965 Also, change ssl_create_cipher_list() (using this new
1966 funcionality) such that between otherwise identical
1967 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1971 *) Change ssl_create_cipher_list() so that it automatically
1972 arranges the ciphersuites in reasonable order before starting
1973 to process the rule string. Thus, the definition for "DEFAULT"
1974 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1975 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1976 This makes it much easier to arrive at a reasonable default order
1977 in applications for which anonymous ciphers are OK (meaning
1978 that you can't actually use DEFAULT).
1979 [Bodo Moeller; suggested by Victor Duchovni]
1981 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1982 processing) into multiple integers instead of setting
1983 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1984 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1985 (These masks as well as the individual bit definitions are hidden
1986 away into the non-exported interface ssl/ssl_locl.h, so this
1987 change to the definition of the SSL_CIPHER structure shouldn't
1988 affect applications.) This give us more bits for each of these
1989 categories, so there is no longer a need to coagulate AES128 and
1990 AES256 into a single algorithm bit, and to coagulate Camellia128
1991 and Camellia256 into a single algorithm bit, which has led to all
1994 Thus, among other things, the kludge introduced in 0.9.7m and
1995 0.9.8e for masking out AES256 independently of AES128 or masking
1996 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1998 With the change, we also introduce new ciphersuite aliases that
1999 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2003 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2004 Use the leftmost N bytes of the signature input if the input is
2005 larger than the prime q (with N being the size in bytes of q).
2008 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2009 it yet and it is largely untested.
2012 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2015 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2016 some compilers (gcc 4.2 and later) reject their use. Safestack is
2017 reimplemented. Update ASN1 to avoid use of legacy functions.
2020 *) Win32/64 targets are linked with Winsock2.
2023 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2024 to external functions. This can be used to increase CRL handling
2025 efficiency especially when CRLs are very large by (for example) storing
2026 the CRL revoked certificates in a database.
2029 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2030 new CRLs added to a directory can be used. New command line option
2031 -verify_return_error to s_client and s_server. This causes real errors
2032 to be returned by the verify callback instead of carrying on no matter
2033 what. This reflects the way a "real world" verify callback would behave.
2036 *) GOST engine, supporting several GOST algorithms and public key formats.
2037 Kindly donated by Cryptocom.
2040 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2041 partitioned by DP are handled but no indirect CRL or reason partitioning
2042 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2043 selected via a scoring technique which handles IDP and AKID in CRLs.
2046 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2047 will ultimately be used for all verify operations: this will remove the
2048 X509_STORE dependency on certificate verification and allow alternative
2049 lookup methods. X509_STORE based implementations of these two callbacks.
2052 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2053 Modify get_crl() to find a valid (unexpired) CRL if possible.
2056 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2057 this would be called X509_CRL_cmp() but that name is already used by
2058 a function that just compares CRL issuer names. Cache several CRL
2059 extensions in X509_CRL structure and cache CRLDP in X509.
2062 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2063 this maps equivalent X509_NAME structures into a consistent structure.
2064 Name comparison can then be performed rapidly using memcmp().
2067 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2071 *) Allow digests to supply their own micalg string for S/MIME type using
2072 the ctrl EVP_MD_CTRL_MICALG.
2075 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2076 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2077 ctrl. It can then customise the structure before and/or after signing
2081 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2082 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2083 to free up any added signature OIDs.
2086 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2087 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2088 digest and cipher tables. New options added to openssl utility:
2089 list-message-digest-algorithms and list-cipher-algorithms.
2092 *) Change the array representation of binary polynomials: the list
2093 of degrees of non-zero coefficients is now terminated with -1.
2094 Previously it was terminated with 0, which was also part of the
2095 value; thus, the array representation was not applicable to
2096 polynomials where t^0 has coefficient zero. This change makes
2097 the array representation useful in a more general context.
2100 *) Various modifications and fixes to SSL/TLS cipher string
2101 handling. For ECC, the code now distinguishes between fixed ECDH
2102 with RSA certificates on the one hand and with ECDSA certificates
2103 on the other hand, since these are separate ciphersuites. The
2104 unused code for Fortezza ciphersuites has been removed.
2106 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2107 (not "ECDHE"). For consistency with the code for DH
2108 certificates, use of ECDH certificates is now considered ECDH
2109 authentication, not RSA or ECDSA authentication (the latter is
2110 merely the CA's signing algorithm and not actively used in the
2113 The temporary ciphersuite alias "ECCdraft" is no longer
2114 available, and ECC ciphersuites are no longer excluded from "ALL"
2115 and "DEFAULT". The following aliases now exist for RFC 4492
2116 ciphersuites, most of these by analogy with the DH case:
2118 kECDHr - ECDH cert, signed with RSA
2119 kECDHe - ECDH cert, signed with ECDSA
2120 kECDH - ECDH cert (signed with either RSA or ECDSA)
2121 kEECDH - ephemeral ECDH
2122 ECDH - ECDH cert or ephemeral ECDH
2128 AECDH - anonymous ECDH
2129 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2133 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2134 Use correct micalg parameters depending on digest(s) in signed message.
2137 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2138 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2141 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2142 an engine to register a method. Add ENGINE lookups for methods and
2143 functional reference processing.
2146 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2147 EVP_{Sign,Verify}* which allow an application to customise the signature
2151 *) New -resign option to smime utility. This adds one or more signers
2152 to an existing PKCS#7 signedData structure. Also -md option to use an
2153 alternative message digest algorithm for signing.
2156 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2157 create PKCS7 structures containing multiple signers. Update smime
2158 application to support multiple signers.
2161 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2165 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2166 Reorganize PBE internals to lookup from a static table using NIDs,
2167 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2168 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2169 PRF which will be automatically used with PBES2.
2172 *) Replace the algorithm specific calls to generate keys in "req" with the
2176 *) Update PKCS#7 enveloped data routines to use new API. This is now
2177 supported by any public key method supporting the encrypt operation. A
2178 ctrl is added to allow the public key algorithm to examine or modify
2179 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2183 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2184 a default digest type to use. In most cases this will be SHA1 but some
2185 algorithms (such as GOST) need to specify an alternative digest. The
2186 return value indicates how strong the prefernce is 1 means optional and
2187 2 is mandatory (that is it is the only supported type). Modify
2188 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2189 use the default md. Update openssl utilities to use the default digest
2190 type for signing if it is not explicitly indicated.
2193 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2194 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2195 signing method from the key type. This effectively removes the link
2196 between digests and public key types.
2199 *) Add an OID cross reference table and utility functions. Its purpose is to
2200 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2201 rsaEncryption. This will allow some of the algorithm specific hackery
2202 needed to use the correct OID to be removed.
2205 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2206 structures for PKCS7_sign(). They are now set up by the relevant public
2210 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2213 *) Add support for key derivation (agreement) in the API, DH method and
2217 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2218 public and private key formats. As a side effect these add additional
2219 command line functionality not previously available: DSA signatures can be
2220 generated and verified using pkeyutl and DH key support and generation in
2225 [Oliver Tappe <zooey@hirschkaefer.de>]
2227 *) New make target "install_html_docs" installs HTML renditions of the
2229 [Oliver Tappe <zooey@hirschkaefer.de>]
2231 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
2232 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2233 support key and parameter generation and add initial key generation
2234 functionality for RSA.
2237 *) Add functions for main EVP_PKEY_method operations. The undocumented
2238 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2239 EVP_PKEY_{encrypt,decrypt}_old.
2242 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2243 key API, doesn't do much yet.
2246 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2247 public key algorithms. New option to openssl utility:
2248 "list-public-key-algorithms" to print out info.
2251 *) Implement the Supported Elliptic Curves Extension for
2252 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2255 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2256 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2259 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2260 utilities such as rsa, dsa, dsaparam etc except they process any key
2264 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2265 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2266 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2270 *) Initial support for pluggable public key ASN1.
2271 De-spaghettify the public key ASN1 handling. Move public and private
2272 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2273 algorithm specific handling to a single module within the relevant
2274 algorithm directory. Add functions to allow (near) opaque processing
2275 of public and private key structures.
2278 *) Implement the Supported Point Formats Extension for
2279 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2282 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2283 for the psk identity [hint] and the psk callback functions to the
2284 SSL_SESSION, SSL and SSL_CTX structure.
2287 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2291 SSL_CTX_use_psk_identity_hint
2292 SSL_get_psk_identity_hint
2293 SSL_get_psk_identity
2294 SSL_use_psk_identity_hint
2296 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2298 *) Add RFC 3161 compliant time stamp request creation, response generation
2299 and response verification functionality.
2300 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2302 *) Add initial support for TLS extensions, specifically for the server_name
2303 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2304 have new members for a host name. The SSL data structure has an
2305 additional member SSL_CTX *initial_ctx so that new sessions can be
2306 stored in that context to allow for session resumption, even after the
2307 SSL has been switched to a new SSL_CTX in reaction to a client's
2308 server_name extension.
2310 New functions (subject to change):
2312 SSL_get_servername()
2313 SSL_get_servername_type()
2316 New CTRL codes and macros (subject to change):
2318 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2319 - SSL_CTX_set_tlsext_servername_callback()
2320 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2321 - SSL_CTX_set_tlsext_servername_arg()
2322 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2324 openssl s_client has a new '-servername ...' option.
2326 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2327 '-key2 ...', '-servername_fatal' (subject to change). This allows
2328 testing the HostName extension for a specific single host name ('-cert'
2329 and '-key' remain fallbacks for handshakes without HostName
2330 negotiation). If the unrecogninzed_name alert has to be sent, this by
2331 default is a warning; it becomes fatal with the '-servername_fatal'
2334 [Peter Sylvester, Remy Allais, Christophe Renou]
2336 *) Whirlpool hash implementation is added.
2339 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2340 bn(64,32). Because of instruction set limitations it doesn't have
2341 any negative impact on performance. This was done mostly in order
2342 to make it possible to share assembler modules, such as bn_mul_mont
2343 implementations, between 32- and 64-bit builds without hassle.
2346 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2347 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2351 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2352 dedicated Montgomery multiplication procedure, is introduced.
2353 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2354 "64-bit" performance on certain 32-bit targets.
2357 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2358 in SSL structures. New SSL ctrl to set maximum send fragment size.
2359 Save memory by seeting the I/O buffer sizes dynamically instead of
2360 using the maximum available value.
2363 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2364 in addition to the text details.
2367 *) Very, very preliminary EXPERIMENTAL support for printing of general
2368 ASN1 structures. This currently produces rather ugly output and doesn't
2369 handle several customised structures at all.
2372 *) Integrated support for PVK file format and some related formats such
2373 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2374 these in the 'rsa' and 'dsa' utilities.
2377 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2380 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2381 place for the (very old) "NETSCAPE" format certificates which are now
2382 handled using new ASN1 code equivalents.
2385 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2386 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2387 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2390 *) Modify CRL distribution points extension code to print out previously
2391 unsupported fields. Enhance extension setting code to allow setting of
2395 *) Add print and set support for Issuing Distribution Point CRL extension.
2398 *) Change 'Configure' script to enable Camellia by default.
2401 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2403 *) When rejecting SSL/TLS records due to an incorrect version number, never
2404 update s->server with a new major version number. As of
2405 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2406 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2407 the previous behavior could result in a read attempt at NULL when
2408 receiving specific incorrect SSL/TLS records once record payload
2409 protection is active. (CVE-2010-0740)
2410 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2412 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2413 could be crashed if the relevant tables were not present (e.g. chrooted).
2414 [Tomas Hoger <thoger@redhat.com>]
2416 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2418 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2419 [Martin Olsson, Neel Mehta]
2421 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2422 accommodate for stack sorting, always a write lock!).
2425 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2426 excessive delays in the RAND_poll(): over a minute. As a workaround
2427 include a time check in the inner Heap32Next loop too.
2430 *) The code that handled flushing of data in SSL/TLS originally used the
2431 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2432 the problem outlined in PR#1949. The fix suggested there however can
2433 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2434 of Apache). So instead simplify the code to flush unconditionally.
2435 This should be fine since flushing with no data to flush is a no op.
2438 *) Handle TLS versions 2.0 and later properly and correctly use the
2439 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2440 off ancient servers have a habit of sticking around for a while...
2443 *) Modify compression code so it frees up structures without using the
2444 ex_data callbacks. This works around a problem where some applications
2445 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2446 restarting) then use compression (e.g. SSL with compression) later.
2447 This results in significant per-connection memory leaks and
2448 has caused some security issues including CVE-2008-1678 and
2452 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2453 change when encrypting or decrypting.
2456 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2457 connect and renegotiate with servers which do not support RI.
2458 Until RI is more widely deployed this option is enabled by default.
2461 *) Add "missing" ssl ctrls to clear options and mode.
2464 *) If client attempts to renegotiate and doesn't support RI respond with
2465 a no_renegotiation alert as required by RFC5746. Some renegotiating
2466 TLS clients will continue a connection gracefully when they receive
2467 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2468 waiting for a server hello which it will never receive. Now we treat a
2469 received no_renegotiation alert as a fatal error. This is because
2470 applications requesting a renegotiation might well expect it to succeed
2471 and would have no code in place to handle the server denying it so the
2472 only safe thing to do is to terminate the connection.
2475 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2476 peer supports secure renegotiation and 0 otherwise. Print out peer
2477 renegotiation support in s_client/s_server.
2480 *) Replace the highly broken and deprecated SPKAC certification method with
2481 the updated NID creation version. This should correctly handle UTF8.
2484 *) Implement RFC5746. Re-enable renegotiation but require the extension
2485 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2486 turns out to be a bad idea. It has been replaced by
2487 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2488 SSL_CTX_set_options(). This is really not recommended unless you
2489 know what you are doing.
2490 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2492 *) Fixes to stateless session resumption handling. Use initial_ctx when
2493 issuing and attempting to decrypt tickets in case it has changed during
2494 servername handling. Use a non-zero length session ID when attempting
2495 stateless session resumption: this makes it possible to determine if
2496 a resumption has occurred immediately after receiving server hello
2497 (several places in OpenSSL subtly assume this) instead of later in
2501 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2502 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2503 fixes for a few places where the return code is not checked
2505 [Julia Lawall <julia@diku.dk>]
2507 *) Add --strict-warnings option to Configure script to include devteam
2508 warnings in other configurations.
2511 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2512 makes it possible to install openssl libraries in locations which
2513 have names other than "lib", for example "/usr/lib64" which some
2515 [Steve Henson, based on patch from Jeremy Utley]
2517 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2518 X690 8.9.12 and can produce some misleading textual output of OIDs.
2519 [Steve Henson, reported by Dan Kaminsky]
2521 *) Delete MD2 from algorithm tables. This follows the recommendation in
2522 several standards that it is not used in new applications due to
2523 several cryptographic weaknesses. For binary compatibility reasons
2524 the MD2 API is still compiled in by default.
2527 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2531 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2532 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2534 [Guenter <lists@gknw.net>]
2536 *) Fix the server certificate chain building code to use X509_verify_cert(),
2537 it used to have an ad-hoc builder which was unable to cope with anything
2538 other than a simple chain.
2539 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2541 *) Don't check self signed certificate signatures in X509_verify_cert()
2542 by default (a flag can override this): it just wastes time without
2543 adding any security. As a useful side effect self signed root CAs
2544 with non-FIPS digests are now usable in FIPS mode.
2547 *) In dtls1_process_out_of_seq_message() the check if the current message
2548 is already buffered was missing. For every new message was memory
2549 allocated, allowing an attacker to perform an denial of service attack
2550 with sending out of seq handshake messages until there is no memory
2551 left. Additionally every future messege was buffered, even if the
2552 sequence number made no sense and would be part of another handshake.
2553 So only messages with sequence numbers less than 10 in advance will be
2554 buffered. (CVE-2009-1378)
2555 [Robin Seggelmann, discovered by Daniel Mentz]
2557 *) Records are buffered if they arrive with a future epoch to be
2558 processed after finishing the corresponding handshake. There is
2559 currently no limitation to this buffer allowing an attacker to perform
2560 a DOS attack with sending records with future epochs until there is no
2561 memory left. This patch adds the pqueue_size() function to detemine
2562 the size of a buffer and limits the record buffer to 100 entries.
2564 [Robin Seggelmann, discovered by Daniel Mentz]
2566 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2567 parent structure is freed. (CVE-2009-1379)
2570 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2571 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2574 [Ilya O. <vrghost@gmail.com>]
2576 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2578 *) Disable renegotiation completely - this fixes a severe security
2579 problem (CVE-2009-3555) at the cost of breaking all
2580 renegotiation. Renegotiation can be re-enabled by setting
2581 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2582 run-time. This is really not recommended unless you know what
2586 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2588 *) Don't set val to NULL when freeing up structures, it is freed up by
2589 underlying code. If sizeof(void *) > sizeof(long) this can result in
2590 zeroing past the valid field. (CVE-2009-0789)
2591 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2593 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2594 checked correctly. This would allow some invalid signed attributes to
2595 appear to verify correctly. (CVE-2009-0591)
2596 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2598 *) Reject UniversalString and BMPString types with invalid lengths. This
2599 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2600 a legal length. (CVE-2009-0590)
2603 *) Set S/MIME signing as the default purpose rather than setting it
2604 unconditionally. This allows applications to override it at the store
2608 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2609 to handle some structures.
2612 *) Improve efficiency of mem_gets: don't search whole buffer each time
2614 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2616 *) New -hex option for openssl rand.
2619 *) Print out UTF8String and NumericString when parsing ASN1.
2622 *) Support NumericString type for name components.
2625 *) Allow CC in the environment to override the automatically chosen
2626 compiler. Note that nothing is done to ensure flags work with the
2630 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2632 *) Properly check EVP_VerifyFinal() and similar return values
2634 [Ben Laurie, Bodo Moeller, Google Security Team]
2636 *) Enable TLS extensions by default.
2639 *) Allow the CHIL engine to be loaded, whether the application is
2640 multithreaded or not. (This does not release the developer from the
2641 obligation to set up the dynamic locking callbacks.)
2642 [Sander Temme <sander@temme.net>]
2644 *) Use correct exit code if there is an error in dgst command.
2645 [Steve Henson; problem pointed out by Roland Dirlewanger]
2647 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2648 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2651 *) Add experimental JPAKE support, including demo authentication in
2652 s_client and s_server.
2655 *) Set the comparison function in v3_addr_canonize().
2656 [Rob Austein <sra@hactrn.net>]
2658 *) Add support for XMPP STARTTLS in s_client.
2659 [Philip Paeps <philip@freebsd.org>]
2661 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2662 to ensure that even with this option, only ciphersuites in the
2663 server's preference list will be accepted. (Note that the option
2664 applies only when resuming a session, so the earlier behavior was
2665 just about the algorithm choice for symmetric cryptography.)
2668 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2670 *) Fix NULL pointer dereference if a DTLS server received
2671 ChangeCipherSpec as first record (CVE-2009-1386).
2674 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2675 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2678 *) The fix in 0.9.8c that supposedly got rid of unsafe
2679 double-checked locking was incomplete for RSA blinding,
2680 addressing just one layer of what turns out to have been
2681 doubly unsafe triple-checked locking.
2683 So now fix this for real by retiring the MONT_HELPER macro
2684 in crypto/rsa/rsa_eay.c.
2686 [Bodo Moeller; problem pointed out by Marius Schilder]
2688 *) Various precautionary measures:
2690 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2692 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2693 (NB: This would require knowledge of the secret session ticket key
2694 to exploit, in which case you'd be SOL either way.)
2696 - Change bn_nist.c so that it will properly handle input BIGNUMs
2697 outside the expected range.
2699 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2702 [Neel Mehta, Bodo Moeller]
2704 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2705 the load fails. Useful for distros.
2706 [Ben Laurie and the FreeBSD team]
2708 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2711 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2714 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2716 This work was sponsored by Logica.
2719 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2720 keystores. Support for SSL/TLS client authentication too.
2721 Not compiled unless enable-capieng specified to Configure.
2723 This work was sponsored by Logica.
2726 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2727 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2728 attribute creation routines such as certifcate requests and PKCS#12
2732 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2734 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2735 handshake which could lead to a cilent crash as found using the
2736 Codenomicon TLS test suite (CVE-2008-1672)
2737 [Steve Henson, Mark Cox]
2739 *) Fix double free in TLS server name extensions which could lead to
2740 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2743 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2745 Clear the error queue to ensure that error entries left from
2746 older function calls do not interfere with the correct operation.
2747 [Lutz Jaenicke, Erik de Castro Lopo]
2749 *) Remove root CA certificates of commercial CAs:
2751 The OpenSSL project does not recommend any specific CA and does not
2752 have any policy with respect to including or excluding any CA.
2753 Therefore it does not make any sense to ship an arbitrary selection
2754 of root CA certificates with the OpenSSL software.
2757 *) RSA OAEP patches to fix two separate invalid memory reads.
2758 The first one involves inputs when 'lzero' is greater than
2759 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2760 before the beginning of from). The second one involves inputs where
2761 the 'db' section contains nothing but zeroes (there is a one-byte
2762 invalid read after the end of 'db').
2763 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2765 *) Partial backport from 0.9.9-dev:
2767 Introduce bn_mul_mont (dedicated Montgomery multiplication
2768 procedure) as a candidate for BIGNUM assembler implementation.
2769 While 0.9.9-dev uses assembler for various architectures, only
2770 x86_64 is available by default here in the 0.9.8 branch, and
2771 32-bit x86 is available through a compile-time setting.
2773 To try the 32-bit x86 assembler implementation, use Configure
2774 option "enable-montasm" (which exists only for this backport).
2776 As "enable-montasm" for 32-bit x86 disclaims code stability
2777 anyway, in this constellation we activate additional code
2778 backported from 0.9.9-dev for further performance improvements,
2779 namely BN_from_montgomery_word. (To enable this otherwise,
2780 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2782 [Andy Polyakov (backport partially by Bodo Moeller)]
2784 *) Add TLS session ticket callback. This allows an application to set
2785 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2786 values. This is useful for key rollover for example where several key
2787 sets may exist with different names.
2790 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2791 This was broken until now in 0.9.8 releases, such that the only way
2792 a registered ENGINE could be used (assuming it initialises
2793 successfully on the host) was to explicitly set it as the default
2794 for the relevant algorithms. This is in contradiction with 0.9.7
2795 behaviour and the documentation. With this fix, when an ENGINE is
2796 registered into a given algorithm's table of implementations, the
2797 'uptodate' flag is reset so that auto-discovery will be used next
2798 time a new context for that algorithm attempts to select an
2800 [Ian Lister (tweaked by Geoff Thorpe)]
2802 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2803 implemention in the following ways:
2805 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2808 Lack of BER streaming support means one pass streaming processing is
2809 only supported if data is detached: setting the streaming flag is
2810 ignored for embedded content.
2812 CMS support is disabled by default and must be explicitly enabled
2813 with the enable-cms configuration option.
2816 *) Update the GMP engine glue to do direct copies between BIGNUM and
2817 mpz_t when openssl and GMP use the same limb size. Otherwise the
2818 existing "conversion via a text string export" trick is still used.
2819 [Paul Sheer <paulsheer@gmail.com>]
2821 *) Zlib compression BIO. This is a filter BIO which compressed and
2822 uncompresses any data passed through it.
2825 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2826 RFC3394 compatible AES key wrapping.
2829 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2830 sets string data without copying. X509_ALGOR_set0() and
2831 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2832 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2833 from an X509_ATTRIBUTE structure optionally checking it occurs only
2834 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2838 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2839 to get the expected BN_FLG_CONSTTIME behavior.
2840 [Bodo Moeller (Google)]
2844 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2845 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2846 - added some more tests to do_tests.pl
2847 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2848 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2849 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2850 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2851 - various changes to netware.pl to enable gcc-cross builds on Win32
2853 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2854 - various changes to fix missing prototype warnings
2855 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2856 - added AES, WHIRLPOOL and CPUID assembler code to build files
2857 - added missing AES assembler make rules to mk1mf.pl
2858 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2859 [Guenter Knauf <eflash@gmx.net>]
2861 *) Implement certificate status request TLS extension defined in RFC3546.
2862 A client can set the appropriate parameters and receive the encoded
2863 OCSP response via a callback. A server can query the supplied parameters
2864 and set the encoded OCSP response in the callback. Add simplified examples
2865 to s_client and s_server.
2868 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2870 *) Fix various bugs:
2871 + Binary incompatibility of ssl_ctx_st structure
2872 + DTLS interoperation with non-compliant servers
2873 + Don't call get_session_cb() without proposed session
2874 + Fix ia64 assembler code
2875 [Andy Polyakov, Steve Henson]
2877 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2879 *) DTLS Handshake overhaul. There were longstanding issues with
2880 OpenSSL DTLS implementation, which were making it impossible for
2881 RFC 4347 compliant client to communicate with OpenSSL server.
2882 Unfortunately just fixing these incompatibilities would "cut off"
2883 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2884 server keeps tolerating non RFC compliant syntax. The opposite is
2885 not true, 0.9.8f client can not communicate with earlier server.
2886 This update even addresses CVE-2007-4995.
2889 *) Changes to avoid need for function casts in OpenSSL: some compilers
2890 (gcc 4.2 and later) reject their use.
2891 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2894 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2895 RFC4507bis. The encrypted ticket format is an encrypted encoded
2896 SSL_SESSION structure, that way new session features are automatically
2899 If a client application caches session in an SSL_SESSION structure
2900 support is transparent because tickets are now stored in the encoded
2903 The SSL_CTX structure automatically generates keys for ticket
2904 protection in servers so again support should be possible
2905 with no application modification.
2907 If a client or server wishes to disable RFC4507 support then the option
2908 SSL_OP_NO_TICKET can be set.
2910 Add a TLS extension debugging callback to allow the contents of any client
2911 or server extensions to be examined.
2913 This work was sponsored by Google.
2916 *) Add initial support for TLS extensions, specifically for the server_name
2917 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2918 have new members for a host name. The SSL data structure has an
2919 additional member SSL_CTX *initial_ctx so that new sessions can be
2920 stored in that context to allow for session resumption, even after the
2921 SSL has been switched to a new SSL_CTX in reaction to a client's
2922 server_name extension.
2924 New functions (subject to change):
2926 SSL_get_servername()
2927 SSL_get_servername_type()
2930 New CTRL codes and macros (subject to change):
2932 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2933 - SSL_CTX_set_tlsext_servername_callback()
2934 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2935 - SSL_CTX_set_tlsext_servername_arg()
2936 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2938 openssl s_client has a new '-servername ...' option.
2940 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2941 '-key2 ...', '-servername_fatal' (subject to change). This allows
2942 testing the HostName extension for a specific single host name ('-cert'
2943 and '-key' remain fallbacks for handshakes without HostName
2944 negotiation). If the unrecogninzed_name alert has to be sent, this by
2945 default is a warning; it becomes fatal with the '-servername_fatal'
2948 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2950 *) Add AES and SSE2 assembly language support to VC++ build.
2953 *) Mitigate attack on final subtraction in Montgomery reduction.
2956 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2957 (which previously caused an internal error).
2960 *) Squeeze another 10% out of IGE mode when in != out.
2963 *) AES IGE mode speedup.
2964 [Dean Gaudet (Google)]
2966 *) Add the Korean symmetric 128-bit cipher SEED (see
2967 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2968 add SEED ciphersuites from RFC 4162:
2970 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2971 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2972 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2973 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2975 To minimize changes between patchlevels in the OpenSSL 0.9.8
2976 series, SEED remains excluded from compilation unless OpenSSL
2977 is configured with 'enable-seed'.
2978 [KISA, Bodo Moeller]
2980 *) Mitigate branch prediction attacks, which can be practical if a
2981 single processor is shared, allowing a spy process to extract
2982 information. For detailed background information, see
2983 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2984 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2985 and Necessary Software Countermeasures"). The core of the change
2986 are new versions BN_div_no_branch() and
2987 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2988 respectively, which are slower, but avoid the security-relevant
2989 conditional branches. These are automatically called by BN_div()
2990 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2991 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2992 remove a conditional branch.
2994 BN_FLG_CONSTTIME is the new name for the previous
2995 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2996 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2997 in the exponent causes BN_mod_exp_mont() to use the alternative
2998 implementation in BN_mod_exp_mont_consttime().) The old name
2999 remains as a deprecated alias.
3001 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3002 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3003 constant-time implementations for more than just exponentiation.
3004 Here too the old name is kept as a deprecated alias.
3006 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3007 the BN_BLINDING structure gets an independent copy of the
3008 modulus. This means that the previous "BIGNUM *m" argument to
3009 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3010 essentially becomes "const BIGNUM *m", although we can't actually
3011 change this in the header file before 0.9.9. It allows
3012 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3013 enable BN_FLG_CONSTTIME.
3015 [Matthew D Wood (Intel Corp)]
3017 *) In the SSL/TLS server implementation, be strict about session ID
3018 context matching (which matters if an application uses a single
3019 external cache for different purposes). Previously,
3020 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3021 set. This did ensure strict client verification, but meant that,
3022 with applications using a single external cache for quite
3023 different requirements, clients could circumvent ciphersuite
3024 restrictions for a given session ID context by starting a session
3025 in a different context.
3028 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3029 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3030 authentication-only ciphersuites.
3033 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3034 not complete and could lead to a possible single byte overflow
3035 (CVE-2007-5135) [Ben Laurie]
3037 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3039 *) Since AES128 and AES256 (and similarly Camellia128 and
3040 Camellia256) share a single mask bit in the logic of
3041 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3042 kludge to work properly if AES128 is available and AES256 isn't
3043 (or if Camellia128 is available and Camellia256 isn't).
3046 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3047 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3048 When a point or a seed is encoded in a BIT STRING, we need to
3049 prevent the removal of trailing zero bits to get the proper DER
3050 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3051 of a NamedBitList, for which trailing 0 bits need to be removed.)
3054 *) Have SSL/TLS server implementation tolerate "mismatched" record
3055 protocol version while receiving ClientHello even if the
3056 ClientHello is fragmented. (The server can't insist on the
3057 particular protocol version it has chosen before the ServerHello
3058 message has informed the client about his choice.)
3061 *) Add RFC 3779 support.
3062 [Rob Austein for ARIN, Ben Laurie]
3064 *) Load error codes if they are not already present instead of using a
3065 static variable. This allows them to be cleanly unloaded and reloaded.
3066 Improve header file function name parsing.
3069 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3070 or CAPABILITY handshake as required by RFCs.
3073 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3075 *) Introduce limits to prevent malicious keys being able to
3076 cause a denial of service. (CVE-2006-2940)
3077 [Steve Henson, Bodo Moeller]
3079 *) Fix ASN.1 parsing of certain invalid structures that can result
3080 in a denial of service. (CVE-2006-2937) [Steve Henson]
3082 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3083 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3085 *) Fix SSL client code which could crash if connecting to a
3086 malicious SSLv2 server. (CVE-2006-4343)
3087 [Tavis Ormandy and Will Drewry, Google Security Team]
3089 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3090 match only those. Before that, "AES256-SHA" would be interpreted
3091 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3092 the same strength classification in 0.9.7h) as we currently only
3093 have a single AES bit in the ciphersuite description bitmap.
3094 That change, however, also applied to ciphersuite strings such as
3095 "RC4-MD5" that intentionally matched multiple ciphersuites --
3096 namely, SSL 2.0 ciphersuites in addition to the more common ones
3097 from SSL 3.0/TLS 1.0.
3099 So we change the selection algorithm again: Naming an explicit
3100 ciphersuite selects this one ciphersuite, and any other similar
3101 ciphersuite (same bitmap) from *other* protocol versions.
3102 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3103 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3105 Since SSL 2.0 does not have any ciphersuites for which the
3106 128/256 bit distinction would be relevant, this works for now.
3107 The proper fix will be to use different bits for AES128 and
3108 AES256, which would have avoided the problems from the beginning;
3109 however, bits are scarce, so we can only do this in a new release
3110 (not just a patchlevel) when we can change the SSL_CIPHER
3111 definition to split the single 'unsigned long mask' bitmap into
3112 multiple values to extend the available space.
3116 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3118 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3119 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3121 *) Add AES IGE and biIGE modes.
3124 *) Change the Unix randomness entropy gathering to use poll() when
3125 possible instead of select(), since the latter has some
3126 undesirable limitations.
3127 [Darryl Miles via Richard Levitte and Bodo Moeller]
3129 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3130 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3131 cannot be implicitly activated as part of, e.g., the "AES" alias.
3132 However, please upgrade to OpenSSL 0.9.9[-dev] for
3133 non-experimental use of the ECC ciphersuites to get TLS extension
3134 support, which is required for curve and point format negotiation
3135 to avoid potential handshake problems.
3138 *) Disable rogue ciphersuites:
3140 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3141 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3142 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3144 The latter two were purportedly from
3145 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3148 Also deactivate the remaining ciphersuites from
3149 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3150 unofficial, and the ID has long expired.
3153 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3154 dual-core machines) and other potential thread-safety issues.
3157 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3158 versions), which is now available for royalty-free use
3159 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3160 Also, add Camellia TLS ciphersuites from RFC 4132.
3162 To minimize changes between patchlevels in the OpenSSL 0.9.8
3163 series, Camellia remains excluded from compilation unless OpenSSL
3164 is configured with 'enable-camellia'.
3167 *) Disable the padding bug check when compression is in use. The padding
3168 bug check assumes the first packet is of even length, this is not
3169 necessarily true if compresssion is enabled and can result in false
3170 positives causing handshake failure. The actual bug test is ancient
3171 code so it is hoped that implementations will either have fixed it by
3172 now or any which still have the bug do not support compression.
3175 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3177 *) When applying a cipher rule check to see if string match is an explicit
3178 cipher suite and only match that one cipher suite if it is.
3181 *) Link in manifests for VC++ if needed.
3182 [Austin Ziegler <halostatue@gmail.com>]
3184 *) Update support for ECC-based TLS ciphersuites according to
3185 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3186 TLS extensions, which are supported starting with the 0.9.9
3187 branch, not in the OpenSSL 0.9.8 branch).
3190 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3191 opaque EVP_CIPHER_CTX handling.
3194 *) Fixes and enhancements to zlib compression code. We now only use
3195 "zlib1.dll" and use the default __cdecl calling convention on Win32
3196 to conform with the standards mentioned here:
3197 http://www.zlib.net/DLL_FAQ.txt
3198 Static zlib linking now works on Windows and the new --with-zlib-include
3199 --with-zlib-lib options to Configure can be used to supply the location
3200 of the headers and library. Gracefully handle case where zlib library
3204 *) Several fixes and enhancements to the OID generation code. The old code
3205 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3206 handle numbers larger than ULONG_MAX, truncated printing and had a
3207 non standard OBJ_obj2txt() behaviour.
3210 *) Add support for building of engines under engine/ as shared libraries
3211 under VC++ build system.
3214 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3215 Hopefully, we will not see any false combination of paths any more.
3218 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3220 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3221 (part of SSL_OP_ALL). This option used to disable the
3222 countermeasure against man-in-the-middle protocol-version
3223 rollback in the SSL 2.0 server implementation, which is a bad
3224 idea. (CVE-2005-2969)
3226 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3227 for Information Security, National Institute of Advanced Industrial
3228 Science and Technology [AIST], Japan)]
3230 *) Add two function to clear and return the verify parameter flags.
3233 *) Keep cipherlists sorted in the source instead of sorting them at
3234 runtime, thus removing the need for a lock.
3237 *) Avoid some small subgroup attacks in Diffie-Hellman.
3238 [Nick Mathewson and Ben Laurie]
3240 *) Add functions for well-known primes.
3243 *) Extended Windows CE support.
3244 [Satoshi Nakamura and Andy Polyakov]
3246 *) Initialize SSL_METHOD structures at compile time instead of during
3247 runtime, thus removing the need for a lock.
3250 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3251 attempting to decrypt each encrypted key in turn. Add support to
3255 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3257 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3260 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3263 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3264 key into the same file any more.
3267 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3270 *) Add -utf8 command line and config file option to 'ca'.
3271 [Stefan <stf@udoma.org]
3273 *) Removed the macro des_crypt(), as it seems to conflict with some
3274 libraries. Use DES_crypt().
3277 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3278 involves renaming the source and generated shared-libs for
3279 both. The engines will accept the corrected or legacy ids
3280 ('ncipher' and '4758_cca' respectively) when binding. NB,
3281 this only applies when building 'shared'.
3282 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3284 *) Add attribute functions to EVP_PKEY structure. Modify
3285 PKCS12_create() to recognize a CSP name attribute and
3286 use it. Make -CSP option work again in pkcs12 utility.
3289 *) Add new functionality to the bn blinding code:
3290 - automatic re-creation of the BN_BLINDING parameters after
3291 a fixed number of uses (currently 32)
3292 - add new function for parameter creation
3293 - introduce flags to control the update behaviour of the
3294 BN_BLINDING parameters
3295 - hide BN_BLINDING structure
3296 Add a second BN_BLINDING slot to the RSA structure to improve
3297 performance when a single RSA object is shared among several
3301 *) Add support for DTLS.
3302 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3304 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3305 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3308 *) Remove buggy and incompletet DH cert support from
3309 ssl/ssl_rsa.c and ssl/s3_both.c
3312 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3313 the apps/openssl applications.
3316 *) Compile clean with "-Wall -Wmissing-prototypes
3317 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3318 DEBUG_SAFESTACK must also be set.
3321 *) Change ./Configure so that certain algorithms can be disabled by default.
3322 The new counterpiece to "no-xxx" is "enable-xxx".
3324 The patented RC5 and MDC2 algorithms will now be disabled unless
3325 "enable-rc5" and "enable-mdc2", respectively, are specified.
3327 (IDEA remains enabled despite being patented. This is because IDEA
3328 is frequently required for interoperability, and there is no license
3329 fee for non-commercial use. As before, "no-idea" can be used to
3330 avoid this algorithm.)
3334 *) Add processing of proxy certificates (see RFC 3820). This work was
3335 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3336 EGEE (Enabling Grids for E-science in Europe).
3339 *) RC4 performance overhaul on modern architectures/implementations, such
3340 as Intel P4, IA-64 and AMD64.
3343 *) New utility extract-section.pl. This can be used specify an alternative
3344 section number in a pod file instead of having to treat each file as
3345 a separate case in Makefile. This can be done by adding two lines to the
3348 =for comment openssl_section:XXX
3350 The blank line is mandatory.
3354 *) New arguments -certform, -keyform and -pass for s_client and s_server
3355 to allow alternative format key and certificate files and passphrase
3359 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3360 update associated structures and add various utility functions.
3362 Add new policy related verify parameters, include policy checking in
3363 standard verify code. Enhance 'smime' application with extra parameters
3364 to support policy checking and print out.
3367 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3368 Nehemiah processors. These extensions support AES encryption in hardware
3369 as well as RNG (though RNG support is currently disabled).
3370 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3372 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3375 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3376 [Andy Polyakov and a number of other people]
3378 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3379 implementation contributed by IBM.
3380 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3382 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3383 exponent rather than 'unsigned long'. There is a corresponding change to
3384 the new 'rsa_keygen' element of the RSA_METHOD structure.
3385 [Jelte Jansen, Geoff Thorpe]
3387 *) Functionality for creating the initial serial number file is now
3388 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3390 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3391 number file to 1, which is bound to cause problems. To avoid
3392 the problems while respecting compatibility between different 0.9.7
3393 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3394 CA.pl for serial number initialization. With the new release 0.9.8,
3395 we can fix the problem directly in the 'ca' utility.)
3398 *) Reduced header interdepencies by declaring more opaque objects in
3399 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3400 give fewer recursive includes, which could break lazy source code - so
3401 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3402 developers should define this symbol when building and using openssl to
3403 ensure they track the recommended behaviour, interfaces, [etc], but
3404 backwards-compatible behaviour prevails when this isn't defined.
3407 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3410 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3411 This will generate a random key of the appropriate length based on the
3412 cipher context. The EVP_CIPHER can provide its own random key generation
3413 routine to support keys of a specific form. This is used in the des and
3414 3des routines to generate a key of the correct parity. Update S/MIME
3415 code to use new functions and hence generate correct parity DES keys.
3416 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3417 valid (weak or incorrect parity).
3420 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3421 as looking them up. This is useful when the verified structure may contain
3422 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3423 present unless the new PKCS7_NO_CRL flag is asserted.
3426 *) Extend ASN1 oid configuration module. It now additionally accepts the
3429 shortName = some long name, 1.2.3.4
3432 *) Reimplemented the BN_CTX implementation. There is now no more static
3433 limitation on the number of variables it can handle nor the depth of the
3434 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3435 information can now expand as required, and rather than having a single
3436 static array of bignums, BN_CTX now uses a linked-list of such arrays
3437 allowing it to expand on demand whilst maintaining the usefulness of
3438 BN_CTX's "bundling".
3441 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3442 to allow all RSA operations to function using a single BN_CTX.
3445 *) Preliminary support for certificate policy evaluation and checking. This
3446 is initially intended to pass the tests outlined in "Conformance Testing
3447 of Relying Party Client Certificate Path Processing Logic" v1.07.
3450 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3451 remained unused and not that useful. A variety of other little bignum
3452 tweaks and fixes have also been made continuing on from the audit (see
3456 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3457 associated ASN1, EVP and SSL functions and old ASN1 macros.
3460 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3461 and this should never fail. So the return value from the use of
3462 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3463 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3466 *) BN_CTX_get() should return zero-valued bignums, providing the same
3467 initialised value as BN_new().
3468 [Geoff Thorpe, suggested by Ulf Möller]
3470 *) Support for inhibitAnyPolicy certificate extension.
3473 *) An audit of the BIGNUM code is underway, for which debugging code is
3474 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3475 is considered valid when processing BIGNUMs, and causes execution to
3476 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3477 further steps are taken to deliberately pollute unused data in BIGNUM
3478 structures to try and expose faulty code further on. For now, openssl will
3479 (in its default mode of operation) continue to tolerate the inconsistent
3480 forms that it has tolerated in the past, but authors and packagers should
3481 consider trying openssl and their own applications when compiled with
3482 these debugging symbols defined. It will help highlight potential bugs in
3483 their own code, and will improve the test coverage for OpenSSL itself. At
3484 some point, these tighter rules will become openssl's default to improve
3485 maintainability, though the assert()s and other overheads will remain only
3486 in debugging configurations. See bn.h for more details.
3487 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3489 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3490 that can only be obtained through BN_CTX_new() (which implicitly
3491 initialises it). The presence of this function only made it possible
3492 to overwrite an existing structure (and cause memory leaks).
3495 *) Because of the callback-based approach for implementing LHASH as a
3496 template type, lh_insert() adds opaque objects to hash-tables and
3497 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3498 to clean up those corresponding objects before destroying the hash table
3499 (and losing the object pointers). So some over-zealous constifications in
3500 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3501 objects as "const" and the lh_doall[_arg] callback wrappers are not
3502 prototyped to have "const" restrictions on the object pointers they are
3503 given (and so aren't required to cast them away any more).
3506 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3507 (speed) prefers to use its own implementation. The two implementations
3508 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3509 its object type properly exposed (MS_TM) instead of casting to/from "char
3510 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3511 aren't necessarily the greatest nomenclatures - but this is what was used
3512 internally to the implementation so I've used that for now.
3515 *) Ensure that deprecated functions do not get compiled when
3516 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3517 the self-tests were still using deprecated key-generation functions so
3518 these have been updated also.
3521 *) Reorganise PKCS#7 code to separate the digest location functionality
3522 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3523 New function PKCS7_set_digest() to set the digest type for PKCS#7
3524 digestedData type. Add additional code to correctly generate the
3525 digestedData type and add support for this type in PKCS7 initialization
3529 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3530 structure of type "other".
3533 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3534 sure the loop does correctly stop and breaking ("division by zero")
3535 modulus operations are not performed. The (pre-generated) prime
3536 table crypto/bn/bn_prime.h was already correct, but it could not be
3537 re-generated on some platforms because of the "division by zero"
3538 situation in the script.
3539 [Ralf S. Engelschall]
3541 *) Update support for ECC-based TLS ciphersuites according to
3542 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3543 SHA-1 now is only used for "small" curves (where the
3544 representation of a field element takes up to 24 bytes); for
3545 larger curves, the field element resulting from ECDH is directly
3546 used as premaster secret.
3547 [Douglas Stebila (Sun Microsystems Laboratories)]
3549 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3550 curve secp160r1 to the tests.
3551 [Douglas Stebila (Sun Microsystems Laboratories)]
3553 *) Add the possibility to load symbols globally with DSO.
3554 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3556 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3557 control of the error stack.
3560 *) Add support for STORE in ENGINE.
3563 *) Add the STORE type. The intention is to provide a common interface
3564 to certificate and key stores, be they simple file-based stores, or
3565 HSM-type store, or LDAP stores, or...
3566 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3569 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3570 pass a list of arguments to any function as well as provide a way
3571 for a function to pass data back to the caller.
3574 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3575 works like BUF_strdup() but can be used to duplicate a portion of
3576 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3580 *) Add the function sk_find_ex() which works like sk_find(), but will
3581 return an index to an element even if an exact match couldn't be
3582 found. The index is guaranteed to point at the element where the
3583 searched-for key would be inserted to preserve sorting order.
3586 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3587 takes an extra flags argument for optional functionality. Currently,
3588 the following flags are defined:
3590 OBJ_BSEARCH_VALUE_ON_NOMATCH
3591 This one gets OBJ_bsearch_ex() to return a pointer to the first
3592 element where the comparing function returns a negative or zero
3595 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3596 This one gets OBJ_bsearch_ex() to return a pointer to the first
3597 element where the comparing function returns zero. This is useful
3598 if there are more than one element where the comparing function
3602 *) Make it possible to create self-signed certificates with 'openssl ca'
3603 in such a way that the self-signed certificate becomes part of the
3604 CA database and uses the same mechanisms for serial number generation
3605 as all other certificate signing. The new flag '-selfsign' enables
3606 this functionality. Adapt CA.sh and CA.pl.in.
3609 *) Add functionality to check the public key of a certificate request
3610 against a given private. This is useful to check that a certificate
3611 request can be signed by that key (self-signing).
3614 *) Make it possible to have multiple active certificates with the same
3615 subject in the CA index file. This is done only if the keyword
3616 'unique_subject' is set to 'no' in the main CA section (default
3617 if 'CA_default') of the configuration file. The value is saved
3618 with the database itself in a separate index attribute file,
3619 named like the index file with '.attr' appended to the name.
3622 *) Generate muti valued AVAs using '+' notation in config files for
3626 *) Support for nameConstraints certificate extension.
3629 *) Support for policyConstraints certificate extension.
3632 *) Support for policyMappings certificate extension.
3635 *) Make sure the default DSA_METHOD implementation only uses its
3636 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3637 and change its own handlers to be NULL so as to remove unnecessary
3638 indirection. This lets alternative implementations fallback to the
3639 default implementation more easily.
3642 *) Support for directoryName in GeneralName related extensions
3646 *) Make it possible to link applications using Makefile.shared.
3647 Make that possible even when linking against static libraries!
3650 *) Support for single pass processing for S/MIME signing. This now
3651 means that S/MIME signing can be done from a pipe, in addition
3652 cleartext signing (multipart/signed type) is effectively streaming
3653 and the signed data does not need to be all held in memory.
3655 This is done with a new flag PKCS7_STREAM. When this flag is set
3656 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3657 is done after the data is output (and digests calculated) in
3658 SMIME_write_PKCS7().
3661 *) Add full support for -rpath/-R, both in shared libraries and
3662 applications, at least on the platforms where it's known how
3666 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3667 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3668 will now compute a table of multiples of the generator that
3669 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3670 faster (notably in the case of a single point multiplication,
3671 scalar * generator).
3672 [Nils Larsch, Bodo Moeller]
3674 *) IPv6 support for certificate extensions. The various extensions
3675 which use the IP:a.b.c.d can now take IPv6 addresses using the
3676 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3680 *) Added an ENGINE that implements RSA by performing private key
3681 exponentiations with the GMP library. The conversions to and from
3682 GMP's mpz_t format aren't optimised nor are any montgomery forms
3683 cached, and on x86 it appears OpenSSL's own performance has caught up.
3684 However there are likely to be other architectures where GMP could
3685 provide a boost. This ENGINE is not built in by default, but it can be
3686 specified at Configure time and should be accompanied by the necessary
3687 linker additions, eg;
3688 ./config -DOPENSSL_USE_GMP -lgmp
3691 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3692 testing availability of engines with "-t" - the old behaviour is
3693 produced by increasing the feature's verbosity with "-tt".
3696 *) ECDSA routines: under certain error conditions uninitialized BN objects
3697 could be freed. Solution: make sure initialization is performed early
3698 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3702 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3703 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3704 software implementations. For DSA and DH, parameter generation can
3705 also be overriden by providing the appropriate method callbacks.
3708 *) Change the "progress" mechanism used in key-generation and
3709 primality testing to functions that take a new BN_GENCB pointer in
3710 place of callback/argument pairs. The new API functions have "_ex"
3711 postfixes and the older functions are reimplemented as wrappers for
3712 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3713 declarations of the old functions to help (graceful) attempts to
3714 migrate to the new functions. Also, the new key-generation API
3715 functions operate on a caller-supplied key-structure and return
3716 success/failure rather than returning a key or NULL - this is to
3717 help make "keygen" another member function of RSA_METHOD etc.
3719 Example for using the new callback interface:
3721 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3725 BN_GENCB_set(&my_cb, my_callback, my_arg);
3727 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3728 /* For the meaning of a, b in calls to my_callback(), see the
3729 * documentation of the function that calls the callback.
3730 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3731 * my_callback should return 1 if it wants BN_is_prime_ex()
3732 * to continue, or 0 to stop.
3737 *) Change the ZLIB compression method to be stateful, and make it
3738 available to TLS with the number defined in
3739 draft-ietf-tls-compression-04.txt.
3742 *) Add the ASN.1 structures and functions for CertificatePair, which
3743 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3745 CertificatePair ::= SEQUENCE {
3746 forward [0] Certificate OPTIONAL,
3747 reverse [1] Certificate OPTIONAL,
3748 -- at least one of the pair shall be present -- }
3750 Also implement the PEM functions to read and write certificate
3751 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3753 This needed to be defined, mostly for the sake of the LDAP
3754 attribute crossCertificatePair, but may prove useful elsewhere as
3758 *) Make it possible to inhibit symlinking of shared libraries in
3759 Makefile.shared, for Cygwin's sake.
3762 *) Extend the BIGNUM API by creating a function
3763 void BN_set_negative(BIGNUM *a, int neg);
3764 and a macro that behave like
3765 int BN_is_negative(const BIGNUM *a);
3767 to avoid the need to access 'a->neg' directly in applications.
3770 *) Implement fast modular reduction for pseudo-Mersenne primes
3771 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3772 EC_GROUP_new_curve_GFp() will now automatically use this
3774 [Nils Larsch <nla@trustcenter.de>]
3776 *) Add new lock type (CRYPTO_LOCK_BN).
3779 *) Change the ENGINE framework to automatically load engines
3780 dynamically from specific directories unless they could be
3781 found to already be built in or loaded. Move all the
3782 current engines except for the cryptodev one to a new
3784 The engines in engines/ are built as shared libraries if
3785 the "shared" options was given to ./Configure or ./config.
3786 Otherwise, they are inserted in libcrypto.a.
3787 /usr/local/ssl/engines is the default directory for dynamic
3788 engines, but that can be overriden at configure time through
3789 the usual use of --prefix and/or --openssldir, and at run
3790 time with the environment variable OPENSSL_ENGINES.
3791 [Geoff Thorpe and Richard Levitte]
3793 *) Add Makefile.shared, a helper makefile to build shared
3794 libraries. Addapt Makefile.org.
3797 *) Add version info to Win32 DLLs.
3798 [Peter 'Luna' Runestig" <peter@runestig.com>]
3800 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3801 can be added using this API to created arbitrary PKCS#12
3802 files while avoiding the low level API.
3804 New options to PKCS12_create(), key or cert can be NULL and
3805 will then be omitted from the output file. The encryption
3806 algorithm NIDs can be set to -1 for no encryption, the mac
3807 iteration count can be set to 0 to omit the mac.
3809 Enhance pkcs12 utility by making the -nokeys and -nocerts
3810 options work when creating a PKCS#12 file. New option -nomac
3811 to omit the mac, NONE can be set for an encryption algorithm.
3812 New code is modified to use the enhanced PKCS12_create()
3813 instead of the low level API.
3816 *) Extend ASN1 encoder to support indefinite length constructed
3817 encoding. This can output sequences tags and octet strings in
3818 this form. Modify pk7_asn1.c to support indefinite length
3819 encoding. This is experimental and needs additional code to
3820 be useful, such as an ASN1 bio and some enhanced streaming
3823 Extend template encode functionality so that tagging is passed
3824 down to the template encoder.
3827 *) Let 'openssl req' fail if an argument to '-newkey' is not
3828 recognized instead of using RSA as a default.
3831 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3832 As these are not official, they are not included in "ALL";
3833 the "ECCdraft" ciphersuite group alias can be used to select them.
3834 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3836 *) Add ECDH engine support.
3837 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3839 *) Add ECDH in new directory crypto/ecdh/.
3840 [Douglas Stebila (Sun Microsystems Laboratories)]
3842 *) Let BN_rand_range() abort with an error after 100 iterations
3843 without success (which indicates a broken PRNG).
3846 *) Change BN_mod_sqrt() so that it verifies that the input value
3847 is really the square of the return value. (Previously,
3848 BN_mod_sqrt would show GIGO behaviour.)
3851 *) Add named elliptic curves over binary fields from X9.62, SECG,
3852 and WAP/WTLS; add OIDs that were still missing.
3854 [Sheueling Chang Shantz and Douglas Stebila
3855 (Sun Microsystems Laboratories)]
3857 *) Extend the EC library for elliptic curves over binary fields
3858 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3861 EC_GF2m_simple_method
3865 EC_GROUP_new_curve_GF2m
3866 EC_GROUP_set_curve_GF2m
3867 EC_GROUP_get_curve_GF2m
3868 EC_POINT_set_affine_coordinates_GF2m
3869 EC_POINT_get_affine_coordinates_GF2m
3870 EC_POINT_set_compressed_coordinates_GF2m
3872 Point compression for binary fields is disabled by default for
3873 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3876 As binary polynomials are represented as BIGNUMs, various members
3877 of the EC_GROUP and EC_POINT data structures can be shared
3878 between the implementations for prime fields and binary fields;
3879 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3880 are essentially identical to their ..._GFp counterparts.
3881 (For simplicity, the '..._GFp' prefix has been dropped from
3882 various internal method names.)
3884 An internal 'field_div' method (similar to 'field_mul' and
3885 'field_sqr') has been added; this is used only for binary fields.
3887 [Sheueling Chang Shantz and Douglas Stebila
3888 (Sun Microsystems Laboratories)]
3890 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3891 through methods ('mul', 'precompute_mult').
3893 The generic implementations (now internally called 'ec_wNAF_mul'
3894 and 'ec_wNAF_precomputed_mult') remain the default if these
3895 methods are undefined.
3897 [Sheueling Chang Shantz and Douglas Stebila
3898 (Sun Microsystems Laboratories)]
3900 *) New function EC_GROUP_get_degree, which is defined through
3901 EC_METHOD. For curves over prime fields, this returns the bit
3902 length of the modulus.
3904 [Sheueling Chang Shantz and Douglas Stebila
3905 (Sun Microsystems Laboratories)]
3907 *) New functions EC_GROUP_dup, EC_POINT_dup.
3908 (These simply call ..._new and ..._copy).
3910 [Sheueling Chang Shantz and Douglas Stebila
3911 (Sun Microsystems Laboratories)]
3913 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3914 Polynomials are represented as BIGNUMs (where the sign bit is not
3915 used) in the following functions [macros]:
3918 BN_GF2m_sub [= BN_GF2m_add]
3919 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3920 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3921 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3923 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3924 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3925 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3926 BN_GF2m_cmp [= BN_ucmp]
3928 (Note that only the 'mod' functions are actually for fields GF(2^m).
3929 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3931 For some functions, an the irreducible polynomial defining a
3932 field can be given as an 'unsigned int[]' with strictly
3933 decreasing elements giving the indices of those bits that are set;
3934 i.e., p[] represents the polynomial
3935 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3937 p[0] > p[1] > ... > p[k] = 0.
3938 This applies to the following functions:
3943 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3944 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3946 BN_GF2m_mod_sqrt_arr
3947 BN_GF2m_mod_solve_quad_arr
3951 Conversion can be performed by the following functions:
3956 bntest.c has additional tests for binary polynomial arithmetic.
3958 Two implementations for BN_GF2m_mod_div() are available.
3959 The default algorithm simply uses BN_GF2m_mod_inv() and
3960 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3961 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3962 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3964 [Sheueling Chang Shantz and Douglas Stebila
3965 (Sun Microsystems Laboratories)]
3967 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3968 functionality is disabled at compile-time.
3969 [Douglas Stebila <douglas.stebila@sun.com>]
3971 *) Change default behaviour of 'openssl asn1parse' so that more
3972 information is visible when viewing, e.g., a certificate:
3974 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3975 mode the content of non-printable OCTET STRINGs is output in a
3976 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3977 avoid the appearance of a printable string.
3978 [Nils Larsch <nla@trustcenter.de>]
3980 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3982 EC_GROUP_set_asn1_flag()
3983 EC_GROUP_get_asn1_flag()
3984 EC_GROUP_set_point_conversion_form()
3985 EC_GROUP_get_point_conversion_form()
3986 These control ASN1 encoding details:
3987 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3988 has been set to OPENSSL_EC_NAMED_CURVE.
3989 - Points are encoded in uncompressed form by default; options for
3990 asn1_for are as for point2oct, namely
3991 POINT_CONVERSION_COMPRESSED
3992 POINT_CONVERSION_UNCOMPRESSED
3993 POINT_CONVERSION_HYBRID
3995 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3998 EC_GROUP_get0_seed()
3999 EC_GROUP_get_seed_len()
4000 This is used only for ASN1 purposes (so far).
4001 [Nils Larsch <nla@trustcenter.de>]
4003 *) Add 'field_type' member to EC_METHOD, which holds the NID
4004 of the appropriate field type OID. The new function
4005 EC_METHOD_get_field_type() returns this value.
4006 [Nils Larsch <nla@trustcenter.de>]
4011 EC_POINT_point2hex()
4012 EC_POINT_hex2point()
4013 providing useful interfaces to EC_POINT_point2oct() and
4014 EC_POINT_oct2point().
4015 [Nils Larsch <nla@trustcenter.de>]
4017 *) Change internals of the EC library so that the functions
4018 EC_GROUP_set_generator()
4019 EC_GROUP_get_generator()
4020 EC_GROUP_get_order()
4021 EC_GROUP_get_cofactor()
4022 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4023 to methods, which would lead to unnecessary code duplication when
4024 adding different types of curves.
4025 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4027 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4028 arithmetic, and such that modified wNAFs are generated
4029 (which avoid length expansion in many cases).
4032 *) Add a function EC_GROUP_check_discriminant() (defined via
4033 EC_METHOD) that verifies that the curve discriminant is non-zero.
4035 Add a function EC_GROUP_check() that makes some sanity tests
4036 on a EC_GROUP, its generator and order. This includes
4037 EC_GROUP_check_discriminant().
4038 [Nils Larsch <nla@trustcenter.de>]
4040 *) Add ECDSA in new directory crypto/ecdsa/.
4042 Add applications 'openssl ecparam' and 'openssl ecdsa'
4043 (these are based on 'openssl dsaparam' and 'openssl dsa').
4045 ECDSA support is also included in various other files across the
4046 library. Most notably,
4047 - 'openssl req' now has a '-newkey ecdsa:file' option;
4048 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4049 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4050 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4051 them suitable for ECDSA where domain parameters must be
4052 extracted before the specific public key;
4053 - ECDSA engine support has been added.
4054 [Nils Larsch <nla@trustcenter.de>]
4056 *) Include some named elliptic curves, and add OIDs from X9.62,
4057 SECG, and WAP/WTLS. Each curve can be obtained from the new
4059 EC_GROUP_new_by_curve_name(),
4060 and the list of available named curves can be obtained with
4061 EC_get_builtin_curves().
4062 Also add a 'curve_name' member to EC_GROUP objects, which can be
4064 EC_GROUP_set_curve_name()
4065 EC_GROUP_get_curve_name()
4066 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4068 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4069 was actually never needed) and in BN_mul(). The removal in BN_mul()
4070 required a small change in bn_mul_part_recursive() and the addition
4071 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4072 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4073 bn_sub_words() and bn_add_words() except they take arrays with
4077 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4079 *) Cleanse PEM buffers before freeing them since they may contain
4081 [Benjamin Bennett <ben@psc.edu>]
4083 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4084 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4085 authentication-only ciphersuites.
4088 *) Since AES128 and AES256 share a single mask bit in the logic of
4089 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4090 kludge to work properly if AES128 is available and AES256 isn't.
4093 *) Expand security boundary to match 1.1.1 module.
4096 *) Remove redundant features: hash file source, editing of test vectors
4097 modify fipsld to use external fips_premain.c signature.
4100 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4101 run algorithm test programs.
4104 *) Make algorithm test programs more tolerant of whitespace.
4107 *) Have SSL/TLS server implementation tolerate "mismatched" record
4108 protocol version while receiving ClientHello even if the
4109 ClientHello is fragmented. (The server can't insist on the
4110 particular protocol version it has chosen before the ServerHello
4111 message has informed the client about his choice.)
4114 *) Load error codes if they are not already present instead of using a
4115 static variable. This allows them to be cleanly unloaded and reloaded.
4118 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4120 *) Introduce limits to prevent malicious keys being able to
4121 cause a denial of service. (CVE-2006-2940)
4122 [Steve Henson, Bodo Moeller]
4124 *) Fix ASN.1 parsing of certain invalid structures that can result
4125 in a denial of service. (CVE-2006-2937) [Steve Henson]
4127 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4128 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4130 *) Fix SSL client code which could crash if connecting to a
4131 malicious SSLv2 server. (CVE-2006-4343)
4132 [Tavis Ormandy and Will Drewry, Google Security Team]
4134 *) Change ciphersuite string processing so that an explicit
4135 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4136 will no longer include "AES128-SHA"), and any other similar
4137 ciphersuite (same bitmap) from *other* protocol versions (so that
4138 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4139 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4140 changes from 0.9.8b and 0.9.8d.
4143 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4145 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4146 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4148 *) Change the Unix randomness entropy gathering to use poll() when
4149 possible instead of select(), since the latter has some
4150 undesirable limitations.
4151 [Darryl Miles via Richard Levitte and Bodo Moeller]
4153 *) Disable rogue ciphersuites:
4155 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4156 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4157 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4159 The latter two were purportedly from
4160 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4163 Also deactive the remaining ciphersuites from
4164 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4165 unofficial, and the ID has long expired.
4168 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4169 dual-core machines) and other potential thread-safety issues.
4172 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4174 *) Adapt fipsld and the build system to link against the validated FIPS
4175 module in FIPS mode.
4178 *) Fixes for VC++ 2005 build under Windows.
4181 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4182 from a Windows bash shell such as MSYS. It is autodetected from the
4183 "config" script when run from a VC++ environment. Modify standard VC++
4184 build to use fipscanister.o from the GNU make build.
4187 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4189 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4190 The value now differs depending on if you build for FIPS or not.
4191 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4192 safely run with a non-FIPSed libcrypto, as it may crash because of
4193 the difference induced by this change.
4196 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4198 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4199 (part of SSL_OP_ALL). This option used to disable the
4200 countermeasure against man-in-the-middle protocol-version
4201 rollback in the SSL 2.0 server implementation, which is a bad
4202 idea. (CVE-2005-2969)
4204 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4205 for Information Security, National Institute of Advanced Industrial
4206 Science and Technology [AIST], Japan)]
4208 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4209 mainly for FIPS compliance and not fully integrated at this stage.
4212 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4213 the exponentiation using a fixed-length exponent. (Otherwise,
4214 the information leaked through timing could expose the secret key
4215 after many signatures; cf. Bleichenbacher's attack on DSA with
4219 *) Make a new fixed-window mod_exp implementation the default for
4220 RSA, DSA, and DH private-key operations so that the sequence of
4221 squares and multiplies and the memory access pattern are
4222 independent of the particular secret key. This will mitigate
4223 cache-timing and potential related attacks.
4225 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4226 and this is automatically used by BN_mod_exp_mont() if the new flag
4227 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4228 will use this BN flag for private exponents unless the flag
4229 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4230 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4232 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4234 *) Change the client implementation for SSLv23_method() and
4235 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4236 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4237 (Previously, the SSL 2.0 backwards compatible Client Hello
4238 message format would be used even with SSL_OP_NO_SSLv2.)
4241 *) Add support for smime-type MIME parameter in S/MIME messages which some
4245 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4246 a threadsafe manner. Modify rsa code to use new function and add calls
4247 to dsa and dh code (which had race conditions before).
4250 *) Include the fixed error library code in the C error file definitions
4251 instead of fixing them up at runtime. This keeps the error code
4252 structures constant.
4255 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4257 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4260 *) Fixes for newer kerberos headers. NB: the casts are needed because
4261 the 'length' field is signed on one version and unsigned on another
4262 with no (?) obvious way to tell the difference, without these VC++
4263 complains. Also the "definition" of FAR (blank) is no longer included
4264 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4265 some needed definitions.
4268 *) Undo Cygwin change.
4271 *) Added support for proxy certificates according to RFC 3820.
4272 Because they may be a security thread to unaware applications,
4273 they must be explicitely allowed in run-time. See
4274 docs/HOWTO/proxy_certificates.txt for further information.
4277 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4279 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4280 server and client random values. Previously
4281 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4282 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4284 This change has negligible security impact because:
4286 1. Server and client random values still have 24 bytes of pseudo random
4289 2. Server and client random values are sent in the clear in the initial
4292 3. The master secret is derived using the premaster secret (48 bytes in
4293 size for static RSA ciphersuites) as well as client server and random
4296 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4299 [Stephen Henson, reported by UK NISCC]
4301 *) Use Windows randomness collection on Cygwin.
4304 *) Fix hang in EGD/PRNGD query when communication socket is closed
4305 prematurely by EGD/PRNGD.
4306 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4308 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4311 *) Back-port of selected performance improvements from development
4312 branch, as well as improved support for PowerPC platforms.
4315 *) Add lots of checks for memory allocation failure, error codes to indicate
4316 failure and freeing up memory if a failure occurs.
4317 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4319 *) Add new -passin argument to dgst.
4322 *) Perform some character comparisons of different types in X509_NAME_cmp:
4323 this is needed for some certificates that reencode DNs into UTF8Strings
4324 (in violation of RFC3280) and can't or wont issue name rollover
4328 *) Make an explicit check during certificate validation to see that
4329 the CA setting in each certificate on the chain is correct. As a
4330 side effect always do the following basic checks on extensions,
4331 not just when there's an associated purpose to the check:
4333 - if there is an unhandled critical extension (unless the user
4334 has chosen to ignore this fault)
4335 - if the path length has been exceeded (if one is set at all)
4336 - that certain extensions fit the associated purpose (if one has
4340 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4342 *) Avoid a race condition when CRLs are checked in a multi threaded
4343 environment. This would happen due to the reordering of the revoked
4344 entries during signature checking and serial number lookup. Now the
4345 encoding is cached and the serial number sort performed under a lock.
4346 Add new STACK function sk_is_sorted().
4349 *) Add Delta CRL to the extension code.
4352 *) Various fixes to s3_pkt.c so alerts are sent properly.
4353 [David Holmes <d.holmes@f5.com>]
4355 *) Reduce the chances of duplicate issuer name and serial numbers (in
4356 violation of RFC3280) using the OpenSSL certificate creation utilities.
4357 This is done by creating a random 64 bit value for the initial serial
4358 number when a serial number file is created or when a self signed
4359 certificate is created using 'openssl req -x509'. The initial serial
4360 number file is created using 'openssl x509 -next_serial' in CA.pl
4361 rather than being initialized to 1.
4364 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4366 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4367 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4368 [Joe Orton, Steve Henson]
4370 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4372 [Joe Orton, Steve Henson]
4374 *) Make it possible to have multiple active certificates with the same
4375 subject in the CA index file. This is done only if the keyword
4376 'unique_subject' is set to 'no' in the main CA section (default
4377 if 'CA_default') of the configuration file. The value is saved
4378 with the database itself in a separate index attribute file,
4379 named like the index file with '.attr' appended to the name.
4382 *) X509 verify fixes. Disable broken certificate workarounds when
4383 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4384 keyUsage extension present. Don't accept CRLs with unhandled critical
4385 extensions: since verify currently doesn't process CRL extensions this
4386 rejects a CRL with *any* critical extensions. Add new verify error codes
4390 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4391 A clarification of RFC2560 will require the use of OCTET STRINGs and
4392 some implementations cannot handle the current raw format. Since OpenSSL
4393 copies and compares OCSP nonces as opaque blobs without any attempt at
4394 parsing them this should not create any compatibility issues.
4397 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4398 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4399 this HMAC (and other) operations are several times slower than OpenSSL
4403 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4404 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4406 *) Use the correct content when signing type "other".
4409 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4411 *) Fix various bugs revealed by running the NISCC test suite:
4413 Stop out of bounds reads in the ASN1 code when presented with
4414 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4416 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4418 If verify callback ignores invalid public key errors don't try to check
4419 certificate signature with the NULL public key.
4423 *) New -ignore_err option in ocsp application to stop the server
4424 exiting on the first error in a request.
4427 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4428 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4432 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4433 extra data after the compression methods not only for TLS 1.0
4434 but also for SSL 3.0 (as required by the specification).
4435 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4437 *) Change X509_certificate_type() to mark the key as exported/exportable
4438 when it's 512 *bits* long, not 512 bytes.
4441 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4442 blocks during encryption.
4445 *) Various fixes to base64 BIO and non blocking I/O. On write
4446 flushes were not handled properly if the BIO retried. On read
4447 data was not being buffered properly and had various logic bugs.
4448 This also affects blocking I/O when the data being decoded is a
4452 *) Various S/MIME bugfixes and compatibility changes:
4453 output correct application/pkcs7 MIME type if
4454 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4455 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4456 of files as .eml work). Correctly handle very long lines in MIME
4460 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4462 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4463 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4464 a protocol version number mismatch like a decryption error
4465 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4468 *) Turn on RSA blinding by default in the default implementation
4469 to avoid a timing attack. Applications that don't want it can call
4470 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4471 They would be ill-advised to do so in most cases.
4472 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4474 *) Change RSA blinding code so that it works when the PRNG is not
4475 seeded (in this case, the secret RSA exponent is abused as
4476 an unpredictable seed -- if it is not unpredictable, there
4477 is no point in blinding anyway). Make RSA blinding thread-safe
4478 by remembering the creator's thread ID in rsa->blinding and
4479 having all other threads use local one-time blinding factors
4480 (this requires more computation than sharing rsa->blinding, but
4481 avoids excessive locking; and if an RSA object is not shared
4482 between threads, blinding will still be very fast).
4485 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4486 ENGINE as defaults for all supported algorithms irrespective of
4487 the 'flags' parameter. 'flags' is now honoured, so applications
4488 should make sure they are passing it correctly.
4491 *) Target "mingw" now allows native Windows code to be generated in
4492 the Cygwin environment as well as with the MinGW compiler.
4495 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4497 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4498 via timing by performing a MAC computation even if incorrrect
4499 block cipher padding has been found. This is a countermeasure
4500 against active attacks where the attacker has to distinguish
4501 between bad padding and a MAC verification error. (CVE-2003-0078)
4503 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4504 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4505 Martin Vuagnoux (EPFL, Ilion)]
4507 *) Make the no-err option work as intended. The intention with no-err
4508 is not to have the whole error stack handling routines removed from
4509 libcrypto, it's only intended to remove all the function name and
4510 reason texts, thereby removing some of the footprint that may not
4511 be interesting if those errors aren't displayed anyway.
4513 NOTE: it's still possible for any application or module to have it's
4514 own set of error texts inserted. The routines are there, just not
4515 used by default when no-err is given.
4518 *) Add support for FreeBSD on IA64.
4519 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4521 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4522 Kerberos function mit_des_cbc_cksum(). Before this change,
4523 the value returned by DES_cbc_cksum() was like the one from
4524 mit_des_cbc_cksum(), except the bytes were swapped.
4525 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4527 *) Allow an application to disable the automatic SSL chain building.
4528 Before this a rather primitive chain build was always performed in
4529 ssl3_output_cert_chain(): an application had no way to send the
4530 correct chain if the automatic operation produced an incorrect result.
4532 Now the chain builder is disabled if either:
4534 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4536 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4538 The reasoning behind this is that an application would not want the
4539 auto chain building to take place if extra chain certificates are
4540 present and it might also want a means of sending no additional
4541 certificates (for example the chain has two certificates and the
4545 *) Add the possibility to build without the ENGINE framework.
4546 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4548 *) Under Win32 gmtime() can return NULL: check return value in
4549 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4552 *) DSA routines: under certain error conditions uninitialized BN objects
4553 could be freed. Solution: make sure initialization is performed early
4554 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4555 Nils Larsch <nla@trustcenter.de> via PR#459)
4558 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4559 checked on reconnect on the client side, therefore session resumption
4560 could still fail with a "ssl session id is different" error. This
4561 behaviour is masked when SSL_OP_ALL is used due to
4562 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4563 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4564 followup to PR #377.
4567 *) IA-32 assembler support enhancements: unified ELF targets, support
4568 for SCO/Caldera platforms, fix for Cygwin shared build.
4571 *) Add support for FreeBSD on sparc64. As a consequence, support for
4572 FreeBSD on non-x86 processors is separate from x86 processors on
4573 the config script, much like the NetBSD support.
4574 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4576 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4578 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4581 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4582 code (06) was taken as the first octet of the session ID and the last
4583 octet was ignored consequently. As a result SSLv2 client side session
4584 caching could not have worked due to the session ID mismatch between
4586 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4590 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4591 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4595 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4596 seems that in spite of existing for more than a year, many application
4597 author have done nothing to provide the necessary callbacks, which
4598 means that this particular engine will not work properly anywhere.
4599 This is a very unfortunate situation which forces us, in the name
4600 of usability, to give the hw_ncipher.c a static lock, which is part
4602 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4603 appear in 0.9.8 or later. We EXPECT application authors to have
4604 dealt properly with this when 0.9.8 is released (unless we actually
4605 make such changes in the libcrypto locking code that changes will
4606 have to be made anyway).
4609 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4610 octets have been read, EOF or an error occurs. Without this change
4611 some truncated ASN1 structures will not produce an error.
4614 *) Disable Heimdal support, since it hasn't been fully implemented.
4615 Still give the possibility to force the use of Heimdal, but with
4616 warnings and a request that patches get sent to openssl-dev.
4619 *) Add the VC-CE target, introduce the WINCE sysname, and add
4620 INSTALL.WCE and appropriate conditionals to make it build.
4621 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4623 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4624 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4625 edit numbers of the version.
4626 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4628 *) Introduce safe string copy and catenation functions
4629 (BUF_strlcpy() and BUF_strlcat()).
4630 [Ben Laurie (CHATS) and Richard Levitte]
4632 *) Avoid using fixed-size buffers for one-line DNs.
4633 [Ben Laurie (CHATS)]
4635 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4636 resizing buffers containing secrets, and use where appropriate.
4637 [Ben Laurie (CHATS)]
4639 *) Avoid using fixed size buffers for configuration file location.
4640 [Ben Laurie (CHATS)]
4642 *) Avoid filename truncation for various CA files.
4643 [Ben Laurie (CHATS)]
4645 *) Use sizeof in preference to magic numbers.
4646 [Ben Laurie (CHATS)]
4648 *) Avoid filename truncation in cert requests.
4649 [Ben Laurie (CHATS)]
4651 *) Add assertions to check for (supposedly impossible) buffer
4653 [Ben Laurie (CHATS)]
4655 *) Don't cache truncated DNS entries in the local cache (this could
4656 potentially lead to a spoofing attack).
4657 [Ben Laurie (CHATS)]
4659 *) Fix various buffers to be large enough for hex/decimal
4660 representations in a platform independent manner.
4661 [Ben Laurie (CHATS)]
4663 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4664 resizing buffers containing secrets, and use where appropriate.
4665 [Ben Laurie (CHATS)]
4667 *) Add BIO_indent() to avoid much slightly worrying code to do
4669 [Ben Laurie (CHATS)]
4671 *) Convert sprintf()/BIO_puts() to BIO_printf().
4672 [Ben Laurie (CHATS)]
4674 *) buffer_gets() could terminate with the buffer only half
4676 [Ben Laurie (CHATS)]
4678 *) Add assertions to prevent user-supplied crypto functions from
4679 overflowing internal buffers by having large block sizes, etc.
4680 [Ben Laurie (CHATS)]
4682 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4684 [Ben Laurie (CHATS)]
4686 *) Eliminate unused copy of key in RC4.
4687 [Ben Laurie (CHATS)]
4689 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4690 [Ben Laurie (CHATS)]
4692 *) Fix off-by-one error in EGD path.
4693 [Ben Laurie (CHATS)]
4695 *) If RANDFILE path is too long, ignore instead of truncating.
4696 [Ben Laurie (CHATS)]
4698 *) Eliminate unused and incorrectly sized X.509 structure
4700 [Ben Laurie (CHATS)]
4702 *) Eliminate unused and dangerous function knumber().
4703 [Ben Laurie (CHATS)]
4705 *) Eliminate unused and dangerous structure, KSSL_ERR.
4706 [Ben Laurie (CHATS)]
4708 *) Protect against overlong session ID context length in an encoded
4709 session object. Since these are local, this does not appear to be
4711 [Ben Laurie (CHATS)]
4713 *) Change from security patch (see 0.9.6e below) that did not affect
4714 the 0.9.6 release series:
4716 Remote buffer overflow in SSL3 protocol - an attacker could
4717 supply an oversized master key in Kerberos-enabled versions.
4719 [Ben Laurie (CHATS)]
4721 *) Change the SSL kerb5 codes to match RFC 2712.
4724 *) Make -nameopt work fully for req and add -reqopt switch.
4725 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4727 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4728 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4730 *) Make sure tests can be performed even if the corresponding algorithms
4731 have been removed entirely. This was also the last step to make
4732 OpenSSL compilable with DJGPP under all reasonable conditions.
4733 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4735 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4736 to allow version independent disabling of normally unselected ciphers,
4737 which may be activated as a side-effect of selecting a single cipher.
4739 (E.g., cipher list string "RSA" enables ciphersuites that are left
4740 out of "ALL" because they do not provide symmetric encryption.
4741 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4742 [Lutz Jaenicke, Bodo Moeller]
4744 *) Add appropriate support for separate platform-dependent build
4745 directories. The recommended way to make a platform-dependent
4746 build directory is the following (tested on Linux), maybe with
4749 # Place yourself outside of the OpenSSL source tree. In
4750 # this example, the environment variable OPENSSL_SOURCE
4751 # is assumed to contain the absolute OpenSSL source directory.
4752 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4753 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4754 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4755 mkdir -p `dirname $F`
4756 ln -s $OPENSSL_SOURCE/$F $F
4759 To be absolutely sure not to disturb the source tree, a "make clean"
4760 is a good thing. If it isn't successfull, don't worry about it,
4761 it probably means the source directory is very clean.
4764 *) Make sure any ENGINE control commands make local copies of string
4765 pointers passed to them whenever necessary. Otherwise it is possible
4766 the caller may have overwritten (or deallocated) the original string
4767 data when a later ENGINE operation tries to use the stored values.
4768 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4770 *) Improve diagnostics in file reading and command-line digests.
4771 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4773 *) Add AES modes CFB and OFB to the object database. Correct an
4774 error in AES-CFB decryption.
4777 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4778 allows existing EVP_CIPHER_CTX structures to be reused after
4779 calling EVP_*Final(). This behaviour is used by encryption
4780 BIOs and some applications. This has the side effect that
4781 applications must explicitly clean up cipher contexts with
4782 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4785 *) Check the values of dna and dnb in bn_mul_recursive before calling
4786 bn_mul_comba (a non zero value means the a or b arrays do not contain
4787 n2 elements) and fallback to bn_mul_normal if either is not zero.
4790 *) Fix escaping of non-ASCII characters when using the -subj option
4791 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4794 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4795 form for "surname", serialNumber has no short form.
4796 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4797 therefore remove "mail" short name for "internet 7".
4798 The OID for unique identifiers in X509 certificates is
4799 x500UniqueIdentifier, not uniqueIdentifier.
4800 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4803 *) Add an "init" command to the ENGINE config module and auto initialize
4804 ENGINEs. Without any "init" command the ENGINE will be initialized
4805 after all ctrl commands have been executed on it. If init=1 the
4806 ENGINE is initailized at that point (ctrls before that point are run
4807 on the uninitialized ENGINE and after on the initialized one). If
4808 init=0 then the ENGINE will not be iniatialized at all.
4811 *) Fix the 'app_verify_callback' interface so that the user-defined
4812 argument is actually passed to the callback: In the
4813 SSL_CTX_set_cert_verify_callback() prototype, the callback
4814 declaration has been changed from
4817 int (*cb)(X509_STORE_CTX *,void *);
4818 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4819 i=s->ctx->app_verify_callback(&ctx)
4820 has been changed into
4821 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4823 To update applications using SSL_CTX_set_cert_verify_callback(),
4824 a dummy argument can be added to their callback functions.
4825 [D. K. Smetters <smetters@parc.xerox.com>]
4827 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4828 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4830 *) Add and OPENSSL_LOAD_CONF define which will cause
4831 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4832 This allows older applications to transparently support certain
4833 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4834 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4835 load the config file and OPENSSL_add_all_algorithms_conf() which will
4836 always load it have also been added.
4839 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4840 Adjust NIDs and EVP layer.
4841 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4843 *) Config modules support in openssl utility.
4845 Most commands now load modules from the config file,
4846 though in a few (such as version) this isn't done
4847 because it couldn't be used for anything.
4849 In the case of ca and req the config file used is
4850 the same as the utility itself: that is the -config
4851 command line option can be used to specify an
4855 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4856 use "openssl_conf" if filename is NULL use default openssl config file.
4859 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4860 config section name. Add a new flag to tolerate a missing config file
4861 and move code to CONF_modules_load_file().
4864 *) Support for crypto accelerator cards from Accelerated Encryption
4865 Processing, www.aep.ie. (Use engine 'aep')
4866 The support was copied from 0.9.6c [engine] and adapted/corrected
4867 to work with the new engine framework.
4868 [AEP Inc. and Richard Levitte]
4870 *) Support for SureWare crypto accelerator cards from Baltimore
4871 Technologies. (Use engine 'sureware')
4872 The support was copied from 0.9.6c [engine] and adapted
4873 to work with the new engine framework.
4876 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4877 make the newer ENGINE framework commands for the CHIL engine work.
4878 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4880 *) Make it possible to produce shared libraries on ReliantUNIX.
4881 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4883 *) Add the configuration target debug-linux-ppro.
4884 Make 'openssl rsa' use the general key loading routines
4885 implemented in apps.c, and make those routines able to
4886 handle the key format FORMAT_NETSCAPE and the variant
4888 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4890 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4891 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4893 *) Add -keyform to rsautl, and document -engine.
4894 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4896 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4897 BIO_R_NO_SUCH_FILE error code rather than the generic
4898 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4901 *) Add new functions
4903 ERR_peek_last_error_line
4904 ERR_peek_last_error_line_data.
4905 These are similar to
4908 ERR_peek_error_line_data,
4909 but report on the latest error recorded rather than the first one
4910 still in the error queue.
4911 [Ben Laurie, Bodo Moeller]
4913 *) default_algorithms option in ENGINE config module. This allows things
4915 default_algorithms = ALL
4916 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4919 *) Prelminary ENGINE config module.
4922 *) New experimental application configuration code.
4925 *) Change the AES code to follow the same name structure as all other
4926 symmetric ciphers, and behave the same way. Move everything to
4927 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4928 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4930 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4931 [Ben Laurie and Theo de Raadt]
4933 *) Add option to output public keys in req command.
4934 [Massimiliano Pala madwolf@openca.org]
4936 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4937 (up to about 10% better than before for P-192 and P-224).
4940 *) New functions/macros
4942 SSL_CTX_set_msg_callback(ctx, cb)
4943 SSL_CTX_set_msg_callback_arg(ctx, arg)
4944 SSL_set_msg_callback(ssl, cb)
4945 SSL_set_msg_callback_arg(ssl, arg)
4947 to request calling a callback function
4949 void cb(int write_p, int version, int content_type,
4950 const void *buf, size_t len, SSL *ssl, void *arg)
4952 whenever a protocol message has been completely received
4953 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4954 protocol version according to which the SSL library interprets
4955 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4956 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4957 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4958 specification (change_cipher_spec(20), alert(21), handshake(22)).
4959 'buf' and 'len' point to the actual message, 'ssl' to the
4960 SSL object, and 'arg' is the application-defined value set by
4961 SSL[_CTX]_set_msg_callback_arg().
4963 'openssl s_client' and 'openssl s_server' have new '-msg' options
4964 to enable a callback that displays all protocol messages.
4967 *) Change the shared library support so shared libraries are built as
4968 soon as the corresponding static library is finished, and thereby get
4969 openssl and the test programs linked against the shared library.
4970 This still only happens when the keyword "shard" has been given to
4971 the configuration scripts.
4973 NOTE: shared library support is still an experimental thing, and
4974 backward binary compatibility is still not guaranteed.
4975 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4977 *) Add support for Subject Information Access extension.
4978 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4980 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4981 additional bytes when new memory had to be allocated, not just
4982 when reusing an existing buffer.
4985 *) New command line and configuration option 'utf8' for the req command.
4986 This allows field values to be specified as UTF8 strings.
4989 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4990 runs for the former and machine-readable output for the latter.
4993 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4994 of the e-mail address in the DN (i.e., it will go into a certificate
4995 extension only). The new configuration file option 'email_in_dn = no'
4996 has the same effect.
4997 [Massimiliano Pala madwolf@openca.org]
4999 *) Change all functions with names starting with des_ to be starting
5000 with DES_ instead. Add wrappers that are compatible with libdes,
5001 but are named _ossl_old_des_*. Finally, add macros that map the
5002 des_* symbols to the corresponding _ossl_old_des_* if libdes
5003 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5004 desired, the des_* symbols will be mapped to DES_*, with one
5007 Since we provide two compatibility mappings, the user needs to
5008 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5009 compatibility is desired. The default (i.e., when that macro
5010 isn't defined) is OpenSSL 0.9.6c compatibility.
5012 There are also macros that enable and disable the support of old
5013 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5014 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5015 are defined, the default will apply: to support the old des routines.
5017 In either case, one must include openssl/des.h to get the correct
5018 definitions. Do not try to just include openssl/des_old.h, that
5021 NOTE: This is a major break of an old API into a new one. Software
5022 authors are encouraged to switch to the DES_ style functions. Some
5023 time in the future, des_old.h and the libdes compatibility functions
5024 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5025 default), and then completely removed.
5028 *) Test for certificates which contain unsupported critical extensions.
5029 If such a certificate is found during a verify operation it is
5030 rejected by default: this behaviour can be overridden by either
5031 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5032 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5033 X509_supported_extension() has also been added which returns 1 if a
5034 particular extension is supported.
5037 *) Modify the behaviour of EVP cipher functions in similar way to digests
5038 to retain compatibility with existing code.
5041 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5042 compatibility with existing code. In particular the 'ctx' parameter does
5043 not have to be to be initialized before the call to EVP_DigestInit() and
5044 it is tidied up after a call to EVP_DigestFinal(). New function
5045 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5046 EVP_MD_CTX_copy() changed to not require the destination to be
5047 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5048 requires the destination to be valid.
5050 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5051 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5054 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5055 so that complete 'Handshake' protocol structures are kept in memory
5056 instead of overwriting 'msg_type' and 'length' with 'body' data.
5059 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5060 [Massimo Santin via Richard Levitte]
5062 *) Major restructuring to the underlying ENGINE code. This includes
5063 reduction of linker bloat, separation of pure "ENGINE" manipulation
5064 (initialisation, etc) from functionality dealing with implementations
5065 of specific crypto iterfaces. This change also introduces integrated
5066 support for symmetric ciphers and digest implementations - so ENGINEs
5067 can now accelerate these by providing EVP_CIPHER and EVP_MD
5068 implementations of their own. This is detailed in crypto/engine/README
5069 as it couldn't be adequately described here. However, there are a few
5070 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5071 were changed in the original introduction of ENGINE code have now
5072 reverted back - the hooking from this code to ENGINE is now a good
5073 deal more passive and at run-time, operations deal directly with
5074 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5075 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5076 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5077 they were not being used by the framework as there is no concept of a
5078 BIGNUM_METHOD and they could not be generalised to the new
5079 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5080 ENGINE_cpy() has been removed as it cannot be consistently defined in
5084 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5087 *) Change mkdef.pl to sort symbols that get the same entry number,
5088 and make sure the automatically generated functions ERR_load_*
5089 become part of libeay.num as well.
5092 *) New function SSL_renegotiate_pending(). This returns true once
5093 renegotiation has been requested (either SSL_renegotiate() call
5094 or HelloRequest/ClientHello receveived from the peer) and becomes
5095 false once a handshake has been completed.
5096 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5097 sends a HelloRequest, but does not ensure that a handshake takes
5098 place. SSL_renegotiate_pending() is useful for checking if the
5099 client has followed the request.)
5102 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5103 By default, clients may request session resumption even during
5104 renegotiation (if session ID contexts permit); with this option,
5105 session resumption is possible only in the first handshake.
5107 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5108 more bits available for options that should not be part of
5109 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5112 *) Add some demos for certificate and certificate request creation.
5115 *) Make maximum certificate chain size accepted from the peer application
5116 settable (SSL*_get/set_max_cert_list()), as proposed by
5117 "Douglas E. Engert" <deengert@anl.gov>.
5120 *) Add support for shared libraries for Unixware-7
5121 (Boyd Lynn Gerber <gerberb@zenez.com>).
5124 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5125 be done prior to destruction. Use this to unload error strings from
5126 ENGINEs that load their own error strings. NB: This adds two new API
5127 functions to "get" and "set" this destroy handler in an ENGINE.
5130 *) Alter all existing ENGINE implementations (except "openssl" and
5131 "openbsd") to dynamically instantiate their own error strings. This
5132 makes them more flexible to be built both as statically-linked ENGINEs
5133 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5134 Also, add stub code to each that makes building them as self-contained
5135 shared-libraries easier (see README.ENGINE).
5138 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5139 implementations into applications that are completely implemented in
5140 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5141 commands that can be used to configure what shared-library to load and
5142 to control aspects of the way it is handled. Also, made an update to
5143 the README.ENGINE file that brings its information up-to-date and
5144 provides some information and instructions on the "dynamic" ENGINE
5145 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5148 *) Make it possible to unload ranges of ERR strings with a new
5149 "ERR_unload_strings" function.
5152 *) Add a copy() function to EVP_MD.
5155 *) Make EVP_MD routines take a context pointer instead of just the
5156 md_data void pointer.
5159 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5160 that the digest can only process a single chunk of data
5161 (typically because it is provided by a piece of
5162 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5163 is only going to provide a single chunk of data, and hence the
5164 framework needn't accumulate the data for oneshot drivers.
5167 *) As with "ERR", make it possible to replace the underlying "ex_data"
5168 functions. This change also alters the storage and management of global
5169 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5170 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5171 index counters. The API functions that use this state have been changed
5172 to take a "class_index" rather than pointers to the class's local STACK
5173 and counter, and there is now an API function to dynamically create new
5174 classes. This centralisation allows us to (a) plug a lot of the
5175 thread-safety problems that existed, and (b) makes it possible to clean
5176 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5177 such data would previously have always leaked in application code and
5178 workarounds were in place to make the memory debugging turn a blind eye
5179 to it. Application code that doesn't use this new function will still
5180 leak as before, but their memory debugging output will announce it now
5181 rather than letting it slide.
5183 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5184 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5185 has a return value to indicate success or failure.
5188 *) Make it possible to replace the underlying "ERR" functions such that the
5189 global state (2 LHASH tables and 2 locks) is only used by the "default"
5190 implementation. This change also adds two functions to "get" and "set"
5191 the implementation prior to it being automatically set the first time
5192 any other ERR function takes place. Ie. an application can call "get",
5193 pass the return value to a module it has just loaded, and that module
5194 can call its own "set" function using that value. This means the
5195 module's "ERR" operations will use (and modify) the error state in the
5196 application and not in its own statically linked copy of OpenSSL code.
5199 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5200 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5201 the operation, and provides a more encapsulated way for external code
5202 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5203 to use these functions rather than manually incrementing the counts.
5205 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5208 *) Add EVP test program.
5211 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5214 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5215 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5216 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5217 These allow a CRL to be built without having to access X509_CRL fields
5218 directly. Modify 'ca' application to use new functions.
5221 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5222 bug workarounds. Rollback attack detection is a security feature.
5223 The problem will only arise on OpenSSL servers when TLSv1 is not
5224 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5225 Software authors not wanting to support TLSv1 will have special reasons
5226 for their choice and can explicitly enable this option.
5227 [Bodo Moeller, Lutz Jaenicke]
5229 *) Rationalise EVP so it can be extended: don't include a union of
5230 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5231 (similar to those existing for EVP_CIPHER_CTX).
5236 EVP_MD_CTX_init(&md); /* new function call */
5237 EVP_DigestInit(&md, EVP_sha1());
5238 EVP_DigestUpdate(&md, in, len);
5239 EVP_DigestFinal(&md, out, NULL);
5240 EVP_MD_CTX_cleanup(&md); /* new function call */
5244 *) Make DES key schedule conform to the usual scheme, as well as
5245 correcting its structure. This means that calls to DES functions
5246 now have to pass a pointer to a des_key_schedule instead of a
5247 plain des_key_schedule (which was actually always a pointer
5250 des_key_schedule ks;
5252 des_set_key_checked(..., &ks);
5253 des_ncbc_encrypt(..., &ks, ...);
5255 (Note that a later change renames 'des_...' into 'DES_...'.)
5258 *) Initial reduction of linker bloat: the use of some functions, such as
5259 PEM causes large amounts of unused functions to be linked in due to
5260 poor organisation. For example pem_all.c contains every PEM function
5261 which has a knock on effect of linking in large amounts of (unused)
5262 ASN1 code. Grouping together similar functions and splitting unrelated
5263 functions prevents this.
5266 *) Cleanup of EVP macros.
5269 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5270 correct _ecb suffix.
5273 *) Add initial OCSP responder support to ocsp application. The
5274 revocation information is handled using the text based index
5275 use by the ca application. The responder can either handle
5276 requests generated internally, supplied in files (for example
5277 via a CGI script) or using an internal minimal server.
5280 *) Add configuration choices to get zlib compression for TLS.
5283 *) Changes to Kerberos SSL for RFC 2712 compliance:
5284 1. Implemented real KerberosWrapper, instead of just using
5285 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5286 2. Implemented optional authenticator field of KerberosWrapper.
5288 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5289 and authenticator structs; see crypto/krb5/.
5291 Generalized Kerberos calls to support multiple Kerberos libraries.
5292 [Vern Staats <staatsvr@asc.hpc.mil>,
5293 Jeffrey Altman <jaltman@columbia.edu>
5294 via Richard Levitte]
5296 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5297 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5298 values for each of the key sizes rather than having just
5299 parameters (and 'speed' generating keys each time).
5302 *) Speed up EVP routines.
5305 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5306 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5307 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5308 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5310 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5311 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5312 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5315 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5317 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5320 *) Added the OS2-EMX target.
5321 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5323 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5324 to support NCONF routines in extension code. New function CONF_set_nconf()
5325 to allow functions which take an NCONF to also handle the old LHASH
5326 structure: this means that the old CONF compatible routines can be
5327 retained (in particular wrt extensions) without having to duplicate the
5328 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5331 *) Enhance the general user interface with mechanisms for inner control
5332 and with possibilities to have yes/no kind of prompts.
5335 *) Change all calls to low level digest routines in the library and
5336 applications to use EVP. Add missing calls to HMAC_cleanup() and
5337 don't assume HMAC_CTX can be copied using memcpy().
5338 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5340 *) Add the possibility to control engines through control names but with
5341 arbitrary arguments instead of just a string.
5342 Change the key loaders to take a UI_METHOD instead of a callback
5343 function pointer. NOTE: this breaks binary compatibility with earlier
5344 versions of OpenSSL [engine].
5345 Adapt the nCipher code for these new conditions and add a card insertion
5349 *) Enhance the general user interface with mechanisms to better support
5350 dialog box interfaces, application-defined prompts, the possibility
5351 to use defaults (for example default passwords from somewhere else)
5352 and interrupts/cancellations.
5355 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5356 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5359 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5360 tidy up some unnecessarily weird code in 'sk_new()').
5361 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5363 *) Change the key loading routines for ENGINEs to use the same kind
5364 callback (pem_password_cb) as all other routines that need this
5368 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5369 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5370 than this minimum value is recommended.
5373 *) New random seeder for OpenVMS, using the system process statistics
5374 that are easily reachable.
5377 *) Windows apparently can't transparently handle global
5378 variables defined in DLLs. Initialisations such as:
5380 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5382 wont compile. This is used by the any applications that need to
5383 declare their own ASN1 modules. This was fixed by adding the option
5384 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5385 needed for static libraries under Win32.
5388 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5389 setting of purpose and trust fields. New X509_STORE trust and
5390 purpose functions and tidy up setting in other SSL functions.
5393 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5394 structure. These are inherited by X509_STORE_CTX when it is
5395 initialised. This allows various defaults to be set in the
5396 X509_STORE structure (such as flags for CRL checking and custom
5397 purpose or trust settings) for functions which only use X509_STORE_CTX
5398 internally such as S/MIME.
5400 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5401 trust settings if they are not set in X509_STORE. This allows X509_STORE
5402 purposes and trust (in S/MIME for example) to override any set by default.
5404 Add command line options for CRL checking to smime, s_client and s_server
5408 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5409 are set then the CRL is looked up in the X509_STORE structure and
5410 its validity and signature checked, then if the certificate is found
5411 in the CRL the verify fails with a revoked error.
5413 Various new CRL related callbacks added to X509_STORE_CTX structure.
5415 Command line options added to 'verify' application to support this.
5417 This needs some additional work, such as being able to handle multiple
5418 CRLs with different times, extension based lookup (rather than just
5419 by subject name) and ultimately more complete V2 CRL extension
5423 *) Add a general user interface API (crypto/ui/). This is designed
5424 to replace things like des_read_password and friends (backward
5425 compatibility functions using this new API are provided).
5426 The purpose is to remove prompting functions from the DES code
5427 section as well as provide for prompting through dialog boxes in
5428 a window system and the like.
5431 *) Add "ex_data" support to ENGINE so implementations can add state at a
5432 per-structure level rather than having to store it globally.
5435 *) Make it possible for ENGINE structures to be copied when retrieved by
5436 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5437 This causes the "original" ENGINE structure to act like a template,
5438 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5439 operational state can be localised to each ENGINE structure, despite the
5440 fact they all share the same "methods". New ENGINE structures returned in
5441 this case have no functional references and the return value is the single
5442 structural reference. This matches the single structural reference returned
5443 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5447 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5448 needs to match any other type at all we need to manually clear the
5452 *) Changes to the "openssl engine" utility to include;
5453 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5454 about an ENGINE's available control commands.
5455 - executing control commands from command line arguments using the
5456 '-pre' and '-post' switches. '-post' is only used if '-t' is
5457 specified and the ENGINE is successfully initialised. The syntax for
5458 the individual commands are colon-separated, for example;
5459 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5462 *) New dynamic control command support for ENGINEs. ENGINEs can now
5463 declare their own commands (numbers), names (strings), descriptions,
5464 and input types for run-time discovery by calling applications. A
5465 subset of these commands are implicitly classed as "executable"
5466 depending on their input type, and only these can be invoked through
5467 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5468 can be based on user input, config files, etc). The distinction is
5469 that "executable" commands cannot return anything other than a boolean
5470 result and can only support numeric or string input, whereas some
5471 discoverable commands may only be for direct use through
5472 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5473 pointers, or other custom uses. The "executable" commands are to
5474 support parameterisations of ENGINE behaviour that can be
5475 unambiguously defined by ENGINEs and used consistently across any
5476 OpenSSL-based application. Commands have been added to all the
5477 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5478 control over shared-library paths without source code alterations.
5481 *) Changed all ENGINE implementations to dynamically allocate their
5482 ENGINEs rather than declaring them statically. Apart from this being
5483 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5484 this also allows the implementations to compile without using the
5485 internal engine_int.h header.
5488 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5489 'const' value. Any code that should be able to modify a RAND_METHOD
5490 should already have non-const pointers to it (ie. they should only
5491 modify their own ones).
5494 *) Made a variety of little tweaks to the ENGINE code.
5495 - "atalla" and "ubsec" string definitions were moved from header files
5496 to C code. "nuron" string definitions were placed in variables
5497 rather than hard-coded - allowing parameterisation of these values
5498 later on via ctrl() commands.
5499 - Removed unused "#if 0"'d code.
5500 - Fixed engine list iteration code so it uses ENGINE_free() to release
5501 structural references.
5502 - Constified the RAND_METHOD element of ENGINE structures.
5503 - Constified various get/set functions as appropriate and added
5504 missing functions (including a catch-all ENGINE_cpy that duplicates
5505 all ENGINE values onto a new ENGINE except reference counts/state).
5506 - Removed NULL parameter checks in get/set functions. Setting a method
5507 or function to NULL is a way of cancelling out a previously set
5508 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5509 and doesn't justify the extra error symbols and code.
5510 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5511 flags from engine_int.h to engine.h.
5512 - Changed prototypes for ENGINE handler functions (init(), finish(),
5513 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5516 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5517 to the algorithm using long division. The binary algorithm can be
5518 used only if the modulus is odd. On 32-bit systems, it is faster
5519 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5520 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5521 up to 450 bits. In 64-bit environments, the binary algorithm
5522 appears to be advantageous for much longer moduli; here we use it
5523 for moduli up to 2048 bits.
5526 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5527 could not support the combine flag in choice fields.
5530 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5531 extensions from a certificate request to the certificate.
5534 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5535 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5536 file: this allows the display of the certificate about to be
5537 signed to be customised, to allow certain fields to be included
5538 or excluded and extension details. The old system didn't display
5539 multicharacter strings properly, omitted fields not in the policy
5540 and couldn't display additional details such as extensions.
5543 *) Function EC_POINTs_mul for multiple scalar multiplication
5544 of an arbitrary number of elliptic curve points
5545 \sum scalars[i]*points[i],
5546 optionally including the generator defined for the EC_GROUP:
5547 scalar*generator + \sum scalars[i]*points[i].
5549 EC_POINT_mul is a simple wrapper function for the typical case
5550 that the point list has just one item (besides the optional
5554 *) First EC_METHODs for curves over GF(p):
5556 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5557 operations and provides various method functions that can also
5558 operate with faster implementations of modular arithmetic.
5560 EC_GFp_mont_method() reuses most functions that are part of
5561 EC_GFp_simple_method, but uses Montgomery arithmetic.
5563 [Bodo Moeller; point addition and point doubling
5564 implementation directly derived from source code provided by
5565 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5567 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5568 crypto/ec/ec_lib.c):
5570 Curves are EC_GROUP objects (with an optional group generator)
5571 based on EC_METHODs that are built into the library.
5573 Points are EC_POINT objects based on EC_GROUP objects.
5575 Most of the framework would be able to handle curves over arbitrary
5576 finite fields, but as there are no obvious types for fields other
5577 than GF(p), some functions are limited to that for now.
5580 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5581 that the file contains a complete HTTP response.
5584 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5585 change the def and num file printf format specifier from "%-40sXXX"
5586 to "%-39s XXX". The latter will always guarantee a space after the
5587 field while the former will cause them to run together if the field
5588 is 40 of more characters long.
5591 *) Constify the cipher and digest 'method' functions and structures
5592 and modify related functions to take constant EVP_MD and EVP_CIPHER
5596 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5597 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5600 *) Modify EVP_Digest*() routines so they now return values. Although the
5601 internal software routines can never fail additional hardware versions
5605 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5607 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5608 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5611 ERR_R_NESTED_ASN1_ERROR
5613 ERR_R_MISSING_ASN1_EOS
5614 were 4 .. 9, conflicting with
5615 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5617 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5618 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5620 Add new error code 'ERR_R_INTERNAL_ERROR'.
5623 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5627 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5628 sets the subject name for a new request or supersedes the
5629 subject name in a given request. Formats that can be parsed are
5630 'CN=Some Name, OU=myOU, C=IT'
5632 'CN=Some Name/OU=myOU/C=IT'.
5634 Add options '-batch' and '-verbose' to 'openssl req'.
5635 [Massimiliano Pala <madwolf@hackmasters.net>]
5637 *) Introduce the possibility to access global variables through
5638 functions on platform were that's the best way to handle exporting
5639 global variables in shared libraries. To enable this functionality,
5640 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5641 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5642 is normally done by Configure or something similar).
5644 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5645 in the source file (foo.c) like this:
5647 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5648 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5650 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5651 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5653 OPENSSL_DECLARE_GLOBAL(int,foo);
5654 #define foo OPENSSL_GLOBAL_REF(foo)
5655 OPENSSL_DECLARE_GLOBAL(double,bar);
5656 #define bar OPENSSL_GLOBAL_REF(bar)
5658 The #defines are very important, and therefore so is including the
5659 header file everywhere where the defined globals are used.
5661 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5662 of ASN.1 items, but that structure is a bit different.
5664 The largest change is in util/mkdef.pl which has been enhanced with
5665 better and easier to understand logic to choose which symbols should
5666 go into the Windows .def files as well as a number of fixes and code
5667 cleanup (among others, algorithm keywords are now sorted
5668 lexicographically to avoid constant rewrites).
5671 *) In BN_div() keep a copy of the sign of 'num' before writing the
5672 result to 'rm' because if rm==num the value will be overwritten
5673 and produce the wrong result if 'num' is negative: this caused
5674 problems with BN_mod() and BN_nnmod().
5677 *) Function OCSP_request_verify(). This checks the signature on an
5678 OCSP request and verifies the signer certificate. The signer
5679 certificate is just checked for a generic purpose and OCSP request
5683 *) Add OCSP_check_validity() function to check the validity of OCSP
5684 responses. OCSP responses are prepared in real time and may only
5685 be a few seconds old. Simply checking that the current time lies
5686 between thisUpdate and nextUpdate max reject otherwise valid responses
5687 caused by either OCSP responder or client clock inaccuracy. Instead
5688 we allow thisUpdate and nextUpdate to fall within a certain period of
5689 the current time. The age of the response can also optionally be
5690 checked. Two new options -validity_period and -status_age added to
5694 *) If signature or public key algorithm is unrecognized print out its
5695 OID rather that just UNKNOWN.
5698 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5699 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5700 ID to be generated from the issuer certificate alone which can then be
5701 passed to OCSP_id_issuer_cmp().
5704 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5705 ASN1 modules to export functions returning ASN1_ITEM pointers
5706 instead of the ASN1_ITEM structures themselves. This adds several
5707 new macros which allow the underlying ASN1 function/structure to
5708 be accessed transparently. As a result code should not use ASN1_ITEM
5709 references directly (such as &X509_it) but instead use the relevant
5710 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5711 use of the new ASN1 code on platforms where exporting structures
5712 is problematical (for example in shared libraries) but exporting
5713 functions returning pointers to structures is not.
5716 *) Add support for overriding the generation of SSL/TLS session IDs.
5717 These callbacks can be registered either in an SSL_CTX or per SSL.
5718 The purpose of this is to allow applications to control, if they wish,
5719 the arbitrary values chosen for use as session IDs, particularly as it
5720 can be useful for session caching in multiple-server environments. A
5721 command-line switch for testing this (and any client code that wishes
5722 to use such a feature) has been added to "s_server".
5723 [Geoff Thorpe, Lutz Jaenicke]
5725 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5726 of the form '#if defined(...) || defined(...) || ...' and
5727 '#if !defined(...) && !defined(...) && ...'. This also avoids
5728 the growing number of special cases it was previously handling.
5731 *) Make all configuration macros available for application by making
5732 sure they are available in opensslconf.h, by giving them names starting
5733 with "OPENSSL_" to avoid conflicts with other packages and by making
5734 sure e_os2.h will cover all platform-specific cases together with
5736 Additionally, it is now possible to define configuration/platform-
5737 specific names (called "system identities"). In the C code, these
5738 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5739 macro with the name beginning with "OPENSSL_SYS_", which is determined
5740 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5744 *) New option -set_serial to 'req' and 'x509' this allows the serial
5745 number to use to be specified on the command line. Previously self
5746 signed certificates were hard coded with serial number 0 and the
5747 CA options of 'x509' had to use a serial number in a file which was
5751 *) New options to 'ca' utility to support V2 CRL entry extensions.
5752 Currently CRL reason, invalidity date and hold instruction are
5753 supported. Add new CRL extensions to V3 code and some new objects.
5756 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5757 disable standard block padding (aka PKCS#5 padding) in the EVP
5758 API, which was previously mandatory. This means that the data is
5759 not padded in any way and so the total length much be a multiple
5760 of the block size, otherwise an error occurs.
5763 *) Initial (incomplete) OCSP SSL support.
5766 *) New function OCSP_parse_url(). This splits up a URL into its host,
5767 port and path components: primarily to parse OCSP URLs. New -url
5768 option to ocsp utility.
5771 *) New nonce behavior. The return value of OCSP_check_nonce() now
5772 reflects the various checks performed. Applications can decide
5773 whether to tolerate certain situations such as an absent nonce
5774 in a response when one was present in a request: the ocsp application
5775 just prints out a warning. New function OCSP_add1_basic_nonce()
5776 this is to allow responders to include a nonce in a response even if
5777 the request is nonce-less.
5780 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5781 skipped when using openssl x509 multiple times on a single input file,
5782 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5785 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5786 set string type: to handle setting ASN1_TIME structures. Fix ca
5787 utility to correctly initialize revocation date of CRLs.
5790 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5791 the clients preferred ciphersuites and rather use its own preferences.
5792 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5793 Internet Explorer by ensuring unchanged hash method during stepup.
5794 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5797 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5798 to aes and add a new 'exist' option to print out symbols that don't
5802 *) Additional options to ocsp utility to allow flags to be set and
5803 additional certificates supplied.
5806 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5807 OCSP client a number of certificate to only verify the response
5811 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5812 handle the new API. Currently only ECB, CBC modes supported. Add new
5815 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5816 Encryption Standard (AES) Ciphersuites for Transport Layer
5817 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5818 not enabled by default and were not part of the "ALL" ciphersuite
5819 alias because they were not yet official; they could be
5820 explicitly requested by specifying the "AESdraft" ciphersuite
5821 group alias. In the final release of OpenSSL 0.9.7, the group
5822 alias is called "AES" and is part of "ALL".)
5823 [Ben Laurie, Steve Henson, Bodo Moeller]
5825 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5826 request to response.
5829 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5830 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5831 extract information from a certificate request. OCSP_response_create()
5832 creates a response and optionally adds a basic response structure.
5833 OCSP_basic_add1_status() adds a complete single response to a basic
5834 response and returns the OCSP_SINGLERESP structure just added (to allow
5835 extensions to be included for example). OCSP_basic_add1_cert() adds a
5836 certificate to a basic response and OCSP_basic_sign() signs a basic
5837 response with various flags. New helper functions ASN1_TIME_check()
5838 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5839 (converts ASN1_TIME to GeneralizedTime).
5842 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5843 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5844 structure from a certificate. X509_pubkey_digest() digests the public_key
5845 contents: this is used in various key identifiers.
5848 *) Make sk_sort() tolerate a NULL argument.
5849 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5851 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5852 passed by the function are trusted implicitly. If any of them signed the
5853 response then it is assumed to be valid and is not verified.
5856 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5857 to data. This was previously part of the PKCS7 ASN1 code. This
5858 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5859 [Steve Henson, reported by Kenneth R. Robinette
5860 <support@securenetterm.com>]
5862 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5863 routines: without these tracing memory leaks is very painful.
5864 Fix leaks in PKCS12 and PKCS7 routines.
5867 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5868 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5869 effectively meant GeneralizedTime would never be used. Now it
5870 is initialised to -1 but X509_time_adj() now has to check the value
5871 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5872 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5873 [Steve Henson, reported by Kenneth R. Robinette
5874 <support@securenetterm.com>]
5876 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5877 result in a zero length in the ASN1_INTEGER structure which was
5878 not consistent with the structure when d2i_ASN1_INTEGER() was used
5879 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5880 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5881 where it did not print out a minus for negative ASN1_INTEGER.
5884 *) Add summary printout to ocsp utility. The various functions which
5885 convert status values to strings have been renamed to:
5886 OCSP_response_status_str(), OCSP_cert_status_str() and
5887 OCSP_crl_reason_str() and are no longer static. New options
5888 to verify nonce values and to disable verification. OCSP response
5889 printout format cleaned up.
5892 *) Add additional OCSP certificate checks. These are those specified
5893 in RFC2560. This consists of two separate checks: the CA of the
5894 certificate being checked must either be the OCSP signer certificate
5895 or the issuer of the OCSP signer certificate. In the latter case the
5896 OCSP signer certificate must contain the OCSP signing extended key
5897 usage. This check is performed by attempting to match the OCSP
5898 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5899 in the OCSP_CERTID structures of the response.
5902 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5903 and related routines. This uses the standard OpenSSL certificate
5904 verify routines to perform initial checks (just CA validity) and
5905 to obtain the certificate chain. Then additional checks will be
5906 performed on the chain. Currently the root CA is checked to see
5907 if it is explicitly trusted for OCSP signing. This is used to set
5908 a root CA as a global signing root: that is any certificate that
5909 chains to that CA is an acceptable OCSP signing certificate.
5912 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5913 extensions from a separate configuration file.
5914 As when reading extensions from the main configuration file,
5915 the '-extensions ...' option may be used for specifying the
5917 [Massimiliano Pala <madwolf@comune.modena.it>]
5919 *) New OCSP utility. Allows OCSP requests to be generated or
5920 read. The request can be sent to a responder and the output
5921 parsed, outputed or printed in text form. Not complete yet:
5922 still needs to check the OCSP response validity.
5925 *) New subcommands for 'openssl ca':
5926 'openssl ca -status <serial>' prints the status of the cert with
5927 the given serial number (according to the index file).
5928 'openssl ca -updatedb' updates the expiry status of certificates
5930 [Massimiliano Pala <madwolf@comune.modena.it>]
5932 *) New '-newreq-nodes' command option to CA.pl. This is like
5933 '-newreq', but calls 'openssl req' with the '-nodes' option
5934 so that the resulting key is not encrypted.
5935 [Damien Miller <djm@mindrot.org>]
5937 *) New configuration for the GNU Hurd.
5938 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5940 *) Initial code to implement OCSP basic response verify. This
5941 is currently incomplete. Currently just finds the signer's
5942 certificate and verifies the signature on the response.
5945 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5946 value of OPENSSLDIR. This is available via the new '-d' option
5947 to 'openssl version', and is also included in 'openssl version -a'.
5950 *) Allowing defining memory allocation callbacks that will be given
5951 file name and line number information in additional arguments
5952 (a const char* and an int). The basic functionality remains, as
5953 well as the original possibility to just replace malloc(),
5954 realloc() and free() by functions that do not know about these
5955 additional arguments. To register and find out the current
5956 settings for extended allocation functions, the following
5957 functions are provided:
5959 CRYPTO_set_mem_ex_functions
5960 CRYPTO_set_locked_mem_ex_functions
5961 CRYPTO_get_mem_ex_functions
5962 CRYPTO_get_locked_mem_ex_functions
5964 These work the same way as CRYPTO_set_mem_functions and friends.
5965 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5966 extended allocation function is enabled.
5967 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5968 a conventional allocation function is enabled.
5969 [Richard Levitte, Bodo Moeller]
5971 *) Finish off removing the remaining LHASH function pointer casts.
5972 There should no longer be any prototype-casting required when using
5973 the LHASH abstraction, and any casts that remain are "bugs". See
5974 the callback types and macros at the head of lhash.h for details
5975 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5978 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5979 If /dev/[u]random devices are not available or do not return enough
5980 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5982 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5983 /etc/entropy will be queried once each in this sequence, quering stops
5984 when enough entropy was collected without querying more sockets.
5987 *) Change the Unix RAND_poll() variant to be able to poll several
5988 random devices, as specified by DEVRANDOM, until a sufficient amount
5989 of data has been collected. We spend at most 10 ms on each file
5990 (select timeout) and read in non-blocking mode. DEVRANDOM now
5991 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5992 (previously it was just the string "/dev/urandom"), so on typical
5993 platforms the 10 ms delay will never occur.
5994 Also separate out the Unix variant to its own file, rand_unix.c.
5995 For VMS, there's a currently-empty rand_vms.c.
5998 *) Move OCSP client related routines to ocsp_cl.c. These
5999 provide utility functions which an application needing
6000 to issue a request to an OCSP responder and analyse the
6001 response will typically need: as opposed to those which an
6002 OCSP responder itself would need which will be added later.
6004 OCSP_request_sign() signs an OCSP request with an API similar
6005 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6006 response. OCSP_response_get1_basic() extracts basic response
6007 from response. OCSP_resp_find_status(): finds and extracts status
6008 information from an OCSP_CERTID structure (which will be created
6009 when the request structure is built). These are built from lower
6010 level functions which work on OCSP_SINGLERESP structures but
6011 wont normally be used unless the application wishes to examine
6012 extensions in the OCSP response for example.
6014 Replace nonce routines with a pair of functions.
6015 OCSP_request_add1_nonce() adds a nonce value and optionally
6016 generates a random value. OCSP_check_nonce() checks the
6017 validity of the nonce in an OCSP response.
6020 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6021 This doesn't copy the supplied OCSP_CERTID and avoids the
6022 need to free up the newly created id. Change return type
6023 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6024 This can then be used to add extensions to the request.
6025 Deleted OCSP_request_new(), since most of its functionality
6026 is now in OCSP_REQUEST_new() (and the case insensitive name
6027 clash) apart from the ability to set the request name which
6028 will be added elsewhere.
6031 *) Update OCSP API. Remove obsolete extensions argument from
6032 various functions. Extensions are now handled using the new
6033 OCSP extension code. New simple OCSP HTTP function which
6034 can be used to send requests and parse the response.
6037 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6038 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6039 uses the special reorder version of SET OF to sort the attributes
6040 and reorder them to match the encoded order. This resolves a long
6041 standing problem: a verify on a PKCS7 structure just after signing
6042 it used to fail because the attribute order did not match the
6043 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6044 it uses the received order. This is necessary to tolerate some broken
6045 software that does not order SET OF. This is handled by encoding
6046 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6047 to produce the required SET OF.
6050 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6051 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6052 files to get correct declarations of the ASN.1 item variables.
6055 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6056 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6057 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6058 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6059 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6060 ASN1_ITEM and no wrapper functions.
6063 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6064 replace the old function pointer based I/O routines. Change most of
6065 the *_d2i_bio() and *_d2i_fp() functions to use these.
6068 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6069 lines, recognice more "algorithms" that can be deselected, and make
6070 it complain about algorithm deselection that isn't recognised.
6073 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6074 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6075 to use new functions. Add NO_ASN1_OLD which can be set to remove
6076 some old style ASN1 functions: this can be used to determine if old
6077 code will still work when these eventually go away.
6080 *) New extension functions for OCSP structures, these follow the
6081 same conventions as certificates and CRLs.
6084 *) New function X509V3_add1_i2d(). This automatically encodes and
6085 adds an extension. Its behaviour can be customised with various
6086 flags to append, replace or delete. Various wrappers added for
6087 certifcates and CRLs.
6090 *) Fix to avoid calling the underlying ASN1 print routine when
6091 an extension cannot be parsed. Correct a typo in the
6092 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6095 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
6096 entries for variables.
6099 *) Add functionality to apps/openssl.c for detecting locking
6100 problems: As the program is single-threaded, all we have
6101 to do is register a locking callback using an array for
6102 storing which locks are currently held by the program.
6105 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6106 SSL_get_ex_data_X509_STORE_idx(), which is used in
6107 ssl_verify_cert_chain() and thus can be called at any time
6108 during TLS/SSL handshakes so that thread-safety is essential.
6109 Unfortunately, the ex_data design is not at all suited
6110 for multi-threaded use, so it probably should be abolished.
6113 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6114 [Broadcom, tweaked and integrated by Geoff Thorpe]
6116 *) Move common extension printing code to new function
6117 X509V3_print_extensions(). Reorganise OCSP print routines and
6118 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6121 *) New function X509_signature_print() to remove duplication in some
6125 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6126 set (this was treated exactly the same as SET OF previously). This
6127 is used to reorder the STACK representing the structure to match the
6128 encoding. This will be used to get round a problem where a PKCS7
6129 structure which was signed could not be verified because the STACK
6130 order did not reflect the encoded order.
6133 *) Reimplement the OCSP ASN1 module using the new code.
6136 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6137 for its ASN1 operations. The old style function pointers still exist
6138 for now but they will eventually go away.
6141 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6142 completely replaces the old ASN1 functionality with a table driven
6143 encoder and decoder which interprets an ASN1_ITEM structure describing
6144 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6145 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6146 has also been converted to the new form.
6149 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6150 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6151 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6152 for negative moduli.
6155 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6156 of not touching the result's sign bit.
6159 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6163 *) Changed the LHASH code to use prototypes for callbacks, and created
6164 macros to declare and implement thin (optionally static) functions
6165 that provide type-safety and avoid function pointer casting for the
6166 type-specific callbacks.
6169 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6171 [Veers Staats <staatsvr@asc.hpc.mil>,
6172 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6174 *) Reformat the FAQ so the different questions and answers can be divided
6175 in sections depending on the subject.
6178 *) Have the zlib compression code load ZLIB.DLL dynamically under
6182 *) New function BN_mod_sqrt for computing square roots modulo a prime
6183 (using the probabilistic Tonelli-Shanks algorithm unless
6184 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6185 be handled deterministically).
6186 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6188 *) Make BN_mod_inverse faster by explicitly handling small quotients
6189 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6190 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6193 *) New function BN_kronecker.
6196 *) Fix BN_gcd so that it works on negative inputs; the result is
6197 positive unless both parameters are zero.
6198 Previously something reasonably close to an infinite loop was
6199 possible because numbers could be growing instead of shrinking
6200 in the implementation of Euclid's algorithm.
6203 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6204 sign of the number in question.
6206 Fix BN_is_word(a,w) to work correctly for w == 0.
6208 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6209 because its test if the absolute value of 'a' equals 'w'.
6210 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6211 it exists mostly for use in the implementations of BN_is_zero(),
6212 BN_is_one(), and BN_is_word().
6215 *) New function BN_swap.
6218 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6219 the exponentiation functions are more likely to produce reasonable
6220 results on negative inputs.
6223 *) Change BN_mod_mul so that the result is always non-negative.
6224 Previously, it could be negative if one of the factors was negative;
6225 I don't think anyone really wanted that behaviour.
6228 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6229 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6230 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6231 and add new functions:
6240 BN_mod_lshift1_quick
6244 These functions always generate non-negative results.
6246 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6247 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6249 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6250 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6251 be reduced modulo m.
6252 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6255 The following entry accidentily appeared in the CHANGES file
6256 distributed with OpenSSL 0.9.7. The modifications described in
6257 it do *not* apply to OpenSSL 0.9.7.
6259 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6260 was actually never needed) and in BN_mul(). The removal in BN_mul()
6261 required a small change in bn_mul_part_recursive() and the addition
6262 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6263 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6264 bn_sub_words() and bn_add_words() except they take arrays with
6269 *) In 'openssl passwd', verify passwords read from the terminal
6270 unless the '-salt' option is used (which usually means that
6271 verification would just waste user's time since the resulting
6272 hash is going to be compared with some given password hash)
6273 or the new '-noverify' option is used.
6275 This is an incompatible change, but it does not affect
6276 non-interactive use of 'openssl passwd' (passwords on the command
6277 line, '-stdin' option, '-in ...' option) and thus should not
6281 *) Remove all references to RSAref, since there's no more need for it.
6284 *) Make DSO load along a path given through an environment variable
6285 (SHLIB_PATH) with shl_load().
6288 *) Constify the ENGINE code as a result of BIGNUM constification.
6289 Also constify the RSA code and most things related to it. In a
6290 few places, most notable in the depth of the ASN.1 code, ugly
6291 casts back to non-const were required (to be solved at a later
6295 *) Make it so the openssl application has all engines loaded by default.
6298 *) Constify the BIGNUM routines a little more.
6301 *) Add the following functions:
6303 ENGINE_load_cswift()
6305 ENGINE_load_atalla()
6307 ENGINE_load_builtin_engines()
6309 That way, an application can itself choose if external engines that
6310 are built-in in OpenSSL shall ever be used or not. The benefit is
6311 that applications won't have to be linked with libdl or other dso
6312 libraries unless it's really needed.
6314 Changed 'openssl engine' to load all engines on demand.
6315 Changed the engine header files to avoid the duplication of some
6316 declarations (they differed!).
6319 *) 'openssl engine' can now list capabilities.
6322 *) Better error reporting in 'openssl engine'.
6325 *) Never call load_dh_param(NULL) in s_server.
6328 *) Add engine application. It can currently list engines by name and
6329 identity, and test if they are actually available.
6332 *) Improve RPM specification file by forcing symbolic linking and making
6333 sure the installed documentation is also owned by root.root.
6334 [Damien Miller <djm@mindrot.org>]
6336 *) Give the OpenSSL applications more possibilities to make use of
6337 keys (public as well as private) handled by engines.
6340 *) Add OCSP code that comes from CertCo.
6343 *) Add VMS support for the Rijndael code.
6346 *) Added untested support for Nuron crypto accelerator.
6349 *) Add support for external cryptographic devices. This code was
6350 previously distributed separately as the "engine" branch.
6351 [Geoff Thorpe, Richard Levitte]
6353 *) Rework the filename-translation in the DSO code. It is now possible to
6354 have far greater control over how a "name" is turned into a filename
6355 depending on the operating environment and any oddities about the
6356 different shared library filenames on each system.
6359 *) Support threads on FreeBSD-elf in Configure.
6362 *) Fix for SHA1 assembly problem with MASM: it produces
6363 warnings about corrupt line number information when assembling
6364 with debugging information. This is caused by the overlapping
6366 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6369 NCONF_get_number() has no error checking at all. As a replacement,
6370 NCONF_get_number_e() is defined (_e for "error checking") and is
6371 promoted strongly. The old NCONF_get_number is kept around for
6372 binary backward compatibility.
6373 Make it possible for methods to load from something other than a BIO,
6374 by providing a function pointer that is given a name instead of a BIO.
6375 For example, this could be used to load configuration data from an
6379 *) Fix for non blocking accept BIOs. Added new I/O special reason
6380 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6381 with non blocking I/O was not possible because no retry code was
6382 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6386 *) Added the beginnings of Rijndael support.
6389 *) Fix for bug in DirectoryString mask setting. Add support for
6390 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6391 to allow certificate printing to more controllable, additional
6392 'certopt' option to 'x509' to allow new printing options to be
6396 *) Clean old EAY MD5 hack from e_os.h.
6399 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6401 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6402 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6403 [Joe Orton, Steve Henson]
6405 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6407 *) Fix additional bug revealed by the NISCC test suite:
6409 Stop bug triggering large recursion when presented with
6410 certain ASN.1 tags (CVE-2003-0851)
6413 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6415 *) Fix various bugs revealed by running the NISCC test suite:
6417 Stop out of bounds reads in the ASN1 code when presented with
6418 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6420 If verify callback ignores invalid public key errors don't try to check
6421 certificate signature with the NULL public key.
6425 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6426 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6430 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6431 extra data after the compression methods not only for TLS 1.0
6432 but also for SSL 3.0 (as required by the specification).
6433 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6435 *) Change X509_certificate_type() to mark the key as exported/exportable
6436 when it's 512 *bits* long, not 512 bytes.
6439 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6441 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6442 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6443 a protocol version number mismatch like a decryption error
6444 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6447 *) Turn on RSA blinding by default in the default implementation
6448 to avoid a timing attack. Applications that don't want it can call
6449 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6450 They would be ill-advised to do so in most cases.
6451 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6453 *) Change RSA blinding code so that it works when the PRNG is not
6454 seeded (in this case, the secret RSA exponent is abused as
6455 an unpredictable seed -- if it is not unpredictable, there
6456 is no point in blinding anyway). Make RSA blinding thread-safe
6457 by remembering the creator's thread ID in rsa->blinding and
6458 having all other threads use local one-time blinding factors
6459 (this requires more computation than sharing rsa->blinding, but
6460 avoids excessive locking; and if an RSA object is not shared
6461 between threads, blinding will still be very fast).
6464 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6466 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6467 via timing by performing a MAC computation even if incorrrect
6468 block cipher padding has been found. This is a countermeasure
6469 against active attacks where the attacker has to distinguish
6470 between bad padding and a MAC verification error. (CVE-2003-0078)
6472 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6473 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6474 Martin Vuagnoux (EPFL, Ilion)]
6476 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6478 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6479 memory from it's contents. This is done with a counter that will
6480 place alternating values in each byte. This can be used to solve
6481 two issues: 1) the removal of calls to memset() by highly optimizing
6482 compilers, and 2) cleansing with other values than 0, since those can
6483 be read through on certain media, for example a swap space on disk.
6486 *) Bugfix: client side session caching did not work with external caching,
6487 because the session->cipher setting was not restored when reloading
6488 from the external cache. This problem was masked, when
6489 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6490 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6493 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6494 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6495 [Zeev Lieber <zeev-l@yahoo.com>]
6497 *) Undo an undocumented change introduced in 0.9.6e which caused
6498 repeated calls to OpenSSL_add_all_ciphers() and
6499 OpenSSL_add_all_digests() to be ignored, even after calling
6503 *) Change the default configuration reader to deal with last line not
6504 being properly terminated.
6507 *) Change X509_NAME_cmp() so it applies the special rules on handling
6508 DN values that are of type PrintableString, as well as RDNs of type
6509 emailAddress where the value has the type ia5String.
6510 [stefank@valicert.com via Richard Levitte]
6512 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6513 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6514 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6515 the bitwise-OR of the two for use by the majority of applications
6516 wanting this behaviour, and update the docs. The documented
6517 behaviour and actual behaviour were inconsistent and had been
6518 changing anyway, so this is more a bug-fix than a behavioural
6520 [Geoff Thorpe, diagnosed by Nadav Har'El]
6522 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6523 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6526 *) Fix initialization code race conditions in
6527 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6528 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6529 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6530 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6531 ssl2_get_cipher_by_char(),
6532 ssl3_get_cipher_by_char().
6533 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6535 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6536 the cached sessions are flushed, as the remove_cb() might use ex_data
6537 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6538 (see [openssl.org #212]).
6539 [Geoff Thorpe, Lutz Jaenicke]
6541 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6542 length, instead of the encoding length to d2i_ASN1_OBJECT.
6545 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6547 *) [In 0.9.6g-engine release:]
6548 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6549 [Lynn Gazis <lgazis@rainbow.com>]
6551 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6553 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6554 and get fix the header length calculation.
6555 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6556 Alon Kantor <alonk@checkpoint.com> (and others),
6559 *) Use proper error handling instead of 'assertions' in buffer
6560 overflow checks added in 0.9.6e. This prevents DoS (the
6561 assertions could call abort()).
6562 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6564 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6566 *) Add various sanity checks to asn1_get_length() to reject
6567 the ASN1 length bytes if they exceed sizeof(long), will appear
6568 negative or the content length exceeds the length of the
6570 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6572 *) Fix cipher selection routines: ciphers without encryption had no flags
6573 for the cipher strength set and where therefore not handled correctly
6574 by the selection routines (PR #130).
6577 *) Fix EVP_dsa_sha macro.
6581 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6582 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6583 that was added in OpenSSL 0.9.6d.
6585 As the countermeasure turned out to be incompatible with some
6586 broken SSL implementations, the new option is part of SSL_OP_ALL.
6587 SSL_OP_ALL is usually employed when compatibility with weird SSL
6588 implementations is desired (e.g. '-bugs' option to 's_client' and
6589 's_server'), so the new option is automatically set in many
6593 *) Changes in security patch:
6595 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6596 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6597 Air Force Materiel Command, USAF, under agreement number
6600 *) Add various sanity checks to asn1_get_length() to reject
6601 the ASN1 length bytes if they exceed sizeof(long), will appear
6602 negative or the content length exceeds the length of the
6603 supplied buffer. (CVE-2002-0659)
6604 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6606 *) Assertions for various potential buffer overflows, not known to
6608 [Ben Laurie (CHATS)]
6610 *) Various temporary buffers to hold ASCII versions of integers were
6611 too small for 64 bit platforms. (CVE-2002-0655)
6612 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6614 *) Remote buffer overflow in SSL3 protocol - an attacker could
6615 supply an oversized session ID to a client. (CVE-2002-0656)
6616 [Ben Laurie (CHATS)]
6618 *) Remote buffer overflow in SSL2 protocol - an attacker could
6619 supply an oversized client master key. (CVE-2002-0656)
6620 [Ben Laurie (CHATS)]
6622 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6624 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6625 encoded as NULL) with id-dsa-with-sha1.
6626 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6628 *) Check various X509_...() return values in apps/req.c.
6629 [Nils Larsch <nla@trustcenter.de>]
6631 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6632 an end-of-file condition would erronously be flagged, when the CRLF
6633 was just at the end of a processed block. The bug was discovered when
6634 processing data through a buffering memory BIO handing the data to a
6635 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6636 <ptsekov@syntrex.com> and Nedelcho Stanev.
6639 *) Implement a countermeasure against a vulnerability recently found
6640 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6641 before application data chunks to avoid the use of known IVs
6642 with data potentially chosen by the attacker.
6645 *) Fix length checks in ssl3_get_client_hello().
6648 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6649 to prevent ssl3_read_internal() from incorrectly assuming that
6650 ssl3_read_bytes() found application data while handshake
6651 processing was enabled when in fact s->s3->in_read_app_data was
6652 merely automatically cleared during the initial handshake.
6653 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6655 *) Fix object definitions for Private and Enterprise: they were not
6656 recognized in their shortname (=lowercase) representation. Extend
6657 obj_dat.pl to issue an error when using undefined keywords instead
6658 of silently ignoring the problem (Svenning Sorensen
6659 <sss@sss.dnsalias.net>).
6662 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6663 generators, i.e. generators other than 2 and 5. (Previously, the
6664 code did not properly initialise the 'add' and 'rem' values to
6665 BN_generate_prime().)
6667 In the new general case, we do not insist that 'generator' is
6668 actually a primitive root: This requirement is rather pointless;
6669 a generator of the order-q subgroup is just as good, if not
6673 *) Map new X509 verification errors to alerts. Discovered and submitted by
6674 Tom Wu <tom@arcot.com>.
6677 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6678 returning non-zero before the data has been completely received
6679 when using non-blocking I/O.
6680 [Bodo Moeller; problem pointed out by John Hughes]
6682 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6683 [Ben Laurie, Lutz Jaenicke]
6685 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6686 Yoram Zahavi <YoramZ@gilian.com>).
6689 *) Add information about CygWin 1.3 and on, and preserve proper
6690 configuration for the versions before that.
6691 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6693 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6694 check whether we deal with a copy of a session and do not delete from
6695 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6696 <izhar@checkpoint.com>.
6699 *) Do not store session data into the internal session cache, if it
6700 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6701 flag is set). Proposed by Aslam <aslam@funk.com>.
6704 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6708 *) [In 0.9.6d-engine release:]
6709 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6710 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6712 *) Add the configuration target linux-s390x.
6713 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6715 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6716 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6717 variable as an indication that a ClientHello message has been
6718 received. As the flag value will be lost between multiple
6719 invocations of ssl3_accept when using non-blocking I/O, the
6720 function may not be aware that a handshake has actually taken
6721 place, thus preventing a new session from being added to the
6724 To avoid this problem, we now set s->new_session to 2 instead of
6725 using a local variable.
6726 [Lutz Jaenicke, Bodo Moeller]
6728 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6729 if the SSL_R_LENGTH_MISMATCH error is detected.
6730 [Geoff Thorpe, Bodo Moeller]
6732 *) New 'shared_ldflag' column in Configure platform table.
6735 *) Fix EVP_CIPHER_mode macro.
6736 ["Dan S. Camper" <dan@bti.net>]
6738 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6739 type, we must throw them away by setting rr->length to 0.
6740 [D P Chang <dpc@qualys.com>]
6742 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6744 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6745 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6746 worked incorrectly for those cases where range = 10..._2 and
6747 3*range is two bits longer than range.)
6750 *) Only add signing time to PKCS7 structures if it is not already
6754 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6755 OBJ_ld_ce should be OBJ_id_ce.
6756 Also some ip-pda OIDs in crypto/objects/objects.txt were
6757 incorrect (cf. RFC 3039).
6758 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6760 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6761 returns early because it has nothing to do.
6762 [Andy Schneider <andy.schneider@bjss.co.uk>]
6764 *) [In 0.9.6c-engine release:]
6765 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6766 [Andy Schneider <andy.schneider@bjss.co.uk>]
6768 *) [In 0.9.6c-engine release:]
6769 Add support for Cryptographic Appliance's keyserver technology.
6770 (Use engine 'keyclient')
6771 [Cryptographic Appliances and Geoff Thorpe]
6773 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6774 is called via tools/c89.sh because arguments have to be
6775 rearranged (all '-L' options must appear before the first object
6777 [Richard Shapiro <rshapiro@abinitio.com>]
6779 *) [In 0.9.6c-engine release:]
6780 Add support for Broadcom crypto accelerator cards, backported
6782 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6784 *) [In 0.9.6c-engine release:]
6785 Add support for SureWare crypto accelerator cards from
6786 Baltimore Technologies. (Use engine 'sureware')
6787 [Baltimore Technologies and Mark Cox]
6789 *) [In 0.9.6c-engine release:]
6790 Add support for crypto accelerator cards from Accelerated
6791 Encryption Processing, www.aep.ie. (Use engine 'aep')
6792 [AEP Inc. and Mark Cox]
6794 *) Add a configuration entry for gcc on UnixWare.
6795 [Gary Benson <gbenson@redhat.com>]
6797 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6798 messages are stored in a single piece (fixed-length part and
6799 variable-length part combined) and fix various bugs found on the way.
6802 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6803 instead. BIO_gethostbyname() does not know what timeouts are
6804 appropriate, so entries would stay in cache even when they have
6806 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6808 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6809 faced with a pathologically small ClientHello fragment that does
6810 not contain client_version: Instead of aborting with an error,
6811 simply choose the highest available protocol version (i.e.,
6812 TLS 1.0 unless it is disabled). In practice, ClientHello
6813 messages are never sent like this, but this change gives us
6814 strictly correct behaviour at least for TLS.
6817 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6818 never resets s->method to s->ctx->method when called from within
6819 one of the SSL handshake functions.
6820 [Bodo Moeller; problem pointed out by Niko Baric]
6822 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6823 (sent using the client's version number) if client_version is
6824 smaller than the protocol version in use. Also change
6825 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6826 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6827 the client will at least see that alert.
6830 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6834 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6835 client receives HelloRequest while in a handshake.
6836 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6838 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6839 should end in 'break', not 'goto end' which circuments various
6840 cleanups done in state SSL_ST_OK. But session related stuff
6841 must be disabled for SSL_ST_OK in the case that we just sent a
6844 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6845 before just sending a HelloRequest.
6846 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6848 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6849 reveal whether illegal block cipher padding was found or a MAC
6850 verification error occured. (Neither SSLerr() codes nor alerts
6851 are directly visible to potential attackers, but the information
6852 may leak via logfiles.)
6854 Similar changes are not required for the SSL 2.0 implementation
6855 because the number of padding bytes is sent in clear for SSL 2.0,
6856 and the extra bytes are just ignored. However ssl/s2_pkt.c
6857 failed to verify that the purported number of padding bytes is in
6861 *) Add OpenUNIX-8 support including shared libraries
6862 (Boyd Lynn Gerber <gerberb@zenez.com>).
6865 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6866 'wristwatch attack' using huge encoding parameters (cf.
6867 James H. Manger's CRYPTO 2001 paper). Note that the
6868 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6869 encoding parameters and hence was not vulnerable.
6872 *) BN_sqr() bug fix.
6873 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6875 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6876 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6877 followed by modular reduction.
6878 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6880 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6881 equivalent based on BN_pseudo_rand() instead of BN_rand().
6884 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6885 This function was broken, as the check for a new client hello message
6886 to handle SGC did not allow these large messages.
6887 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6890 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6893 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6894 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6897 *) Rework the configuration and shared library support for Tru64 Unix.
6898 The configuration part makes use of modern compiler features and
6899 still retains old compiler behavior for those that run older versions
6900 of the OS. The shared library support part includes a variant that
6901 uses the RPATH feature, and is available through the special
6902 configuration target "alpha-cc-rpath", which will never be selected
6904 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6906 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6907 with the same message size as in ssl3_get_certificate_request().
6908 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6909 messages might inadvertently be reject as too long.
6910 [Petr Lampa <lampa@fee.vutbr.cz>]
6912 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6915 *) Modified SSL library such that the verify_callback that has been set
6916 specificly for an SSL object with SSL_set_verify() is actually being
6917 used. Before the change, a verify_callback set with this function was
6918 ignored and the verify_callback() set in the SSL_CTX at the time of
6919 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6920 to allow the necessary settings.
6923 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6924 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6925 done automatically (in contradiction to the requirements of the C
6926 standard). This made problems when used from OpenSSH.
6929 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6930 dh->length and always used
6932 BN_rand_range(priv_key, dh->p).
6934 BN_rand_range() is not necessary for Diffie-Hellman, and this
6935 specific range makes Diffie-Hellman unnecessarily inefficient if
6936 dh->length (recommended exponent length) is much smaller than the
6937 length of dh->p. We could use BN_rand_range() if the order of
6938 the subgroup was stored in the DH structure, but we only have
6943 BN_rand(priv_key, l, ...)
6945 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6951 RSA_eay_public_encrypt
6952 RSA_eay_private_decrypt
6953 RSA_eay_private_encrypt (signing)
6954 RSA_eay_public_decrypt (signature verification)
6956 (default implementations for RSA_public_encrypt,
6957 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6958 always reject numbers >= n.
6961 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6962 to synchronize access to 'locking_thread'. This is necessary on
6963 systems where access to 'locking_thread' (an 'unsigned long'
6964 variable) is not atomic.
6967 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6968 *before* setting the 'crypto_lock_rand' flag. The previous code had
6969 a race condition if 0 is a valid thread ID.
6970 [Travis Vitek <vitek@roguewave.com>]
6972 *) Add support for shared libraries under Irix.
6973 [Albert Chin-A-Young <china@thewrittenword.com>]
6975 *) Add configuration option to build on Linux on both big-endian and
6977 [Ralf Baechle <ralf@uni-koblenz.de>]
6979 *) Add the possibility to create shared libraries on HP-UX.
6982 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6984 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6985 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6986 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6987 PRNG state recovery was possible based on the output of
6988 one PRNG request appropriately sized to gain knowledge on
6989 'md' followed by enough consecutive 1-byte PRNG requests
6990 to traverse all of 'state'.
6992 1. When updating 'md_local' (the current thread's copy of 'md')
6993 during PRNG output generation, hash all of the previous
6994 'md_local' value, not just the half used for PRNG output.
6996 2. Make the number of bytes from 'state' included into the hash
6997 independent from the number of PRNG bytes requested.
6999 The first measure alone would be sufficient to avoid
7000 Markku-Juhani's attack. (Actually it had never occurred
7001 to me that the half of 'md_local' used for chaining was the
7002 half from which PRNG output bytes were taken -- I had always
7003 assumed that the secret half would be used.) The second
7004 measure makes sure that additional data from 'state' is never
7005 mixed into 'md_local' in small portions; this heuristically
7006 further strengthens the PRNG.
7009 *) Fix crypto/bn/asm/mips3.s.
7012 *) When only the key is given to "enc", the IV is undefined. Print out
7013 an error message in this case.
7016 *) Handle special case when X509_NAME is empty in X509 printing routines.
7019 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7020 positive and less than q.
7023 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7024 used: it isn't thread safe and the add_lock_callback should handle
7026 [Paul Rose <Paul.Rose@bridge.com>]
7028 *) Verify that incoming data obeys the block size in
7029 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7033 [Ulf Möller, Bodo Möller]
7035 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7036 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7037 when fixing the server behaviour for backwards-compatible 'client
7038 hello' messages. (Note that the attack is impractical against
7039 SSL 3.0 and TLS 1.0 anyway because length and version checking
7040 means that the probability of guessing a valid ciphertext is
7041 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7044 Before 0.9.5, the countermeasure (hide the error by generating a
7045 random 'decryption result') did not work properly because
7046 ERR_clear_error() was missing, meaning that SSL_get_error() would
7047 detect the supposedly ignored error.
7049 Both problems are now fixed.
7052 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7053 (previously it was 1024).
7056 *) Fix for compatibility mode trust settings: ignore trust settings
7057 unless some valid trust or reject settings are present.
7060 *) Fix for blowfish EVP: its a variable length cipher.
7063 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7064 parameters in DSA public key structures and return an error in the
7065 DSA routines if parameters are absent.
7068 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7069 in the current directory if neither $RANDFILE nor $HOME was set.
7070 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7071 caused some confusion to Windows users who haven't defined $HOME.
7072 Thus RAND_file_name() is changed again: e_os.h can define a
7073 DEFAULT_HOME, which will be used if $HOME is not set.
7074 For Windows, we use "C:"; on other platforms, we still require
7075 environment variables.
7077 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7078 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7079 having multiple threads call RAND_poll() concurrently.
7082 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7083 combination of a flag and a thread ID variable.
7084 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7085 flag), *other* threads can enter ssleay_add_bytes without obeying
7086 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7087 that they do not hold after the first thread unsets add_do_not_lock).
7090 *) Change bctest again: '-x' expressions are not available in all
7094 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7096 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7097 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7099 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7100 the default extension for executables, if any. Also, make the perl
7101 scripts that use symlink() to test if it really exists and use "cp"
7102 if it doesn't. All this made OpenSSL compilable and installable in
7106 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7107 If SEQUENCE is length is indefinite just set c->slen to the total
7108 amount of data available.
7109 [Steve Henson, reported by shige@FreeBSD.org]
7110 [This change does not apply to 0.9.7.]
7112 *) Change bctest to avoid here-documents inside command substitution
7113 (workaround for FreeBSD /bin/sh bug).
7114 For compatibility with Ultrix, avoid shell functions (introduced
7115 in the bctest version that searches along $PATH).
7118 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7119 with des_encrypt() defined on some operating systems, like Solaris
7123 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7124 On the Importance of Eliminating Errors in Cryptographic
7125 Computations, J. Cryptology 14 (2001) 2, 101-119,
7126 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7129 *) MIPS assembler BIGNUM division bug fix.
7132 *) Disabled incorrect Alpha assembler code.
7135 *) Fix PKCS#7 decode routines so they correctly update the length
7136 after reading an EOC for the EXPLICIT tag.
7138 [This change does not apply to 0.9.7.]
7140 *) Fix bug in PKCS#12 key generation routines. This was triggered
7141 if a 3DES key was generated with a 0 initial byte. Include
7142 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7143 (but broken) behaviour.
7146 *) Enhance bctest to search for a working bc along $PATH and print
7148 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7150 *) Fix memory leaks in err.c: free err_data string if necessary;
7151 don't write to the wrong index in ERR_set_error_data.
7154 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7158 *) Replace rdtsc with _emit statements for VC++ version 5.
7159 [Jeremy Cooper <jeremy@baymoo.org>]
7161 *) Make it possible to reuse SSLv2 sessions.
7164 *) In copy_email() check for >= 0 as a return value for
7165 X509_NAME_get_index_by_NID() since 0 is a valid index.
7166 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7168 *) Avoid coredump with unsupported or invalid public keys by checking if
7169 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7170 PKCS7_verify() fails with non detached data.
7173 *) Don't use getenv in library functions when run as setuid/setgid.
7174 New function OPENSSL_issetugid().
7177 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7178 due to incorrect handling of multi-threading:
7180 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7182 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7184 3. Count how many times MemCheck_off() has been called so that
7185 nested use can be treated correctly. This also avoids
7186 inband-signalling in the previous code (which relied on the
7187 assumption that thread ID 0 is impossible).
7190 *) Add "-rand" option also to s_client and s_server.
7193 *) Fix CPU detection on Irix 6.x.
7194 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7195 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7197 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7200 [This change does not apply to 0.9.7.]
7202 *) Use the cached encoding of an X509_NAME structure rather than
7203 copying it. This is apparently the reason for the libsafe "errors"
7204 but the code is actually correct.
7207 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7208 Bleichenbacher's DSA attack.
7209 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7210 to be set and top=0 forces the highest bit to be set; top=-1 is new
7211 and leaves the highest bit random.
7212 [Ulf Moeller, Bodo Moeller]
7214 *) In the NCONF_...-based implementations for CONF_... queries
7215 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7216 a temporary CONF structure with the data component set to NULL
7217 (which gives segmentation faults in lh_retrieve).
7218 Instead, use NULL for the CONF pointer in CONF_get_string and
7219 CONF_get_number (which may use environment variables) and directly
7220 return NULL from CONF_get_section.
7223 *) Fix potential buffer overrun for EBCDIC.
7226 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7227 keyUsage if basicConstraints absent for a CA.
7230 *) Make SMIME_write_PKCS7() write mail header values with a format that
7231 is more generally accepted (no spaces before the semicolon), since
7232 some programs can't parse those values properly otherwise. Also make
7233 sure BIO's that break lines after each write do not create invalid
7237 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7238 macros previously used would not encode an empty SEQUENCE OF
7239 and break the signature.
7241 [This change does not apply to 0.9.7.]
7243 *) Zero the premaster secret after deriving the master secret in
7247 *) Add some EVP_add_digest_alias registrations (as found in
7248 OpenSSL_add_all_digests()) to SSL_library_init()
7249 aka OpenSSL_add_ssl_algorithms(). This provides improved
7250 compatibility with peers using X.509 certificates
7251 with unconventional AlgorithmIdentifier OIDs.
7254 *) Fix for Irix with NO_ASM.
7255 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7257 *) ./config script fixes.
7258 [Ulf Moeller, Richard Levitte]
7260 *) Fix 'openssl passwd -1'.
7263 *) Change PKCS12_key_gen_asc() so it can cope with non null
7264 terminated strings whose length is passed in the passlen
7265 parameter, for example from PEM callbacks. This was done
7266 by adding an extra length parameter to asc2uni().
7267 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7269 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7270 call failed, free the DSA structure.
7273 *) Fix to uni2asc() to cope with zero length Unicode strings.
7274 These are present in some PKCS#12 files.
7277 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7278 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7279 when writing a 32767 byte record.
7280 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7282 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7283 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7285 (RSA objects have a reference count access to which is protected
7286 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7287 so they are meant to be shared between threads.)
7288 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7289 "Reddie, Steven" <Steven.Reddie@ca.com>]
7291 *) Fix a deadlock in CRYPTO_mem_leaks().
7294 *) Use better test patterns in bntest.
7297 *) rand_win.c fix for Borland C.
7300 *) BN_rshift bugfix for n == 0.
7303 *) Add a 'bctest' script that checks for some known 'bc' bugs
7304 so that 'make test' does not abort just because 'bc' is broken.
7307 *) Store verify_result within SSL_SESSION also for client side to
7308 avoid potential security hole. (Re-used sessions on the client side
7309 always resulted in verify_result==X509_V_OK, not using the original
7310 result of the server certificate verification.)
7313 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7314 SSL3_RT_APPLICATION_DATA, return 0.
7315 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7319 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7320 releases, have been re-implemented by renaming the previous
7321 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7322 and ssl3_read_internal, respectively, and adding 'peek' parameters
7323 to them. The new ssl[23]_{read,peek} functions are calls to
7324 ssl[23]_read_internal with the 'peek' flag set appropriately.
7325 A 'peek' parameter has also been added to ssl3_read_bytes, which
7326 does the actual work for ssl3_read_internal.
7329 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7330 the method-specific "init()" handler. Also clean up ex_data after
7331 calling the method-specific "finish()" handler. Previously, this was
7332 happening the other way round.
7335 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7336 The previous value, 12, was not always sufficient for BN_mod_exp().
7339 *) Make sure that shared libraries get the internal name engine with
7340 the full version number and not just 0. This should mark the
7341 shared libraries as not backward compatible. Of course, this should
7342 be changed again when we can guarantee backward binary compatibility.
7345 *) Fix typo in get_cert_by_subject() in by_dir.c
7346 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7348 *) Rework the system to generate shared libraries:
7350 - Make note of the expected extension for the shared libraries and
7351 if there is a need for symbolic links from for example libcrypto.so.0
7352 to libcrypto.so.0.9.7. There is extended info in Configure for
7355 - Make as few rebuilds of the shared libraries as possible.
7357 - Still avoid linking the OpenSSL programs with the shared libraries.
7359 - When installing, install the shared libraries separately from the
7363 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7365 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7366 and not in SSL_clear because the latter is also used by the
7367 accept/connect functions; previously, the settings made by
7368 SSL_set_read_ahead would be lost during the handshake.
7369 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7371 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7372 Previously, it would create entries for disableed algorithms no
7376 *) Added several new manual pages for SSL_* function.
7379 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7381 *) In ssl23_get_client_hello, generate an error message when faced
7382 with an initial SSL 3.0/TLS record that is too small to contain the
7383 first two bytes of the ClientHello message, i.e. client_version.
7384 (Note that this is a pathologic case that probably has never happened
7385 in real life.) The previous approach was to use the version number
7386 from the record header as a substitute; but our protocol choice
7387 should not depend on that one because it is not authenticated
7388 by the Finished messages.
7391 *) More robust randomness gathering functions for Windows.
7392 [Jeffrey Altman <jaltman@columbia.edu>]
7394 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7395 not set then we don't setup the error code for issuer check errors
7396 to avoid possibly overwriting other errors which the callback does
7397 handle. If an application does set the flag then we assume it knows
7398 what it is doing and can handle the new informational codes
7402 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7403 a general "ANY" type, as such it should be able to decode anything
7404 including tagged types. However it didn't check the class so it would
7405 wrongly interpret tagged types in the same way as their universal
7406 counterpart and unknown types were just rejected. Changed so that the
7407 tagged and unknown types are handled in the same way as a SEQUENCE:
7408 that is the encoding is stored intact. There is also a new type
7409 "V_ASN1_OTHER" which is used when the class is not universal, in this
7410 case we have no idea what the actual type is so we just lump them all
7414 *) On VMS, stdout may very well lead to a file that is written to
7415 in a record-oriented fashion. That means that every write() will
7416 write a separate record, which will be read separately by the
7417 programs trying to read from it. This can be very confusing.
7419 The solution is to put a BIO filter in the way that will buffer
7420 text until a linefeed is reached, and then write everything a
7421 line at a time, so every record written will be an actual line,
7422 not chunks of lines and not (usually doesn't happen, but I've
7423 seen it once) several lines in one record. BIO_f_linebuffer() is
7426 Currently, it's a VMS-only method, because that's where it has
7427 been tested well enough.
7430 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7431 it can return incorrect results.
7432 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7433 but it was in 0.9.6-beta[12].)
7436 *) Disable the check for content being present when verifying detached
7437 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7438 include zero length content when signing messages.
7441 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7442 BIO_ctrl (for BIO pairs).
7445 *) Add DSO method for VMS.
7448 *) Bug fix: Montgomery multiplication could produce results with the
7452 *) Add RPM specification openssl.spec and modify it to build three
7453 packages. The default package contains applications, application
7454 documentation and run-time libraries. The devel package contains
7455 include files, static libraries and function documentation. The
7456 doc package contains the contents of the doc directory. The original
7457 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7460 *) Add a large number of documentation files for many SSL routines.
7461 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7463 *) Add a configuration entry for Sony News 4.
7464 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7466 *) Don't set the two most significant bits to one when generating a
7467 random number < q in the DSA library.
7470 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7471 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7472 the underlying transport is blocking) if a handshake took place.
7473 (The default behaviour is needed by applications such as s_client
7474 and s_server that use select() to determine when to use SSL_read;
7475 but for applications that know in advance when to expect data, it
7476 just makes things more complicated.)
7479 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7483 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7484 work better on such systems.
7485 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7487 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7488 Update PKCS12_parse() so it copies the friendlyName and the
7489 keyid to the certificates aux info.
7492 *) Fix bug in PKCS7_verify() which caused an infinite loop
7493 if there was more than one signature.
7494 [Sven Uszpelkat <su@celocom.de>]
7496 *) Major change in util/mkdef.pl to include extra information
7497 about each symbol, as well as presentig variables as well
7498 as functions. This change means that there's n more need
7499 to rebuild the .num files when some algorithms are excluded.
7502 *) Allow the verify time to be set by an application,
7503 rather than always using the current time.
7506 *) Phase 2 verify code reorganisation. The certificate
7507 verify code now looks up an issuer certificate by a
7508 number of criteria: subject name, authority key id
7509 and key usage. It also verifies self signed certificates
7510 by the same criteria. The main comparison function is
7511 X509_check_issued() which performs these checks.
7513 Lot of changes were necessary in order to support this
7514 without completely rewriting the lookup code.
7516 Authority and subject key identifier are now cached.
7518 The LHASH 'certs' is X509_STORE has now been replaced
7519 by a STACK_OF(X509_OBJECT). This is mainly because an
7520 LHASH can't store or retrieve multiple objects with
7521 the same hash value.
7523 As a result various functions (which were all internal
7524 use only) have changed to handle the new X509_STORE
7525 structure. This will break anything that messed round
7526 with X509_STORE internally.
7528 The functions X509_STORE_add_cert() now checks for an
7529 exact match, rather than just subject name.
7531 The X509_STORE API doesn't directly support the retrieval
7532 of multiple certificates matching a given criteria, however
7533 this can be worked round by performing a lookup first
7534 (which will fill the cache with candidate certificates)
7535 and then examining the cache for matches. This is probably
7536 the best we can do without throwing out X509_LOOKUP
7537 entirely (maybe later...).
7539 The X509_VERIFY_CTX structure has been enhanced considerably.
7541 All certificate lookup operations now go via a get_issuer()
7542 callback. Although this currently uses an X509_STORE it
7543 can be replaced by custom lookups. This is a simple way
7544 to bypass the X509_STORE hackery necessary to make this
7545 work and makes it possible to use more efficient techniques
7546 in future. A very simple version which uses a simple
7547 STACK for its trusted certificate store is also provided
7548 using X509_STORE_CTX_trusted_stack().
7550 The verify_cb() and verify() callbacks now have equivalents
7551 in the X509_STORE_CTX structure.
7553 X509_STORE_CTX also has a 'flags' field which can be used
7554 to customise the verify behaviour.
7557 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7558 excludes S/MIME capabilities.
7561 *) When a certificate request is read in keep a copy of the
7562 original encoding of the signed data and use it when outputing
7563 again. Signatures then use the original encoding rather than
7564 a decoded, encoded version which may cause problems if the
7565 request is improperly encoded.
7568 *) For consistency with other BIO_puts implementations, call
7569 buffer_write(b, ...) directly in buffer_puts instead of calling
7572 In BIO_puts, increment b->num_write as in BIO_write.
7573 [Peter.Sylvester@EdelWeb.fr]
7575 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7576 BN_zero, we may not return a BIGNUM with an array consisting of
7580 *) Avoid calling abort() from within the library when problems are
7581 detected, except if preprocessor symbols have been defined
7582 (such as REF_CHECK, BN_DEBUG etc.).
7585 *) New openssl application 'rsautl'. This utility can be
7586 used for low level RSA operations. DER public key
7587 BIO/fp routines also added.
7590 *) New Configure entry and patches for compiling on QNX 4.
7591 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7593 *) A demo state-machine implementation was sponsored by
7594 Nuron (http://www.nuron.com/) and is now available in
7595 demos/state_machine.
7598 *) New options added to the 'dgst' utility for signature
7599 generation and verification.
7602 *) Unrecognized PKCS#7 content types are now handled via a
7603 catch all ASN1_TYPE structure. This allows unsupported
7604 types to be stored as a "blob" and an application can
7605 encode and decode it manually.
7608 *) Fix various signed/unsigned issues to make a_strex.c
7610 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7612 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7613 length if passed a buffer. ASN1_INTEGER_to_BN failed
7614 if passed a NULL BN and its argument was negative.
7615 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7617 *) Modification to PKCS#7 encoding routines to output definite
7618 length encoding. Since currently the whole structures are in
7619 memory there's not real point in using indefinite length
7620 constructed encoding. However if OpenSSL is compiled with
7621 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7624 *) Added BIO_vprintf() and BIO_vsnprintf().
7627 *) Added more prefixes to parse for in the the strings written
7628 through a logging bio, to cover all the levels that are available
7629 through syslog. The prefixes are now:
7631 PANIC, EMERG, EMR => LOG_EMERG
7632 ALERT, ALR => LOG_ALERT
7633 CRIT, CRI => LOG_CRIT
7634 ERROR, ERR => LOG_ERR
7635 WARNING, WARN, WAR => LOG_WARNING
7636 NOTICE, NOTE, NOT => LOG_NOTICE
7637 INFO, INF => LOG_INFO
7638 DEBUG, DBG => LOG_DEBUG
7640 and as before, if none of those prefixes are present at the
7641 beginning of the string, LOG_ERR is chosen.
7643 On Win32, the LOG_* levels are mapped according to this:
7645 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7646 LOG_WARNING => EVENTLOG_WARNING_TYPE
7647 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7651 *) Made it possible to reconfigure with just the configuration
7652 argument "reconf" or "reconfigure". The command line arguments
7653 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7654 and are retrieved from there when reconfiguring.
7658 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7660 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7663 *) The obj_dat.pl script was messing up the sorting of object
7664 names. The reason was that it compared the quoted version
7665 of strings as a result "OCSP" > "OCSP Signing" because
7666 " > SPACE. Changed script to store unquoted versions of
7667 names and add quotes on output. It was also omitting some
7668 names from the lookup table if they were given a default
7669 value (that is if SN is missing it is given the same
7670 value as LN and vice versa), these are now added on the
7671 grounds that if an object has a name we should be able to
7672 look it up. Finally added warning output when duplicate
7673 short or long names are found.
7676 *) Changes needed for Tandem NSK.
7677 [Scott Uroff <scott@xypro.com>]
7679 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7680 RSA_padding_check_SSLv23(), special padding was never detected
7681 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7682 version rollback attacks was not effective.
7684 In s23_clnt.c, don't use special rollback-attack detection padding
7685 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7686 client; similarly, in s23_srvr.c, don't do the rollback check if
7687 SSL 2.0 is the only protocol enabled in the server.
7690 *) Make it possible to get hexdumps of unprintable data with 'openssl
7691 asn1parse'. By implication, the functions ASN1_parse_dump() and
7692 BIO_dump_indent() are added.
7695 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7696 these print out strings and name structures based on various
7697 flags including RFC2253 support and proper handling of
7698 multibyte characters. Added options to the 'x509' utility
7699 to allow the various flags to be set.
7702 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7703 Also change the functions X509_cmp_current_time() and
7704 X509_gmtime_adj() work with an ASN1_TIME structure,
7705 this will enable certificates using GeneralizedTime in validity
7706 dates to be checked.
7709 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7710 negative public key encodings) on by default,
7711 NO_NEG_PUBKEY_BUG can be set to disable it.
7714 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7715 content octets. An i2c_ASN1_OBJECT is unnecessary because
7716 the encoding can be trivially obtained from the structure.
7719 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7720 not read locks (CRYPTO_r_[un]lock).
7723 *) A first attempt at creating official support for shared
7724 libraries through configuration. I've kept it so the
7725 default is static libraries only, and the OpenSSL programs
7726 are always statically linked for now, but there are
7727 preparations for dynamic linking in place.
7728 This has been tested on Linux and Tru64.
7731 *) Randomness polling function for Win9x, as described in:
7732 Peter Gutmann, Software Generation of Practically Strong
7736 *) Fix so PRNG is seeded in req if using an already existing
7740 *) New options to smime application. -inform and -outform
7741 allow alternative formats for the S/MIME message including
7742 PEM and DER. The -content option allows the content to be
7743 specified separately. This should allow things like Netscape
7744 form signing output easier to verify.
7747 *) Fix the ASN1 encoding of tags using the 'long form'.
7750 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7751 STRING types. These convert content octets to and from the
7752 underlying type. The actual tag and length octets are
7753 already assumed to have been read in and checked. These
7754 are needed because all other string types have virtually
7755 identical handling apart from the tag. By having versions
7756 of the ASN1 functions that just operate on content octets
7757 IMPLICIT tagging can be handled properly. It also allows
7758 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7759 and ASN1_INTEGER are identical apart from the tag.
7762 *) Change the handling of OID objects as follows:
7764 - New object identifiers are inserted in objects.txt, following
7765 the syntax given in objects.README.
7766 - objects.pl is used to process obj_mac.num and create a new
7768 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7771 This is currently kind of a hack, and the perl code in objects.pl
7772 isn't very elegant, but it works as I intended. The simplest way
7773 to check that it worked correctly is to look in obj_dat.h and
7774 check the array nid_objs and make sure the objects haven't moved
7775 around (this is important!). Additions are OK, as well as
7776 consistent name changes.
7779 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7782 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7783 The given file adds to whatever has already been seeded into the
7784 random pool through the RANDFILE configuration file option or
7785 environment variable, or the default random state file.
7788 *) mkstack.pl now sorts each macro group into lexical order.
7789 Previously the output order depended on the order the files
7790 appeared in the directory, resulting in needless rewriting
7794 *) Patches to make OpenSSL compile under Win32 again. Mostly
7795 work arounds for the VC++ problem that it treats func() as
7796 func(void). Also stripped out the parts of mkdef.pl that
7797 added extra typesafe functions: these no longer exist.
7800 *) Reorganisation of the stack code. The macros are now all
7801 collected in safestack.h . Each macro is defined in terms of
7802 a "stack macro" of the form SKM_<name>(type, a, b). The
7803 DEBUG_SAFESTACK is now handled in terms of function casts,
7804 this has the advantage of retaining type safety without the
7805 use of additional functions. If DEBUG_SAFESTACK is not defined
7806 then the non typesafe macros are used instead. Also modified the
7807 mkstack.pl script to handle the new form. Needs testing to see
7808 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7809 the default if no major problems. Similar behaviour for ASN1_SET_OF
7810 and PKCS12_STACK_OF.
7813 *) When some versions of IIS use the 'NET' form of private key the
7814 key derivation algorithm is different. Normally MD5(password) is
7815 used as a 128 bit RC4 key. In the modified case
7816 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7817 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7818 as the old Netscape_RSA functions except they have an additional
7819 'sgckey' parameter which uses the modified algorithm. Also added
7820 an -sgckey command line option to the rsa utility. Thanks to
7821 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7822 algorithm to openssl-dev.
7825 *) The evp_local.h macros were using 'c.##kname' which resulted in
7826 invalid expansion on some systems (SCO 5.0.5 for example).
7827 Corrected to 'c.kname'.
7828 [Phillip Porch <root@theporch.com>]
7830 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7831 a STACK of email addresses from a certificate or request, these look
7832 in the subject name and the subject alternative name extensions and
7833 omit any duplicate addresses.
7836 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7837 This makes DSA verification about 2 % faster.
7840 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7841 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7842 plus overhead for 1024 bit moduli).
7843 This makes exponentiations about 0.5 % faster for 1024 bit
7844 exponents (as measured by "openssl speed rsa2048").
7847 *) Rename memory handling macros to avoid conflicts with other
7849 Malloc => OPENSSL_malloc
7850 Malloc_locked => OPENSSL_malloc_locked
7851 Realloc => OPENSSL_realloc
7852 Free => OPENSSL_free
7855 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7856 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7859 *) CygWin32 support.
7860 [John Jarvie <jjarvie@newsguy.com>]
7862 *) The type-safe stack code has been rejigged. It is now only compiled
7863 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7864 by default all type-specific stack functions are "#define"d back to
7865 standard stack functions. This results in more streamlined output
7866 but retains the type-safety checking possibilities of the original
7870 *) The STACK code has been cleaned up, and certain type declarations
7871 that didn't make a lot of sense have been brought in line. This has
7872 also involved a cleanup of sorts in safestack.h to more correctly
7873 map type-safe stack functions onto their plain stack counterparts.
7874 This work has also resulted in a variety of "const"ifications of
7875 lots of the code, especially "_cmp" operations which should normally
7876 be prototyped with "const" parameters anyway.
7879 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7880 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7881 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7882 where all of 'md' is used each time the PRNG is used, but 'state'
7883 is used only indexed by a cyclic counter. As entropy may not be
7884 well distributed from the beginning, 'md' is important as a
7885 chaining variable. However, the output function chains only half
7886 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7887 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7888 in all of 'state' being rewritten, with the new values depending
7889 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7892 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7893 the handshake is continued after ssl_verify_cert_chain();
7894 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7895 can lead to 'unexplainable' connection aborts later.
7896 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7898 *) Major EVP API cipher revision.
7899 Add hooks for extra EVP features. This allows various cipher
7900 parameters to be set in the EVP interface. Support added for variable
7901 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7902 setting of RC2 and RC5 parameters.
7904 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7907 Remove lots of duplicated code from the EVP library. For example *every*
7908 cipher init() function handles the 'iv' in the same way according to the
7909 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7910 for CFB and OFB modes they zero ctx->num.
7912 New functionality allows removal of S/MIME code RC2 hack.
7914 Most of the routines have the same form and so can be declared in terms
7917 By shifting this to the top level EVP_CipherInit() it can be removed from
7918 all individual ciphers. If the cipher wants to handle IVs or keys
7919 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7922 Change lots of functions like EVP_EncryptUpdate() to now return a
7923 value: although software versions of the algorithms cannot fail
7924 any installed hardware versions can.
7927 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7928 this option is set, tolerate broken clients that send the negotiated
7929 protocol version number instead of the requested protocol version
7933 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7934 i.e. non-zero for export ciphersuites, zero otherwise.
7935 Previous versions had this flag inverted, inconsistent with
7936 rsa_tmp_cb (..._TMP_RSA_CB).
7937 [Bodo Moeller; problem reported by Amit Chopra]
7939 *) Add missing DSA library text string. Work around for some IIS
7940 key files with invalid SEQUENCE encoding.
7943 *) Add a document (doc/standards.txt) that list all kinds of standards
7944 and so on that are implemented in OpenSSL.
7947 *) Enhance c_rehash script. Old version would mishandle certificates
7948 with the same subject name hash and wouldn't handle CRLs at all.
7949 Added -fingerprint option to crl utility, to support new c_rehash
7953 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7956 *) Fix for SSL server purpose checking. Server checking was
7957 rejecting certificates which had extended key usage present
7958 but no ssl client purpose.
7959 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7961 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7962 is a little unclear about how a blank password is handled.
7963 Since the password in encoded as a BMPString with terminating
7964 double NULL a zero length password would end up as just the
7965 double NULL. However no password at all is different and is
7966 handled differently in the PKCS#12 key generation code. NS
7967 treats a blank password as zero length. MSIE treats it as no
7968 password on export: but it will try both on import. We now do
7969 the same: PKCS12_parse() tries zero length and no password if
7970 the password is set to "" or NULL (NULL is now a valid password:
7971 it wasn't before) as does the pkcs12 application.
7974 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7975 perror when PEM_read_bio_X509_REQ fails, the error message must
7976 be obtained from the error queue.
7979 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7980 it in ERR_remove_state if appropriate, and change ERR_get_state
7981 accordingly to avoid race conditions (this is necessary because
7982 thread_hash is no longer constant once set).
7985 *) Bugfix for linux-elf makefile.one.
7988 *) RSA_get_default_method() will now cause a default
7989 RSA_METHOD to be chosen if one doesn't exist already.
7990 Previously this was only set during a call to RSA_new()
7991 or RSA_new_method(NULL) meaning it was possible for
7992 RSA_get_default_method() to return NULL.
7995 *) Added native name translation to the existing DSO code
7996 that will convert (if the flag to do so is set) filenames
7997 that are sufficiently small and have no path information
7998 into a canonical native form. Eg. "blah" converted to
7999 "libblah.so" or "blah.dll" etc.
8002 *) New function ERR_error_string_n(e, buf, len) which is like
8003 ERR_error_string(e, buf), but writes at most 'len' bytes
8004 including the 0 terminator. For ERR_error_string_n, 'buf'
8006 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8008 *) CONF library reworked to become more general. A new CONF
8009 configuration file reader "class" is implemented as well as a
8010 new functions (NCONF_*, for "New CONF") to handle it. The now
8011 old CONF_* functions are still there, but are reimplemented to
8012 work in terms of the new functions. Also, a set of functions
8013 to handle the internal storage of the configuration data is
8014 provided to make it easier to write new configuration file
8015 reader "classes" (I can definitely see something reading a
8016 configuration file in XML format, for example), called _CONF_*,
8017 or "the configuration storage API"...
8019 The new configuration file reading functions are:
8021 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8022 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8024 NCONF_default, NCONF_WIN32
8026 NCONF_dump_fp, NCONF_dump_bio
8028 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8029 NCONF_new creates a new CONF object. This works in the same way
8030 as other interfaces in OpenSSL, like the BIO interface.
8031 NCONF_dump_* dump the internal storage of the configuration file,
8032 which is useful for debugging. All other functions take the same
8033 arguments as the old CONF_* functions wth the exception of the
8034 first that must be a `CONF *' instead of a `LHASH *'.
8036 To make it easer to use the new classes with the old CONF_* functions,
8037 the function CONF_set_default_method is provided.
8040 *) Add '-tls1' option to 'openssl ciphers', which was already
8041 mentioned in the documentation but had not been implemented.
8042 (This option is not yet really useful because even the additional
8043 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8046 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8047 OpenSSL-based applications) load shared libraries and bind to
8048 them in a portable way.
8049 [Geoff Thorpe, with contributions from Richard Levitte]
8051 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8053 *) Make sure _lrotl and _lrotr are only used with MSVC.
8055 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8056 (the default implementation of RAND_status).
8058 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8059 to '-clrext' (= clear extensions), as intended and documented.
8060 [Bodo Moeller; inconsistency pointed out by Michael Attili
8063 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8064 was larger than the MD block size.
8065 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8067 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8068 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8069 using the passed key: if the passed key was a private key the result
8070 of X509_print(), for example, would be to print out all the private key
8074 *) des_quad_cksum() byte order bug fix.
8075 [Ulf Möller, using the problem description in krb4-0.9.7, where
8076 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8078 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8080 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8082 *) For easily testing in shell scripts whether some command
8083 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8084 returns with exit code 0 iff no command of the given name is available.
8085 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8086 the output goes to stdout and nothing is printed to stderr.
8087 Additional arguments are always ignored.
8089 Since for each cipher there is a command of the same name,
8090 the 'no-cipher' compilation switches can be tested this way.
8092 ('openssl no-XXX' is not able to detect pseudo-commands such
8093 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8096 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8099 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8100 is set; it will be thrown away anyway because each handshake creates
8102 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8103 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8104 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
8105 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8108 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8109 'Q' and 'R' lose their special meanings (quit/renegotiate).
8110 This is part of what -quiet does; unlike -quiet, -ign_eof
8111 does not suppress any output.
8114 *) Add compatibility options to the purpose and trust code. The
8115 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8116 accepts a certificate or CA, this was the previous behaviour,
8117 with all the associated security issues.
8119 X509_TRUST_COMPAT is the old trust behaviour: only and
8120 automatically trust self signed roots in certificate store. A
8121 new trust setting X509_TRUST_DEFAULT is used to specify that
8122 a purpose has no associated trust setting and it should instead
8123 use the value in the default purpose.
8126 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8127 and fix a memory leak.
8130 *) In util/mkerr.pl (which implements 'make errors'), preserve
8131 reason strings from the previous version of the .c file, as
8132 the default to have only downcase letters (and digits) in
8133 automatically generated reasons codes is not always appropriate.
8136 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8137 using strerror. Previously, ERR_reason_error_string() returned
8138 library names as reason strings for SYSerr; but SYSerr is a special
8139 case where small numbers are errno values, not library numbers.
8142 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8143 converts DSA parameters into DH parameters. (When creating parameters,
8144 DSA_generate_parameters is used.)
8147 *) Include 'length' (recommended exponent length) in C code generated
8148 by 'openssl dhparam -C'.
8151 *) The second argument to set_label in perlasm was already being used
8152 so couldn't be used as a "file scope" flag. Moved to third argument
8156 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8157 instead of RAND_bytes for encryption IVs and salts.
8160 *) Include RAND_status() into RAND_METHOD instead of implementing
8161 it only for md_rand.c Otherwise replacing the PRNG by calling
8162 RAND_set_rand_method would be impossible.
8165 *) Don't let DSA_generate_key() enter an infinite loop if the random
8166 number generation fails.
8169 *) New 'rand' application for creating pseudo-random output.
8172 *) Added configuration support for Linux/IA64
8173 [Rolf Haberrecker <rolf@suse.de>]
8175 *) Assembler module support for Mingw32.
8178 *) Shared library support for HPUX (in shlib/).
8179 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8181 *) Shared library support for Solaris gcc.
8182 [Lutz Behnke <behnke@trustcenter.de>]
8184 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8186 *) PKCS7_encrypt() was adding text MIME headers twice because they
8187 were added manually and by SMIME_crlf_copy().
8190 *) In bntest.c don't call BN_rand with zero bits argument.
8191 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8193 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8194 case was implemented. This caused BN_div_recp() to fail occasionally.
8197 *) Add an optional second argument to the set_label() in the perl
8198 assembly language builder. If this argument exists and is set
8199 to 1 it signals that the assembler should use a symbol whose
8200 scope is the entire file, not just the current function. This
8201 is needed with MASM which uses the format label:: for this scope.
8202 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8204 *) Change the ASN1 types so they are typedefs by default. Before
8205 almost all types were #define'd to ASN1_STRING which was causing
8206 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8210 *) Change names of new functions to the new get1/get0 naming
8211 convention: After 'get1', the caller owns a reference count
8212 and has to call ..._free; 'get0' returns a pointer to some
8213 data structure without incrementing reference counters.
8214 (Some of the existing 'get' functions increment a reference
8215 counter, some don't.)
8216 Similarly, 'set1' and 'add1' functions increase reference
8217 counters or duplicate objects.
8220 *) Allow for the possibility of temp RSA key generation failure:
8221 the code used to assume it always worked and crashed on failure.
8224 *) Fix potential buffer overrun problem in BIO_printf().
8225 [Ulf Möller, using public domain code by Patrick Powell; problem
8226 pointed out by David Sacerdote <das33@cornell.edu>]
8228 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8229 RAND_egd() and RAND_status(). In the command line application,
8230 the EGD socket can be specified like a seed file using RANDFILE
8234 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8235 Some CAs (e.g. Verisign) distribute certificates in this form.
8238 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8239 list to exclude them. This means that no special compilation option
8240 is needed to use anonymous DH: it just needs to be included in the
8244 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8245 EVP_MD_type. The old functionality is available in a new macro called
8246 EVP_MD_md(). Change code that uses it and update docs.
8249 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8250 where the 'void *' argument is replaced by a function pointer argument.
8251 Previously 'void *' was abused to point to functions, which works on
8252 many platforms, but is not correct. As these functions are usually
8253 called by macros defined in OpenSSL header files, most source code
8254 should work without changes.
8257 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8258 sections with information on -D... compiler switches used for
8259 compiling the library so that applications can see them. To enable
8260 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8261 must be defined. E.g.,
8262 #define OPENSSL_ALGORITHM_DEFINES
8263 #include <openssl/opensslconf.h>
8264 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8265 [Richard Levitte, Ulf and Bodo Möller]
8267 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8271 *) Change the 'other' type in certificate aux info to a STACK_OF
8272 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8273 the required ASN1 format: arbitrary types determined by an OID.
8276 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8277 argument to 'req'. This is not because the function is newer or
8278 better than others it just uses the work 'NEW' in the certificate
8279 request header lines. Some software needs this.
8282 *) Reorganise password command line arguments: now passwords can be
8283 obtained from various sources. Delete the PEM_cb function and make
8284 it the default behaviour: i.e. if the callback is NULL and the
8285 usrdata argument is not NULL interpret it as a null terminated pass
8286 phrase. If usrdata and the callback are NULL then the pass phrase
8287 is prompted for as usual.
8290 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8291 the support is automatically enabled. The resulting binaries will
8292 autodetect the card and use it if present.
8293 [Ben Laurie and Compaq Inc.]
8295 *) Work around for Netscape hang bug. This sends certificate request
8296 and server done in one record. Since this is perfectly legal in the
8297 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8298 the bugs/SSLv3 entry for more info.
8301 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8304 *) Add -rand argument to smime and pkcs12 applications and read/write
8308 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8311 *) Add command line password options to the remaining applications.
8314 *) Bug fix for BN_div_recp() for numerators with an even number of
8318 *) More tests in bntest.c, and changed test_bn output.
8321 *) ./config recognizes MacOS X now.
8324 *) Bug fix for BN_div() when the first words of num and divsor are
8325 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8328 *) Add support for various broken PKCS#8 formats, and command line
8329 options to produce them.
8332 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8333 get temporary BIGNUMs from a BN_CTX.
8336 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8340 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8341 include a #define from the old name to the new. The original intent
8342 was that statically linked binaries could for example just call
8343 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8344 link with digests. This never worked becayse SSLeay_add_all_digests()
8345 and SSLeay_add_all_ciphers() were in the same source file so calling
8346 one would link with the other. They are now in separate source files.
8349 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8352 *) Use a less unusual form of the Miller-Rabin primality test (it used
8353 a binary algorithm for exponentiation integrated into the Miller-Rabin
8354 loop, our standard modexp algorithms are faster).
8357 *) Support for the EBCDIC character set completed.
8358 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8360 *) Source code cleanups: use const where appropriate, eliminate casts,
8361 use void * instead of char * in lhash.
8364 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8365 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8366 this the server could overwrite ephemeral keys that the client
8370 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8371 using 50 iterations of the Rabin-Miller test.
8373 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8374 iterations of the Rabin-Miller test as required by the appendix
8375 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8376 As BN_is_prime_fasttest includes trial division, DSA parameter
8377 generation becomes much faster.
8379 This implies a change for the callback functions in DSA_is_prime
8380 and DSA_generate_parameters: The callback function is called once
8381 for each positive witness in the Rabin-Miller test, not just
8382 occasionally in the inner loop; and the parameters to the
8383 callback function now provide an iteration count for the outer
8384 loop rather than for the current invocation of the inner loop.
8385 DSA_generate_parameters additionally can call the callback
8386 function with an 'iteration count' of -1, meaning that a
8387 candidate has passed the trial division test (when q is generated
8388 from an application-provided seed, trial division is skipped).
8391 *) New function BN_is_prime_fasttest that optionally does trial
8392 division before starting the Rabin-Miller test and has
8393 an additional BN_CTX * argument (whereas BN_is_prime always
8394 has to allocate at least one BN_CTX).
8395 'callback(1, -1, cb_arg)' is called when a number has passed the
8396 trial division stage.
8399 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8403 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8406 *) New function BN_pseudo_rand().
8409 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8410 bignum version of BN_from_montgomery() with the working code from
8411 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8415 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8416 made it impossible to use the same SSL_SESSION data structure in
8417 SSL2 clients in multiple threads.
8420 *) The return value of RAND_load_file() no longer counts bytes obtained
8421 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8422 to seed the PRNG (previously an explicit byte count was required).
8423 [Ulf Möller, Bodo Möller]
8425 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8426 used (char *) instead of (void *) and had casts all over the place.
8429 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8432 *) Retain source code compatibility for BN_prime_checks macro:
8433 BN_is_prime(..., BN_prime_checks, ...) now uses
8434 BN_prime_checks_for_size to determine the appropriate number of
8435 Rabin-Miller iterations.
8438 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8439 DH_CHECK_P_NOT_SAFE_PRIME.
8440 (Check if this is true? OpenPGP calls them "strong".)
8443 *) Merge the functionality of "dh" and "gendh" programs into a new program
8444 "dhparam". The old programs are retained for now but will handle DH keys
8445 (instead of parameters) in future.
8448 *) Make the ciphers, s_server and s_client programs check the return values
8449 when a new cipher list is set.
8452 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8453 ciphers. Before when the 56bit ciphers were enabled the sorting was
8456 The syntax for the cipher sorting has been extended to support sorting by
8457 cipher-strength (using the strength_bits hard coded in the tables).
8458 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8460 Fix a bug in the cipher-command parser: when supplying a cipher command
8461 string with an "undefined" symbol (neither command nor alphanumeric
8462 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8463 an error is flagged.
8465 Due to the strength-sorting extension, the code of the
8466 ssl_create_cipher_list() function was completely rearranged. I hope that
8467 the readability was also increased :-)
8468 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8470 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8471 for the first serial number and places 2 in the serial number file. This
8472 avoids problems when the root CA is created with serial number zero and
8473 the first user certificate has the same issuer name and serial number
8477 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8478 the new code. Add documentation for this stuff.
8481 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8482 X509_*() to X509at_*() on the grounds that they don't handle X509
8483 structures and behave in an analagous way to the X509v3 functions:
8484 they shouldn't be called directly but wrapper functions should be used
8487 So we also now have some wrapper functions that call the X509at functions
8488 when passed certificate requests. (TO DO: similar things can be done with
8489 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8490 things. Some of these need some d2i or i2d and print functionality
8491 because they handle more complex structures.)
8494 *) Add missing #ifndefs that caused missing symbols when building libssl
8495 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8496 NO_RSA in ssl/s2*.c.
8497 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8499 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8500 has a return value which indicates the quality of the random data
8501 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8502 error queue. New function RAND_pseudo_bytes() generates output that is
8503 guaranteed to be unique but not unpredictable. RAND_add is like
8504 RAND_seed, but takes an extra argument for an entropy estimate
8505 (RAND_seed always assumes full entropy).
8508 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8509 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8510 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8511 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8512 false-positive rate of at most 2^-80 for random input.
8515 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8518 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8519 in the 0.9.5 release), this returns the chain
8520 from an X509_CTX structure with a dup of the stack and all
8521 the X509 reference counts upped: so the stack will exist
8522 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8525 Also make SSL_SESSION_print() print out the verify return
8529 *) Add manpage for the pkcs12 command. Also change the default
8530 behaviour so MAC iteration counts are used unless the new
8531 -nomaciter option is used. This improves file security and
8532 only older versions of MSIE (4.0 for example) need it.
8535 *) Honor the no-xxx Configure options when creating .DEF files.
8538 *) Add PKCS#10 attributes to field table: challengePassword,
8539 unstructuredName and unstructuredAddress. These are taken from
8540 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8541 international characters are used.
8543 More changes to X509_ATTRIBUTE code: allow the setting of types
8544 based on strings. Remove the 'loc' parameter when adding
8545 attributes because these will be a SET OF encoding which is sorted
8549 *) Initial changes to the 'req' utility to allow request generation
8550 automation. This will allow an application to just generate a template
8551 file containing all the field values and have req construct the
8554 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8555 used all over the place including certificate requests and PKCS#7
8556 structures. They are currently handled manually where necessary with
8557 some primitive wrappers for PKCS#7. The new functions behave in a
8558 manner analogous to the X509 extension functions: they allow
8559 attributes to be looked up by NID and added.
8561 Later something similar to the X509V3 code would be desirable to
8562 automatically handle the encoding, decoding and printing of the
8563 more complex types. The string types like challengePassword can
8564 be handled by the string table functions.
8566 Also modified the multi byte string table handling. Now there is
8567 a 'global mask' which masks out certain types. The table itself
8568 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8569 is useful when for example there is only one permissible type
8570 (as in countryName) and using the mask might result in no valid
8574 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8575 SSL_get_peer_finished to allow applications to obtain the latest
8576 Finished messages sent to the peer or expected from the peer,
8577 respectively. (SSL_get_peer_finished is usually the Finished message
8578 actually received from the peer, otherwise the protocol will be aborted.)
8580 As the Finished message are message digests of the complete handshake
8581 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8582 be used for external authentication procedures when the authentication
8583 provided by SSL/TLS is not desired or is not enough.
8586 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8587 the host supports BWX extension and if Compaq C is present on the
8588 $PATH. Just exploiting of the BWX extension results in 20-30%
8589 performance kick for some algorithms, e.g. DES and RC4 to mention
8590 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8594 *) Add support for MS "fast SGC". This is arguably a violation of the
8595 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8596 weak crypto and after checking the certificate is SGC a second one
8597 with strong crypto. MS SGC stops the first handshake after receiving
8598 the server certificate message and sends a second client hello. Since
8599 a server will typically do all the time consuming operations before
8600 expecting any further messages from the client (server key exchange
8601 is the most expensive) there is little difference between the two.
8603 To get OpenSSL to support MS SGC we have to permit a second client
8604 hello message after we have sent server done. In addition we have to
8605 reset the MAC if we do get this second client hello.
8608 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8609 if a DER encoded private key is RSA or DSA traditional format. Changed
8610 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8611 format DER encoded private key. Newer code should use PKCS#8 format which
8612 has the key type encoded in the ASN1 structure. Added DER private key
8613 support to pkcs8 application.
8616 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8617 ciphersuites has been selected (as required by the SSL 3/TLS 1
8618 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8619 is set, we interpret this as a request to violate the specification
8620 (the worst that can happen is a handshake failure, and 'correct'
8621 behaviour would result in a handshake failure anyway).
8624 *) In SSL_CTX_add_session, take into account that there might be multiple
8625 SSL_SESSION structures with the same session ID (e.g. when two threads
8626 concurrently obtain them from an external cache).
8627 The internal cache can handle only one SSL_SESSION with a given ID,
8628 so if there's a conflict, we now throw out the old one to achieve
8632 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8633 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8634 some routines that use cipher OIDs: some ciphers do not have OIDs
8635 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8639 *) Simplify the trust setting structure and code. Now we just have
8640 two sequences of OIDs for trusted and rejected settings. These will
8641 typically have values the same as the extended key usage extension
8642 and any application specific purposes.
8644 The trust checking code now has a default behaviour: it will just
8645 check for an object with the same NID as the passed id. Functions can
8646 be provided to override either the default behaviour or the behaviour
8647 for a given id. SSL client, server and email already have functions
8648 in place for compatibility: they check the NID and also return "trusted"
8649 if the certificate is self signed.
8652 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8653 traditional format into an EVP_PKEY structure.
8656 *) Add a password callback function PEM_cb() which either prompts for
8657 a password if usr_data is NULL or otherwise assumes it is a null
8658 terminated password. Allow passwords to be passed on command line
8659 environment or config files in a few more utilities.
8662 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8663 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8664 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8665 Update documentation.
8668 *) Support for ASN1 "NULL" type. This could be handled before by using
8669 ASN1_TYPE but there wasn't any function that would try to read a NULL
8670 and produce an error if it couldn't. For compatibility we also have
8671 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8672 don't allocate anything because they don't need to.
8675 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8677 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8679 *) Rebuild of the memory allocation routines used by OpenSSL code and
8680 possibly others as well. The purpose is to make an interface that
8681 provide hooks so anyone can build a separate set of allocation and
8682 deallocation routines to be used by OpenSSL, for example memory
8683 pool implementations, or something else, which was previously hard
8684 since Malloc(), Realloc() and Free() were defined as macros having
8685 the values malloc, realloc and free, respectively (except for Win32
8686 compilations). The same is provided for memory debugging code.
8687 OpenSSL already comes with functionality to find memory leaks, but
8688 this gives people a chance to debug other memory problems.
8690 With these changes, a new set of functions and macros have appeared:
8692 CRYPTO_set_mem_debug_functions() [F]
8693 CRYPTO_get_mem_debug_functions() [F]
8694 CRYPTO_dbg_set_options() [F]
8695 CRYPTO_dbg_get_options() [F]
8696 CRYPTO_malloc_debug_init() [M]
8698 The memory debug functions are NULL by default, unless the library
8699 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8700 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8701 gives the standard debugging functions that come with OpenSSL) or
8702 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8703 provided by the library user) must be used. When the standard
8704 debugging functions are used, CRYPTO_dbg_set_options can be used to
8705 request additional information:
8706 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8707 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8709 Also, things like CRYPTO_set_mem_functions will always give the
8710 expected result (the new set of functions is used for allocation
8711 and deallocation) at all times, regardless of platform and compiler
8714 To finish it up, some functions that were never use in any other
8715 way than through macros have a new API and new semantic:
8718 CRYPTO_dbg_realloc()
8721 All macros of value have retained their old syntax.
8722 [Richard Levitte and Bodo Moeller]
8724 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8725 ordering of SMIMECapabilities wasn't in "strength order" and there
8726 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8730 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8731 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8732 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8734 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8735 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8736 functionality to handle multipart/signed properly) and a utility
8737 called 'smime' to call all this stuff. This is based on code I
8738 originally wrote for Celo who have kindly allowed it to be
8739 included in OpenSSL.
8742 *) Add variants des_set_key_checked and des_set_key_unchecked of
8743 des_set_key (aka des_key_sched). Global variable des_check_key
8744 decides which of these is called by des_set_key; this way
8745 des_check_key behaves as it always did, but applications and
8746 the library itself, which was buggy for des_check_key == 1,
8747 have a cleaner way to pick the version they need.
8750 *) New function PKCS12_newpass() which changes the password of a
8754 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8755 dynamic mix. In both cases the ids can be used as an index into the
8756 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8757 functions so they accept a list of the field values and the
8758 application doesn't need to directly manipulate the X509_TRUST
8762 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8766 *) Modify the way the V3 extension code looks up extensions. This now
8767 works in a similar way to the object code: we have some "standard"
8768 extensions in a static table which is searched with OBJ_bsearch()
8769 and the application can add dynamic ones if needed. The file
8770 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8771 updated whenever a new extension is added to the core code and kept
8772 in ext_nid order. There is a simple program 'tabtest.c' which checks
8773 this. New extensions are not added too often so this file can readily
8774 be maintained manually.
8776 There are two big advantages in doing things this way. The extensions
8777 can be looked up immediately and no longer need to be "added" using
8778 X509V3_add_standard_extensions(): this function now does nothing.
8779 [Side note: I get *lots* of email saying the extension code doesn't
8780 work because people forget to call this function]
8781 Also no dynamic allocation is done unless new extensions are added:
8782 so if we don't add custom extensions there is no need to call
8783 X509V3_EXT_cleanup().
8786 *) Modify enc utility's salting as follows: make salting the default. Add a
8787 magic header, so unsalted files fail gracefully instead of just decrypting
8788 to garbage. This is because not salting is a big security hole, so people
8789 should be discouraged from doing it.
8792 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8793 digest to be passed on the command line but it only used this
8794 parameter when signing a certificate. Modified so all relevant
8795 operations are affected by the digest parameter including the
8796 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8797 DSA key was used because it didn't fix the digest.
8800 *) Initial certificate chain verify code. Currently tests the untrusted
8801 certificates for consistency with the verify purpose (which is set
8802 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8804 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8805 this is because it will reject chains with invalid extensions whereas
8806 every previous version of OpenSSL and SSLeay made no checks at all.
8808 Trust code: checks the root CA for the relevant trust settings. Trust
8809 settings have an initial value consistent with the verify purpose: e.g.
8810 if the verify purpose is for SSL client use it expects the CA to be
8811 trusted for SSL client use. However the default value can be changed to
8812 permit custom trust settings: one example of this would be to only trust
8813 certificates from a specific "secure" set of CAs.
8815 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8816 which should be used for version portability: especially since the
8817 verify structure is likely to change more often now.
8819 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8820 to set them. If not set then assume SSL clients will verify SSL servers
8823 Two new options to the verify program: -untrusted allows a set of
8824 untrusted certificates to be passed in and -purpose which sets the
8825 intended purpose of the certificate. If a purpose is set then the
8826 new chain verify code is used to check extension consistency.
8829 *) Support for the authority information access extension.
8832 *) Modify RSA and DSA PEM read routines to transparently handle
8833 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8834 public keys in a format compatible with certificate
8835 SubjectPublicKeyInfo structures. Unfortunately there were already
8836 functions called *_PublicKey_* which used various odd formats so
8837 these are retained for compatibility: however the DSA variants were
8838 never in a public release so they have been deleted. Changed dsa/rsa
8839 utilities to handle the new format: note no releases ever handled public
8840 keys so we should be OK.
8842 The primary motivation for this change is to avoid the same fiasco
8843 that dogs private keys: there are several incompatible private key
8844 formats some of which are standard and some OpenSSL specific and
8845 require various evil hacks to allow partial transparent handling and
8846 even then it doesn't work with DER formats. Given the option anything
8847 other than PKCS#8 should be dumped: but the other formats have to
8848 stay in the name of compatibility.
8850 With public keys and the benefit of hindsight one standard format
8851 is used which works with EVP_PKEY, RSA or DSA structures: though
8852 it clearly returns an error if you try to read the wrong kind of key.
8854 Added a -pubkey option to the 'x509' utility to output the public key.
8855 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8856 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8857 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8858 that do the same as the EVP_PKEY_assign_*() except they up the
8859 reference count of the added key (they don't "swallow" the
8863 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8864 CRLs would fail if the file contained no certificates or no CRLs:
8865 added a new function to read in both types and return the number
8866 read: this means that if none are read it will be an error. The
8867 DER versions of the certificate and CRL reader would always fail
8868 because it isn't possible to mix certificates and CRLs in DER format
8869 without choking one or the other routine. Changed this to just read
8870 a certificate: this is the best we can do. Also modified the code
8871 in apps/verify.c to take notice of return codes: it was previously
8872 attempting to read in certificates from NULL pointers and ignoring
8873 any errors: this is one reason why the cert and CRL reader seemed
8874 to work. It doesn't check return codes from the default certificate
8875 routines: these may well fail if the certificates aren't installed.
8878 *) Code to support otherName option in GeneralName.
8881 *) First update to verify code. Change the verify utility
8882 so it warns if it is passed a self signed certificate:
8883 for consistency with the normal behaviour. X509_verify
8884 has been modified to it will now verify a self signed
8885 certificate if *exactly* the same certificate appears
8886 in the store: it was previously impossible to trust a
8887 single self signed certificate. This means that:
8888 openssl verify ss.pem
8889 now gives a warning about a self signed certificate but
8890 openssl verify -CAfile ss.pem ss.pem
8894 *) For servers, store verify_result in SSL_SESSION data structure
8895 (and add it to external session representation).
8896 This is needed when client certificate verifications fails,
8897 but an application-provided verification callback (set by
8898 SSL_CTX_set_cert_verify_callback) allows accepting the session
8899 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8900 but returns 1): When the session is reused, we have to set
8901 ssl->verify_result to the appropriate error code to avoid
8903 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8905 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8906 case in PKCS7_dataInit() where the signed PKCS7 structure
8907 didn't contain any existing data because it was being created.
8908 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8910 *) Add a salt to the key derivation routines in enc.c. This
8911 forms the first 8 bytes of the encrypted file. Also add a
8912 -S option to allow a salt to be input on the command line.
8915 *) New function X509_cmp(). Oddly enough there wasn't a function
8916 to compare two certificates. We do this by working out the SHA1
8917 hash and comparing that. X509_cmp() will be needed by the trust
8921 *) SSL_get1_session() is like SSL_get_session(), but increments
8922 the reference count in the SSL_SESSION returned.
8923 [Geoff Thorpe <geoff@eu.c2.net>]
8925 *) Fix for 'req': it was adding a null to request attributes.
8926 Also change the X509_LOOKUP and X509_INFO code to handle
8927 certificate auxiliary information.
8930 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8934 *) Add the possibility to add extra information to the memory leak
8935 detecting output, to form tracebacks, showing from where each
8936 allocation was originated: CRYPTO_push_info("constant string") adds
8937 the string plus current file name and line number to a per-thread
8938 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8939 is like calling CYRPTO_pop_info() until the stack is empty.
8940 Also updated memory leak detection code to be multi-thread-safe.
8943 *) Add options -text and -noout to pkcs7 utility and delete the
8944 encryption options which never did anything. Update docs.
8947 *) Add options to some of the utilities to allow the pass phrase
8948 to be included on either the command line (not recommended on
8949 OSes like Unix) or read from the environment. Update the
8950 manpages and fix a few bugs.
8953 *) Add a few manpages for some of the openssl commands.
8956 *) Fix the -revoke option in ca. It was freeing up memory twice,
8957 leaking and not finding already revoked certificates.
8960 *) Extensive changes to support certificate auxiliary information.
8961 This involves the use of X509_CERT_AUX structure and X509_AUX
8962 functions. An X509_AUX function such as PEM_read_X509_AUX()
8963 can still read in a certificate file in the usual way but it
8964 will also read in any additional "auxiliary information". By
8965 doing things this way a fair degree of compatibility can be
8966 retained: existing certificates can have this information added
8967 using the new 'x509' options.
8969 Current auxiliary information includes an "alias" and some trust
8970 settings. The trust settings will ultimately be used in enhanced
8971 certificate chain verification routines: currently a certificate
8972 can only be trusted if it is self signed and then it is trusted
8976 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8977 The problem was that one of the replacement routines had not been working
8978 since SSLeay releases. For now the offending routine has been replaced
8979 with non-optimised assembler. Even so, this now gives around 95%
8980 performance improvement for 1024 bit RSA signs.
8983 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8984 handling. Most clients have the effective key size in bits equal to
8985 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8986 A few however don't do this and instead use the size of the decrypted key
8987 to determine the RC2 key length and the AlgorithmIdentifier to determine
8988 the effective key length. In this case the effective key length can still
8989 be 40 bits but the key length can be 168 bits for example. This is fixed
8990 by manually forcing an RC2 key into the EVP_PKEY structure because the
8991 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8992 the key length and effective key length are equal.
8995 *) Add a bunch of functions that should simplify the creation of
8996 X509_NAME structures. Now you should be able to do:
8997 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8998 and have it automatically work out the correct field type and fill in
8999 the structures. The more adventurous can try:
9000 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9001 and it will (hopefully) work out the correct multibyte encoding.
9004 *) Change the 'req' utility to use the new field handling and multibyte
9005 copy routines. Before the DN field creation was handled in an ad hoc
9006 way in req, ca, and x509 which was rather broken and didn't support
9007 BMPStrings or UTF8Strings. Since some software doesn't implement
9008 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9009 using the dirstring_type option. See the new comment in the default
9010 openssl.cnf for more info.
9013 *) Make crypto/rand/md_rand.c more robust:
9014 - Assure unique random numbers after fork().
9015 - Make sure that concurrent threads access the global counter and
9016 md serializably so that we never lose entropy in them
9017 or use exactly the same state in multiple threads.
9018 Access to the large state is not always serializable because
9019 the additional locking could be a performance killer, and
9020 md should be large enough anyway.
9023 *) New file apps/app_rand.c with commonly needed functionality
9024 for handling the random seed file.
9026 Use the random seed file in some applications that previously did not:
9028 dsaparam -genkey (which also ignored its '-rand' option),
9031 x509 (when signing).
9032 Except on systems with /dev/urandom, it is crucial to have a random
9033 seed file at least for key creation, DSA signing, and for DH exchanges;
9034 for RSA signatures we could do without one.
9036 gendh and gendsa (unlike genrsa) used to read only the first byte
9037 of each file listed in the '-rand' option. The function as previously
9038 found in genrsa is now in app_rand.c and is used by all programs
9039 that support '-rand'.
9042 *) In RAND_write_file, use mode 0600 for creating files;
9043 don't just chmod when it may be too late.
9046 *) Report an error from X509_STORE_load_locations
9047 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9050 *) New function ASN1_mbstring_copy() this copies a string in either
9051 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9052 into an ASN1_STRING type. A mask of permissible types is passed
9053 and it chooses the "minimal" type to use or an error if not type
9057 *) Add function equivalents to the various macros in asn1.h. The old
9058 macros are retained with an M_ prefix. Code inside the library can
9059 use the M_ macros. External code (including the openssl utility)
9060 should *NOT* in order to be "shared library friendly".
9063 *) Add various functions that can check a certificate's extensions
9064 to see if it usable for various purposes such as SSL client,
9065 server or S/MIME and CAs of these types. This is currently
9066 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9067 verification. Also added a -purpose flag to x509 utility to
9068 print out all the purposes.
9071 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9075 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9076 for, obtain and decode and extension and obtain its critical flag.
9077 This allows all the necessary extension code to be handled in a
9078 single function call.
9081 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9082 platforms. See crypto/rc4/rc4_enc.c for further details.
9085 *) New -noout option to asn1parse. This causes no output to be produced
9086 its main use is when combined with -strparse and -out to extract data
9087 from a file (which may not be in ASN.1 format).
9090 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9091 when producing the local key id.
9092 [Richard Levitte <levitte@stacken.kth.se>]
9094 *) New option -dhparam in s_server. This allows a DH parameter file to be
9095 stated explicitly. If it is not stated then it tries the first server
9096 certificate file. The previous behaviour hard coded the filename
9100 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9101 a public key to be input or output. For example:
9102 openssl rsa -in key.pem -pubout -out pubkey.pem
9103 Also added necessary DSA public key functions to handle this.
9106 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9107 in the message. This was handled by allowing
9108 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9109 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9111 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9112 to the end of the strings whereas this didn't. This would cause problems
9113 if strings read with d2i_ASN1_bytes() were later modified.
9114 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9116 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9117 data and it contains EOF it will end up returning an error. This is
9118 caused by input 46 bytes long. The cause is due to the way base64
9119 BIOs find the start of base64 encoded data. They do this by trying a
9120 trial decode on each line until they find one that works. When they
9121 do a flag is set and it starts again knowing it can pass all the
9122 data directly through the decoder. Unfortunately it doesn't reset
9123 the context it uses. This means that if EOF is reached an attempt
9124 is made to pass two EOFs through the context and this causes the
9125 resulting error. This can also cause other problems as well. As is
9126 usual with these problems it takes *ages* to find and the fix is
9127 trivial: move one line.
9128 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9130 *) Ugly workaround to get s_client and s_server working under Windows. The
9131 old code wouldn't work because it needed to select() on sockets and the
9132 tty (for keypresses and to see if data could be written). Win32 only
9133 supports select() on sockets so we select() with a 1s timeout on the
9134 sockets and then see if any characters are waiting to be read, if none
9135 are present then we retry, we also assume we can always write data to
9136 the tty. This isn't nice because the code then blocks until we've
9137 received a complete line of data and it is effectively polling the
9138 keyboard at 1s intervals: however it's quite a bit better than not
9139 working at all :-) A dedicated Windows application might handle this
9140 with an event loop for example.
9143 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9144 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9145 will be called when RSA_sign() and RSA_verify() are used. This is useful
9146 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9147 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9148 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9149 This necessitated the support of an extra signature type NID_md5_sha1
9150 for SSL signatures and modifications to the SSL library to use it instead
9151 of calling RSA_public_decrypt() and RSA_private_encrypt().
9154 *) Add new -verify -CAfile and -CApath options to the crl program, these
9155 will lookup a CRL issuers certificate and verify the signature in a
9156 similar way to the verify program. Tidy up the crl program so it
9157 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9158 less strict. It will now permit CRL extensions even if it is not
9159 a V2 CRL: this will allow it to tolerate some broken CRLs.
9162 *) Initialize all non-automatic variables each time one of the openssl
9163 sub-programs is started (this is necessary as they may be started
9164 multiple times from the "OpenSSL>" prompt).
9165 [Lennart Bang, Bodo Moeller]
9167 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9168 removing all other RSA functionality (this is what NO_RSA does). This
9169 is so (for example) those in the US can disable those operations covered
9170 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9174 *) Non-copying interface to BIO pairs.
9175 (still largely untested)
9178 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9179 ASCII string. This was handled independently in various places before.
9182 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9183 UTF8 strings a character at a time.
9186 *) Use client_version from client hello to select the protocol
9187 (s23_srvr.c) and for RSA client key exchange verification
9188 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9191 *) Add various utility functions to handle SPKACs, these were previously
9192 handled by poking round in the structure internals. Added new function
9193 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9194 print, verify and generate SPKACs. Based on an original idea from
9195 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9198 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9201 *) Allow the config file extension section to be overwritten on the
9202 command line. Based on an original idea from Massimiliano Pala
9203 <madwolf@comune.modena.it>. The new option is called -extensions
9204 and can be applied to ca, req and x509. Also -reqexts to override
9205 the request extensions in req and -crlexts to override the crl extensions
9209 *) Add new feature to the SPKAC handling in ca. Now you can include
9210 the same field multiple times by preceding it by "XXXX." for example:
9213 this is the same syntax as used in the req config file.
9216 *) Allow certificate extensions to be added to certificate requests. These
9217 are specified in a 'req_extensions' option of the req section of the
9218 config file. They can be printed out with the -text option to req but
9219 are otherwise ignored at present.
9222 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9223 data read consists of only the final block it would not decrypted because
9224 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9225 A misplaced 'break' also meant the decrypted final block might not be
9226 copied until the next read.
9229 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9230 a few extra parameters to the DH structure: these will be useful if
9231 for example we want the value of 'q' or implement X9.42 DH.
9234 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9235 provides hooks that allow the default DSA functions or functions on a
9236 "per key" basis to be replaced. This allows hardware acceleration and
9237 hardware key storage to be handled without major modification to the
9238 library. Also added low level modexp hooks and CRYPTO_EX structure and
9239 associated functions.
9242 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9243 as "read only": it can't be written to and the buffer it points to will
9244 not be freed. Reading from a read only BIO is much more efficient than
9245 a normal memory BIO. This was added because there are several times when
9246 an area of memory needs to be read from a BIO. The previous method was
9247 to create a memory BIO and write the data to it, this results in two
9248 copies of the data and an O(n^2) reading algorithm. There is a new
9249 function BIO_new_mem_buf() which creates a read only memory BIO from
9250 an area of memory. Also modified the PKCS#7 routines to use read only
9254 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9255 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9256 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9257 but a retry condition occured while trying to read the rest.
9260 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9261 NID_pkcs7_encrypted by default: this was wrong since this should almost
9262 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9263 the encrypted data type: this is a more sensible place to put it and it
9264 allows the PKCS#12 code to be tidied up that duplicated this
9268 *) Changed obj_dat.pl script so it takes its input and output files on
9269 the command line. This should avoid shell escape redirection problems
9273 *) Initial support for certificate extension requests, these are included
9274 in things like Xenroll certificate requests. Included functions to allow
9275 extensions to be obtained and added.
9278 *) -crlf option to s_client and s_server for sending newlines as
9279 CRLF (as required by many protocols).
9282 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9284 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9285 [Ralf S. Engelschall]
9287 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9288 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9290 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9294 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9295 DH parameters/keys (q is lost during that conversion, but the resulting
9296 DH parameters contain its length).
9298 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9299 much faster than DH_generate_parameters (which creates parameters
9300 where p = 2*q + 1), and also the smaller q makes DH computations
9301 much more efficient (160-bit exponentiation instead of 1024-bit
9302 exponentiation); so this provides a convenient way to support DHE
9303 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9304 utter importance to use
9305 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9307 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9308 when such DH parameters are used, because otherwise small subgroup
9309 attacks may become possible!
9312 *) Avoid memory leak in i2d_DHparams.
9315 *) Allow the -k option to be used more than once in the enc program:
9316 this allows the same encrypted message to be read by multiple recipients.
9319 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9320 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9321 it will always use the numerical form of the OID, even if it has a short
9325 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9326 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9327 otherwise bn_mod_exp was called. In the case of hardware keys for example
9328 no private key components need be present and it might store extra data
9329 in the RSA structure, which cannot be accessed from bn_mod_exp.
9330 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9331 private key operations.
9334 *) Added support for SPARC Linux.
9337 *) pem_password_cb function type incompatibly changed from
9338 typedef int pem_password_cb(char *buf, int size, int rwflag);
9340 ....(char *buf, int size, int rwflag, void *userdata);
9341 so that applications can pass data to their callbacks:
9342 The PEM[_ASN1]_{read,write}... functions and macros now take an
9343 additional void * argument, which is just handed through whenever
9344 the password callback is called.
9345 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9347 New function SSL_CTX_set_default_passwd_cb_userdata.
9349 Compatibility note: As many C implementations push function arguments
9350 onto the stack in reverse order, the new library version is likely to
9351 interoperate with programs that have been compiled with the old
9352 pem_password_cb definition (PEM_whatever takes some data that
9353 happens to be on the stack as its last argument, and the callback
9354 just ignores this garbage); but there is no guarantee whatsoever that
9357 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9358 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9359 problems not only on Windows, but also on some Unix platforms.
9360 To avoid problematic command lines, these definitions are now in an
9361 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9362 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9365 *) MIPS III/IV assembler module is reimplemented.
9368 *) More DES library cleanups: remove references to srand/rand and
9369 delete an unused file.
9372 *) Add support for the the free Netwide assembler (NASM) under Win32,
9373 since not many people have MASM (ml) and it can be hard to obtain.
9374 This is currently experimental but it seems to work OK and pass all
9375 the tests. Check out INSTALL.W32 for info.
9378 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9379 without temporary keys kept an extra copy of the server key,
9380 and connections with temporary keys did not free everything in case
9384 *) New function RSA_check_key and new openssl rsa option -check
9385 for verifying the consistency of RSA keys.
9386 [Ulf Moeller, Bodo Moeller]
9388 *) Various changes to make Win32 compile work:
9389 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9390 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9391 comparison" warnings.
9392 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9395 *) Add a debugging option to PKCS#5 v2 key generation function: when
9396 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9397 derived keys are printed to stderr.
9400 *) Copy the flags in ASN1_STRING_dup().
9401 [Roman E. Pavlov <pre@mo.msk.ru>]
9403 *) The x509 application mishandled signing requests containing DSA
9404 keys when the signing key was also DSA and the parameters didn't match.
9406 It was supposed to omit the parameters when they matched the signing key:
9407 the verifying software was then supposed to automatically use the CA's
9408 parameters if they were absent from the end user certificate.
9410 Omitting parameters is no longer recommended. The test was also
9411 the wrong way round! This was probably due to unusual behaviour in
9412 EVP_cmp_parameters() which returns 1 if the parameters match.
9413 This meant that parameters were omitted when they *didn't* match and
9414 the certificate was useless. Certificates signed with 'ca' didn't have
9416 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9418 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9419 The interface is as follows:
9420 Applications can use
9421 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9422 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9423 "off" is now the default.
9424 The library internally uses
9425 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9426 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9427 to disable memory-checking temporarily.
9429 Some inconsistent states that previously were possible (and were
9430 even the default) are now avoided.
9432 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9433 with each memory chunk allocated; this is occasionally more helpful
9434 than just having a counter.
9436 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9438 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9442 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9443 which largely parallels "options", but is for changing API behaviour,
9444 whereas "options" are about protocol behaviour.
9445 Initial "mode" flags are:
9447 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9448 a single record has been written.
9449 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9450 retries use the same buffer location.
9451 (But all of the contents must be
9455 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9458 *) Fix problems with no-hmac etc.
9459 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9461 *) New functions RSA_get_default_method(), RSA_set_method() and
9462 RSA_get_method(). These allows replacement of RSA_METHODs without having
9463 to mess around with the internals of an RSA structure.
9466 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9467 Also really enable memory leak checks in openssl.c and in some
9469 [Chad C. Mulligan, Bodo Moeller]
9471 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9472 up the length of negative integers. This has now been simplified to just
9473 store the length when it is first determined and use it later, rather
9474 than trying to keep track of where data is copied and updating it to
9476 [Steve Henson, reported by Brien Wheeler
9477 <bwheeler@authentica-security.com>]
9479 *) Add a new function PKCS7_signatureVerify. This allows the verification
9480 of a PKCS#7 signature but with the signing certificate passed to the
9481 function itself. This contrasts with PKCS7_dataVerify which assumes the
9482 certificate is present in the PKCS#7 structure. This isn't always the
9483 case: certificates can be omitted from a PKCS#7 structure and be
9484 distributed by "out of band" means (such as a certificate database).
9487 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9488 function prototypes in pem.h, also change util/mkdef.pl to add the
9489 necessary function names.
9492 *) mk1mf.pl (used by Windows builds) did not properly read the
9493 options set by Configure in the top level Makefile, and Configure
9494 was not even able to write more than one option correctly.
9495 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9498 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9499 file to be loaded from a BIO or FILE pointer. The BIO version will
9500 for example allow memory BIOs to contain config info.
9503 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9504 Whoever hopes to achieve shared-library compatibility across versions
9505 must use this, not the compile-time macro.
9506 (Exercise 0.9.4: Which is the minimum library version required by
9508 Note: All this applies only to multi-threaded programs, others don't
9512 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9513 through a BIO pair triggered the default case, i.e.
9514 SSLerr(...,SSL_R_UNKNOWN_STATE).
9517 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9518 can use the SSL library even if none of the specific BIOs is
9522 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9523 for the encoded length.
9524 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9526 *) Add initial documentation of the X509V3 functions.
9529 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9530 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9531 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9532 secure PKCS#8 private key format with a high iteration count.
9535 *) Fix determination of Perl interpreter: A perl or perl5
9536 _directory_ in $PATH was also accepted as the interpreter.
9537 [Ralf S. Engelschall]
9539 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9540 wrong with it but it was very old and did things like calling
9541 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9545 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9546 to use the new extension code.
9549 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9550 with macros. This should make it easier to change their form, add extra
9551 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9555 *) Add to configuration table a new entry that can specify an alternative
9556 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9557 according to Mark Crispin <MRC@Panda.COM>.
9561 *) DES CBC did not update the IV. Weird.
9564 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9565 Changing the behaviour of the former might break existing programs --
9566 where IV updating is needed, des_ncbc_encrypt can be used.
9569 *) When bntest is run from "make test" it drives bc to check its
9570 calculations, as well as internally checking them. If an internal check
9571 fails, it needs to cause bc to give a non-zero result or make test carries
9572 on without noticing the failure. Fixed.
9575 *) DES library cleanups.
9578 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9579 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9580 ciphers. NOTE: although the key derivation function has been verified
9581 against some published test vectors it has not been extensively tested
9582 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9586 *) Instead of "mkdir -p", which is not fully portable, use new
9587 Perl script "util/mkdir-p.pl".
9590 *) Rewrite the way password based encryption (PBE) is handled. It used to
9591 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9592 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9593 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9594 the 'parameter' field of the AlgorithmIdentifier is passed to the
9595 underlying key generation function so it must do its own ASN1 parsing.
9596 This has also changed the EVP_PBE_CipherInit() function which now has a
9597 'parameter' argument instead of literal salt and iteration count values
9598 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9601 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9602 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9603 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9604 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9605 value was just used as a "magic string" and not used directly its
9606 value doesn't matter.
9609 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9613 *) "linux-sparc64" configuration (ultrapenguin).
9614 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9615 "linux-sparc" configuration.
9616 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9618 *) config now generates no-xxx options for missing ciphers.
9621 *) Support the EBCDIC character set (work in progress).
9622 File ebcdic.c not yet included because it has a different license.
9623 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9625 *) Support BS2000/OSD-POSIX.
9626 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9628 *) Make callbacks for key generation use void * instead of char *.
9631 *) Make S/MIME samples compile (not yet tested).
9634 *) Additional typesafe stacks.
9637 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9641 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9643 *) New configuration variant "sco5-gcc".
9645 *) Updated some demos.
9646 [Sean O Riordain, Wade Scholine]
9648 *) Add missing BIO_free at exit of pkcs12 application.
9651 *) Fix memory leak in conf.c.
9654 *) Updates for Win32 to assembler version of MD5.
9657 *) Set #! path to perl in apps/der_chop to where we found it
9658 instead of using a fixed path.
9661 *) SHA library changes for irix64-mips4-cc.
9664 *) Improvements for VMS support.
9668 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9670 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9671 This also avoids the problems with SC4.2 and unpatched SC5.
9672 [Andy Polyakov <appro@fy.chalmers.se>]
9674 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9675 These are required because of the typesafe stack would otherwise break
9676 existing code. If old code used a structure member which used to be STACK
9677 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9678 sk_num or sk_value it would produce an error because the num, data members
9679 are not present in STACK_OF. Now it just produces a warning. sk_set
9680 replaces the old method of assigning a value to sk_value
9681 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9682 that does this will no longer work (and should use sk_set instead) but
9683 this could be regarded as a "questionable" behaviour anyway.
9686 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9687 correctly handle encrypted S/MIME data.
9690 *) Change type of various DES function arguments from des_cblock
9691 (which means, in function argument declarations, pointer to char)
9692 to des_cblock * (meaning pointer to array with 8 char elements),
9693 which allows the compiler to do more typechecking; it was like
9694 that back in SSLeay, but with lots of ugly casts.
9696 Introduce new type const_des_cblock.
9699 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9700 problems: find RecipientInfo structure that matches recipient certificate
9701 and initialise the ASN1 structures properly based on passed cipher.
9704 *) Belatedly make the BN tests actually check the results.
9707 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9708 to and from BNs: it was completely broken. New compilation option
9709 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9710 key elements as negative integers.
9713 *) Reorganize and speed up MD5.
9714 [Andy Polyakov <appro@fy.chalmers.se>]
9717 [Richard Levitte <richard@levitte.org>]
9719 *) New option -out to asn1parse to allow the parsed structure to be
9720 output to a file. This is most useful when combined with the -strparse
9721 option to examine the output of things like OCTET STRINGS.
9724 *) Make SSL library a little more fool-proof by not requiring any longer
9725 that SSL_set_{accept,connect}_state be called before
9726 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9727 in many applications because usually everything *appeared* to work as
9728 intended anyway -- now it really works as intended).
9731 *) Move openssl.cnf out of lib/.
9734 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9735 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9736 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9737 [Ralf S. Engelschall]
9739 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9740 handle PKCS#7 enveloped data properly.
9741 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9743 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9744 copying pointers. The cert_st handling is changed by this in
9745 various ways (and thus what used to be known as ctx->default_cert
9746 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9747 any longer when s->cert does not give us what we need).
9748 ssl_cert_instantiate becomes obsolete by this change.
9749 As soon as we've got the new code right (possibly it already is?),
9750 we have solved a couple of bugs of the earlier code where s->cert
9751 was used as if it could not have been shared with other SSL structures.
9753 Note that using the SSL API in certain dirty ways now will result
9754 in different behaviour than observed with earlier library versions:
9755 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9756 does not influence s as it used to.
9758 In order to clean up things more thoroughly, inside SSL_SESSION
9759 we don't use CERT any longer, but a new structure SESS_CERT
9760 that holds per-session data (if available); currently, this is
9761 the peer's certificate chain and, for clients, the server's certificate
9762 and temporary key. CERT holds only those values that can have
9763 meaningful defaults in an SSL_CTX.
9766 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9767 from the internal representation. Various PKCS#7 fixes: remove some
9768 evil casts and set the enc_dig_alg field properly based on the signing
9772 *) Allow PKCS#12 password to be set from the command line or the
9773 environment. Let 'ca' get its config file name from the environment
9774 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9778 *) Allow certificate policies extension to use an IA5STRING for the
9779 organization field. This is contrary to the PKIX definition but
9780 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9784 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9785 without disallowing inline assembler and the like for non-pedantic builds.
9788 *) Support Borland C++ builder.
9789 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9794 *) SHA-1 cleanups and performance enhancements.
9795 [Andy Polyakov <appro@fy.chalmers.se>]
9797 *) Sparc v8plus assembler for the bignum library.
9798 [Andy Polyakov <appro@fy.chalmers.se>]
9800 *) Accept any -xxx and +xxx compiler options in Configure.
9803 *) Update HPUX configuration.
9806 *) Add missing sk_<type>_unshift() function to safestack.h
9807 [Ralf S. Engelschall]
9809 *) New function SSL_CTX_use_certificate_chain_file that sets the
9810 "extra_cert"s in addition to the certificate. (This makes sense
9811 only for "PEM" format files, as chains as a whole are not
9815 *) Support verify_depth from the SSL API.
9816 x509_vfy.c had what can be considered an off-by-one-error:
9817 Its depth (which was not part of the external interface)
9818 was actually counting the number of certificates in a chain;
9819 now it really counts the depth.
9822 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9823 instead of X509err, which often resulted in confusing error
9824 messages since the error codes are not globally unique
9825 (e.g. an alleged error in ssl3_accept when a certificate
9826 didn't match the private key).
9828 *) New function SSL_CTX_set_session_id_context that allows to set a default
9829 value (so that you don't need SSL_set_session_id_context for each
9830 connection using the SSL_CTX).
9833 *) OAEP decoding bug fix.
9836 *) Support INSTALL_PREFIX for package builders, as proposed by
9840 *) New Configure options "threads" and "no-threads". For systems
9841 where the proper compiler options are known (currently Solaris
9842 and Linux), "threads" is the default.
9845 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9848 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9849 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9850 such as /usr/local/bin.
9853 *) "make linux-shared" to build shared libraries.
9854 [Niels Poppe <niels@netbox.org>]
9856 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9859 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9860 extension adding in x509 utility.
9863 *) Remove NOPROTO sections and error code comments.
9866 *) Partial rewrite of the DEF file generator to now parse the ANSI
9870 *) New Configure options --prefix=DIR and --openssldir=DIR.
9873 *) Complete rewrite of the error code script(s). It is all now handled
9874 by one script at the top level which handles error code gathering,
9875 header rewriting and C source file generation. It should be much better
9876 than the old method: it now uses a modified version of Ulf's parser to
9877 read the ANSI prototypes in all header files (thus the old K&R definitions
9878 aren't needed for error creation any more) and do a better job of
9879 translating function codes into names. The old 'ASN1 error code imbedded
9880 in a comment' is no longer necessary and it doesn't use .err files which
9881 have now been deleted. Also the error code call doesn't have to appear all
9882 on one line (which resulted in some large lines...).
9885 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9888 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9889 0 (which usually indicates a closed connection), but continue reading.
9892 *) Fix some race conditions.
9895 *) Add support for CRL distribution points extension. Add Certificate
9896 Policies and CRL distribution points documentation.
9899 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9902 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9903 8 of keying material. Merlin has also confirmed interop with this fix
9904 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9905 [Merlin Hughes <merlin@baltimore.ie>]
9907 *) Fix lots of warnings.
9908 [Richard Levitte <levitte@stacken.kth.se>]
9910 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9911 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9912 [Richard Levitte <levitte@stacken.kth.se>]
9914 *) Fix problems with sizeof(long) == 8.
9915 [Andy Polyakov <appro@fy.chalmers.se>]
9917 *) Change functions to ANSI C.
9920 *) Fix typos in error codes.
9921 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9923 *) Remove defunct assembler files from Configure.
9926 *) SPARC v8 assembler BIGNUM implementation.
9927 [Andy Polyakov <appro@fy.chalmers.se>]
9929 *) Support for Certificate Policies extension: both print and set.
9930 Various additions to support the r2i method this uses.
9933 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9934 return a const string when you are expecting an allocated buffer.
9937 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9938 types DirectoryString and DisplayText.
9941 *) Add code to allow r2i extensions to access the configuration database,
9942 add an LHASH database driver and add several ctx helper functions.
9945 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9946 fail when they extended the size of a BIGNUM.
9949 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9950 support typesafe stack.
9953 *) Fix typo in SSL_[gs]et_options().
9954 [Nils Frostberg <nils@medcom.se>]
9956 *) Delete various functions and files that belonged to the (now obsolete)
9957 old X509V3 handling code.
9960 *) New Configure option "rsaref".
9963 *) Don't auto-generate pem.h.
9966 *) Introduce type-safe ASN.1 SETs.
9969 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9970 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9972 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9973 that links with OpenSSL (well at least cause lots of warnings), but fear
9974 not: the conversion is trivial, and it eliminates loads of evil casts. A
9975 few STACKed things have been converted already. Feel free to convert more.
9976 In the fullness of time, I'll do away with the STACK type altogether.
9979 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9980 specified in <certfile> by updating the entry in the index.txt file.
9981 This way one no longer has to edit the index.txt file manually for
9982 revoking a certificate. The -revoke option does the gory details now.
9983 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9985 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9986 `-text' option at all and this way the `-noout -text' combination was
9987 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9988 [Ralf S. Engelschall]
9990 *) Make sure a corresponding plain text error message exists for the
9991 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9992 verify callback function determined that a certificate was revoked.
9993 [Ralf S. Engelschall]
9995 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9996 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9997 all available cipers including rc5, which was forgotten until now.
9998 In order to let the testing shell script know which algorithms
9999 are available, a new (up to now undocumented) command
10000 "openssl list-cipher-commands" is used.
10003 *) Bugfix: s_client occasionally would sleep in select() when
10004 it should have checked SSL_pending() first.
10007 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10008 the raw DSA values prior to ASN.1 encoding.
10011 *) Tweaks to Configure
10012 [Niels Poppe <niels@netbox.org>]
10014 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10018 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10021 *) New config option to avoid instructions that are illegal on the 80386.
10022 The default code is faster, but requires at least a 486.
10025 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10026 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10027 same as SSL2_VERSION anyway.
10030 *) New "-showcerts" option for s_client.
10033 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10034 application. Various cleanups and fixes.
10037 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10038 modify error routines to work internally. Add error codes and PBE init
10039 to library startup routines.
10042 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10043 packing functions to asn1 and evp. Changed function names and error
10044 codes along the way.
10047 *) PKCS12 integration: and so it begins... First of several patches to
10048 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10049 objects to objects.h
10052 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10053 and display support for Thawte strong extranet extension.
10056 *) Add LinuxPPC support.
10057 [Jeff Dubrule <igor@pobox.org>]
10059 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10060 bn_div_words in alpha.s.
10061 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10063 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10064 OAEP isn't supported when OpenSSL is built with RSAref.
10065 [Ulf Moeller <ulf@fitug.de>]
10067 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10068 so they no longer are missing under -DNOPROTO.
10069 [Soren S. Jorvang <soren@t.dk>]
10072 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10074 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10075 doesn't work when the session is reused. Coming soon!
10078 *) Fix a security hole, that allows sessions to be reused in the wrong
10079 context thus bypassing client cert protection! All software that uses
10080 client certs and session caches in multiple contexts NEEDS PATCHING to
10081 allow session reuse! A fuller solution is in the works.
10082 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10084 *) Some more source tree cleanups (removed obsolete files
10085 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10086 permission on "config" script to be executable) and a fix for the INSTALL
10088 [Ulf Moeller <ulf@fitug.de>]
10090 *) Remove some legacy and erroneous uses of malloc, free instead of
10092 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10094 *) Make rsa_oaep_test return non-zero on error.
10095 [Ulf Moeller <ulf@fitug.de>]
10097 *) Add support for native Solaris shared libraries. Configure
10098 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10099 if someone would make that last step automatic.
10100 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10102 *) ctx_size was not built with the right compiler during "make links". Fixed.
10105 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10106 except NULL ciphers". This means the default cipher list will no longer
10107 enable NULL ciphers. They need to be specifically enabled e.g. with
10108 the string "DEFAULT:eNULL".
10111 *) Fix to RSA private encryption routines: if p < q then it would
10112 occasionally produce an invalid result. This will only happen with
10113 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10116 *) Be less restrictive and allow also `perl util/perlpath.pl
10117 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10118 because this way one can also use an interpreter named `perl5' (which is
10119 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10120 installed as `perl').
10121 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10123 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10124 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10126 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10127 advapi32.lib to Win32 build and change the pem test comparision
10128 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10129 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10130 and crypto/des/ede_cbcm_enc.c.
10133 *) DES quad checksum was broken on big-endian architectures. Fixed.
10136 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10137 Win32 test batch file so it (might) work again. The Win32 test batch file
10138 is horrible: I feel ill....
10141 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10142 in e_os.h. Audit of header files to check ANSI and non ANSI
10143 sections: 10 functions were absent from non ANSI section and not exported
10144 from Windows DLLs. Fixed up libeay.num for new functions.
10147 *) Make `openssl version' output lines consistent.
10148 [Ralf S. Engelschall]
10150 *) Fix Win32 symbol export lists for BIO functions: Added
10151 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10152 to ms/libeay{16,32}.def.
10153 [Ralf S. Engelschall]
10155 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10156 fine under Unix and passes some trivial tests I've now added. But the
10157 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10158 added to make sure no one expects that this stuff really works in the
10159 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10160 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10162 [Ralf S. Engelschall]
10164 *) Fix the generation of two part addresses in perl.
10165 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10167 *) Add config entry for Linux on MIPS.
10168 [John Tobey <jtobey@channel1.com>]
10170 *) Make links whenever Configure is run, unless we are on Windoze.
10173 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10174 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10178 *) Add a useful kludge to allow package maintainers to specify compiler and
10179 other platforms details on the command line without having to patch the
10180 Configure script everytime: One now can use ``perl Configure
10181 <id>:<details>'', i.e. platform ids are allowed to have details appended
10182 to them (seperated by colons). This is treated as there would be a static
10183 pre-configured entry in Configure's %table under key <id> with value
10184 <details> and ``perl Configure <id>'' is called. So, when you want to
10185 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10186 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10187 now, which overrides the FreeBSD-elf entry on-the-fly.
10188 [Ralf S. Engelschall]
10190 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10193 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10194 on the `perl Configure ...' command line. This way one can compile
10195 OpenSSL libraries with Position Independent Code (PIC) which is needed
10196 for linking it into DSOs.
10197 [Ralf S. Engelschall]
10199 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10203 *) Cleaned up the LICENSE document: The official contact for any license
10204 questions now is the OpenSSL core team under openssl-core@openssl.org.
10205 And add a paragraph about the dual-license situation to make sure people
10206 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10207 to the OpenSSL toolkit.
10208 [Ralf S. Engelschall]
10210 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10211 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10212 Additonally cleaned up the `make links' target: Remove unnecessary
10213 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10214 to speed processing and no longer clutter the display with confusing
10215 stuff. Instead only the actually done links are displayed.
10216 [Ralf S. Engelschall]
10218 *) Permit null encryption ciphersuites, used for authentication only. It used
10219 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10220 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10224 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10225 signed attributes when verifying signatures (this would break them),
10226 the detached data encoding was wrong and public keys obtained using
10227 X509_get_pubkey() weren't freed.
10230 *) Add text documentation for the BUFFER functions. Also added a work around
10231 to a Win95 console bug. This was triggered by the password read stuff: the
10232 last character typed gets carried over to the next fread(). If you were
10233 generating a new cert request using 'req' for example then the last
10234 character of the passphrase would be CR which would then enter the first
10238 *) Added the new `Includes OpenSSL Cryptography Software' button as
10239 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10240 button and can be used by applications based on OpenSSL to show the
10241 relationship to the OpenSSL project.
10242 [Ralf S. Engelschall]
10244 *) Remove confusing variables in function signatures in files
10245 ssl/ssl_lib.c and ssl/ssl.h.
10246 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10248 *) Don't install bss_file.c under PREFIX/include/
10249 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10251 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10252 functions that return function pointers and has support for NT specific
10253 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10254 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10255 unsigned to signed types: this was killing the Win32 compile.
10258 *) Add new certificate file to stack functions,
10259 SSL_add_dir_cert_subjects_to_stack() and
10260 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10261 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10262 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10263 This means that Apache-SSL and similar packages don't have to mess around
10264 to add as many CAs as they want to the preferred list.
10267 *) Experiment with doxygen documentation. Currently only partially applied to
10269 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10270 openssl.doxy as the configuration file.
10273 *) Get rid of remaining C++-style comments which strict C compilers hate.
10274 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10276 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10277 compiled in by default: it has problems with large keys.
10280 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10281 DH private keys and/or callback functions which directly correspond to
10282 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10283 is needed for applications which have to configure certificates on a
10284 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10286 For the RSA certificate situation is makes no difference, but
10287 for the DSA certificate situation this fixes the "no shared cipher"
10288 problem where the OpenSSL cipher selection procedure failed because the
10289 temporary keys were not overtaken from the context and the API provided
10290 no way to reconfigure them.
10291 The new functions now let applications reconfigure the stuff and they
10292 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10293 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10294 non-public-API function ssl_cert_instantiate() is used as a helper
10295 function and also to reduce code redundancy inside ssl_rsa.c.
10296 [Ralf S. Engelschall]
10298 *) Move s_server -dcert and -dkey options out of the undocumented feature
10299 area because they are useful for the DSA situation and should be
10300 recognized by the users.
10301 [Ralf S. Engelschall]
10303 *) Fix the cipher decision scheme for export ciphers: the export bits are
10304 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10305 SSL_EXP_MASK. So, the original variable has to be used instead of the
10306 already masked variable.
10307 [Richard Levitte <levitte@stacken.kth.se>]
10309 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10310 [Richard Levitte <levitte@stacken.kth.se>]
10312 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10313 from `int' to `unsigned int' because it's a length and initialized by
10314 EVP_DigestFinal() which expects an `unsigned int *'.
10315 [Richard Levitte <levitte@stacken.kth.se>]
10317 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10318 script. Instead use the usual Shell->Perl transition trick.
10319 [Ralf S. Engelschall]
10321 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10322 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10323 -noout -modulus' as it's already the case for `openssl rsa -noout
10324 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10325 currently the public key is printed (a decision which was already done by
10326 `openssl dsa -modulus' in the past) which serves a similar purpose.
10327 Additionally the NO_RSA no longer completely removes the whole -modulus
10328 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10330 [Ralf S. Engelschall]
10332 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10333 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10334 [Arne Ansper <arne@ats.cyber.ee>]
10336 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10337 to be added. Now both 'req' and 'ca' can use new objects defined in the
10341 *) Add cool BIO that does syslog (or event log on NT).
10342 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10344 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10345 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10346 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10347 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10350 *) Add preliminary config info for new extension code.
10353 *) Make RSA_NO_PADDING really use no padding.
10354 [Ulf Moeller <ulf@fitug.de>]
10356 *) Generate errors when private/public key check is done.
10359 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10360 for some CRL extensions and new objects added.
10363 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10364 key usage extension and fuller support for authority key id.
10367 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10368 padding method for RSA, which is recommended for new applications in PKCS
10369 #1 v2.0 (RFC 2437, October 1998).
10370 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10371 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10372 against Bleichbacher's attack on RSA.
10373 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10376 *) Updates to the new SSL compression code
10377 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10379 *) Fix so that the version number in the master secret, when passed
10380 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10381 (because the server will not accept higher), that the version number
10382 is 0x03,0x01, not 0x03,0x00
10383 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10385 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10386 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10387 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10390 *) Support for RAW extensions where an arbitrary extension can be
10391 created by including its DER encoding. See apps/openssl.cnf for
10395 *) Make sure latest Perl versions don't interpret some generated C array
10396 code as Perl array code in the crypto/err/err_genc.pl script.
10397 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10399 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10400 not many people have the assembler. Various Win32 compilation fixes and
10401 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10402 build instructions.
10405 *) Modify configure script 'Configure' to automatically create crypto/date.h
10406 file under Win32 and also build pem.h from pem.org. New script
10407 util/mkfiles.pl to create the MINFO file on environments that can't do a
10408 'make files': perl util/mkfiles.pl >MINFO should work.
10411 *) Major rework of DES function declarations, in the pursuit of correctness
10412 and purity. As a result, many evil casts evaporated, and some weirdness,
10413 too. You may find this causes warnings in your code. Zapping your evil
10414 casts will probably fix them. Mostly.
10417 *) Fix for a typo in asn1.h. Bug fix to object creation script
10418 obj_dat.pl. It considered a zero in an object definition to mean
10419 "end of object": none of the objects in objects.h have any zeros
10420 so it wasn't spotted.
10421 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10423 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10424 Masking (CBCM). In the absence of test vectors, the best I have been able
10425 to do is check that the decrypt undoes the encrypt, so far. Send me test
10426 vectors if you have them.
10429 *) Correct calculation of key length for export ciphers (too much space was
10430 allocated for null ciphers). This has not been tested!
10433 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10434 message is now correct (it understands "crypto" and "ssl" on its
10435 command line). There is also now an "update" option. This will update
10436 the util/ssleay.num and util/libeay.num files with any new functions.
10438 perl util/mkdef.pl crypto ssl update
10439 it will update them.
10442 *) Overhauled the Perl interface (perl/*):
10443 - ported BN stuff to OpenSSL's different BN library
10444 - made the perl/ source tree CVS-aware
10445 - renamed the package from SSLeay to OpenSSL (the files still contain
10446 their history because I've copied them in the repository)
10447 - removed obsolete files (the test scripts will be replaced
10448 by better Test::Harness variants in the future)
10449 [Ralf S. Engelschall]
10451 *) First cut for a very conservative source tree cleanup:
10452 1. merge various obsolete readme texts into doc/ssleay.txt
10453 where we collect the old documents and readme texts.
10454 2. remove the first part of files where I'm already sure that we no
10455 longer need them because of three reasons: either they are just temporary
10456 files which were left by Eric or they are preserved original files where
10457 I've verified that the diff is also available in the CVS via "cvs diff
10458 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10459 the crypto/md/ stuff).
10460 [Ralf S. Engelschall]
10462 *) More extension code. Incomplete support for subject and issuer alt
10463 name, issuer and authority key id. Change the i2v function parameters
10464 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10465 what that's for :-) Fix to ASN1 macro which messed up
10466 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10469 *) Preliminary support for ENUMERATED type. This is largely copied from the
10473 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10474 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10476 *) Make sure `make rehash' target really finds the `openssl' program.
10477 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10479 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10480 like to hear about it if this slows down other processors.
10483 *) Add CygWin32 platform information to Configure script.
10484 [Alan Batie <batie@aahz.jf.intel.com>]
10486 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10487 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10489 *) New program nseq to manipulate netscape certificate sequences
10492 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10496 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10497 but the BN code had some problems that would cause failures when
10498 doing certificate verification and some other functions.
10499 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10501 *) Add ASN1 and PEM code to support netscape certificate sequences.
10504 *) Add ASN1 and PEM code to support netscape certificate sequences.
10507 *) Add several PKIX and private extended key usage OIDs.
10510 *) Modify the 'ca' program to handle the new extension code. Modify
10511 openssl.cnf for new extension format, add comments.
10514 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10515 and add a sample to openssl.cnf so req -x509 now adds appropriate
10519 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10520 error code, add initial support to X509_print() and x509 application.
10523 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10524 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10525 stuff is currently isolated and isn't even compiled yet.
10528 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10529 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10530 Removed the versions check from X509 routines when loading extensions:
10531 this allows certain broken certificates that don't set the version
10532 properly to be processed.
10535 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10536 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10537 can still be regenerated with "make depend".
10540 *) Spelling mistake in C version of CAST-128.
10541 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10543 *) Changes to the error generation code. The perl script err-code.pl
10544 now reads in the old error codes and retains the old numbers, only
10545 adding new ones if necessary. It also only changes the .err files if new
10546 codes are added. The makefiles have been modified to only insert errors
10547 when needed (to avoid needlessly modifying header files). This is done
10548 by only inserting errors if the .err file is newer than the auto generated
10549 C file. To rebuild all the error codes from scratch (the old behaviour)
10550 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10551 or delete all the .err files.
10554 *) CAST-128 was incorrectly implemented for short keys. The C version has
10555 been fixed, but is untested. The assembler versions are also fixed, but
10556 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10557 to regenerate it if needed.
10558 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10559 Hagino <itojun@kame.net>]
10561 *) File was opened incorrectly in randfile.c.
10562 [Ulf Möller <ulf@fitug.de>]
10564 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10565 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10566 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10567 al: it's just almost always a UTCTime. Note this patch adds new error
10568 codes so do a "make errors" if there are problems.
10571 *) Correct Linux 1 recognition in config.
10572 [Ulf Möller <ulf@fitug.de>]
10574 *) Remove pointless MD5 hash when using DSA keys in ca.
10575 [Anonymous <nobody@replay.com>]
10577 *) Generate an error if given an empty string as a cert directory. Also
10578 generate an error if handed NULL (previously returned 0 to indicate an
10579 error, but didn't set one).
10580 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10582 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10585 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10586 parameters. This was causing a warning which killed off the Win32 compile.
10589 *) Remove C++ style comments from crypto/bn/bn_local.h.
10590 [Neil Costigan <neil.costigan@celocom.com>]
10592 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10593 based on a text string, looking up short and long names and finally
10594 "dot" format. The "dot" format stuff didn't work. Added new function
10595 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10596 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10597 OID is not part of the table.
10600 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10601 X509_LOOKUP_by_alias().
10604 *) Sort openssl functions by name.
10607 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10608 encryption from sample DSA keys (in case anyone is interested the password
10612 *) Make _all_ *_free functions accept a NULL pointer.
10613 [Frans Heymans <fheymans@isaserver.be>]
10615 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10617 [Anonymous <nobody@replay.com>]
10619 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10620 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10622 *) Don't blow it for numeric -newkey arguments to apps/req.
10623 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10625 *) Temp key "for export" tests were wrong in s3_srvr.c.
10626 [Anonymous <nobody@replay.com>]
10628 *) Add prototype for temp key callback functions
10629 SSL_CTX_set_tmp_{rsa,dh}_callback().
10632 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10633 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10636 *) X509_name_add_entry() freed the wrong thing after an error.
10637 [Arne Ansper <arne@ats.cyber.ee>]
10639 *) rsa_eay.c would attempt to free a NULL context.
10640 [Arne Ansper <arne@ats.cyber.ee>]
10642 *) BIO_s_socket() had a broken should_retry() on Windoze.
10643 [Arne Ansper <arne@ats.cyber.ee>]
10645 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10646 [Arne Ansper <arne@ats.cyber.ee>]
10648 *) Make sure the already existing X509_STORE->depth variable is initialized
10649 in X509_STORE_new(), but document the fact that this variable is still
10650 unused in the certificate verification process.
10651 [Ralf S. Engelschall]
10653 *) Fix the various library and apps files to free up pkeys obtained from
10654 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10657 *) Fix reference counting in X509_PUBKEY_get(). This makes
10658 demos/maurice/example2.c work, amongst others, probably.
10659 [Steve Henson and Ben Laurie]
10661 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10662 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10663 are no longer created. This way we have a single and consistent command
10664 line interface `openssl <command>', similar to `cvs <command>'.
10665 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10667 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10668 BIT STRING wrapper always have zero unused bits.
10671 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10674 *) Make the top-level INSTALL documentation easier to understand.
10677 *) Makefiles updated to exit if an error occurs in a sub-directory
10678 make (including if user presses ^C) [Paul Sutton]
10680 *) Make Montgomery context stuff explicit in RSA data structure.
10683 *) Fix build order of pem and err to allow for generated pem.h.
10686 *) Fix renumbering bug in X509_NAME_delete_entry().
10689 *) Enhanced the err-ins.pl script so it makes the error library number
10690 global and can add a library name. This is needed for external ASN1 and
10691 other error libraries.
10694 *) Fixed sk_insert which never worked properly.
10697 *) Fix ASN1 macros so they can handle indefinite length construted
10698 EXPLICIT tags. Some non standard certificates use these: they can now
10702 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10703 into a single doc/ssleay.txt bundle. This way the information is still
10704 preserved but no longer messes up this directory. Now it's new room for
10705 the new set of documenation files.
10706 [Ralf S. Engelschall]
10708 *) SETs were incorrectly DER encoded. This was a major pain, because they
10709 shared code with SEQUENCEs, which aren't coded the same. This means that
10710 almost everything to do with SETs or SEQUENCEs has either changed name or
10711 number of arguments.
10712 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10714 *) Fix test data to work with the above.
10717 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10718 was already fixed by Eric for 0.9.1 it seems.
10719 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10721 *) Autodetect FreeBSD3.
10724 *) Fix various bugs in Configure. This affects the following platforms:
10728 unixware-2.0-pentium
10732 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10733 before they are needed.
10736 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10740 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10742 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10743 changed SSLeay to OpenSSL in version strings.
10744 [Ralf S. Engelschall]
10746 *) Some fixups to the top-level documents.
10749 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10750 because the symlink to include/ was missing.
10751 [Ralf S. Engelschall]
10753 *) Incorporated the popular no-RSA/DSA-only patches
10754 which allow to compile a RSA-free SSLeay.
10755 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10757 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10758 when "ssleay" is still not found.
10759 [Ralf S. Engelschall]
10761 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10762 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10764 *) Updated the README file.
10765 [Ralf S. Engelschall]
10767 *) Added various .cvsignore files in the CVS repository subdirs
10768 to make a "cvs update" really silent.
10769 [Ralf S. Engelschall]
10771 *) Recompiled the error-definition header files and added
10772 missing symbols to the Win32 linker tables.
10773 [Ralf S. Engelschall]
10775 *) Cleaned up the top-level documents;
10776 o new files: CHANGES and LICENSE
10777 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10778 o merged COPYRIGHT into LICENSE
10779 o removed obsolete TODO file
10780 o renamed MICROSOFT to INSTALL.W32
10781 [Ralf S. Engelschall]
10783 *) Removed dummy files from the 0.9.1b source tree:
10784 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10785 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10786 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10787 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10788 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10789 [Ralf S. Engelschall]
10791 *) Added various platform portability fixes.
10794 *) The Genesis of the OpenSSL rpject:
10795 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10796 Young and Tim J. Hudson created while they were working for C2Net until
10798 [The OpenSSL Project]
10801 Changes between 0.9.0b and 0.9.1b [not released]
10803 *) Updated a few CA certificates under certs/
10806 *) Changed some BIGNUM api stuff.
10809 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10810 DGUX x86, Linux Alpha, etc.
10813 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10814 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10818 *) Add -strparse option to asn1pars program which parses nested
10820 [Dr Stephen Henson <shenson@bigfoot.com>]
10822 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10825 *) DSA fix for "ca" program.
10828 *) Added "-genkey" option to "dsaparam" program.
10831 *) Added RIPE MD160 (rmd160) message digest.
10834 *) Added -a (all) option to "ssleay version" command.
10837 *) Added PLATFORM define which is the id given to Configure.
10840 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10843 *) Extended the ASN.1 parser routines.
10846 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10849 *) Added a BN_CTX to the BN library.
10852 *) Fixed the weak key values in DES library
10855 *) Changed API in EVP library for cipher aliases.
10858 *) Added support for RC2/64bit cipher.
10861 *) Converted the lhash library to the crypto/mem.c functions.
10864 *) Added more recognized ASN.1 object ids.
10867 *) Added more RSA padding checks for SSL/TLS.
10870 *) Added BIO proxy/filter functionality.
10873 *) Added extra_certs to SSL_CTX which can be used
10874 send extra CA certificates to the client in the CA cert chain sending
10875 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10878 *) Now Fortezza is denied in the authentication phase because
10879 this is key exchange mechanism is not supported by SSLeay at all.
10882 *) Additional PKCS1 checks.
10885 *) Support the string "TLSv1" for all TLS v1 ciphers.
10888 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10889 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10892 *) Fixed a few memory leaks.
10895 *) Fixed various code and comment typos.
10898 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10899 bytes sent in the client random.
10900 [Edward Bishop <ebishop@spyglass.com>]