5 Changes between 1.0.1s and 1.0.1t [3 May 2016]
7 *) Prevent padding oracle in AES-NI CBC MAC check
9 A MITM attacker can use a padding oracle attack to decrypt traffic
10 when the connection uses an AES CBC cipher and the server support
13 This issue was introduced as part of the fix for Lucky 13 padding
14 attack (CVE-2013-0169). The padding check was rewritten to be in
15 constant time by making sure that always the same bytes are read and
16 compared against either the MAC or padding bytes. But it no longer
17 checked that there was enough data to have both the MAC and padding
20 This issue was reported by Juraj Somorovsky using TLS-Attacker.
24 *) Fix EVP_EncodeUpdate overflow
26 An overflow can occur in the EVP_EncodeUpdate() function which is used for
27 Base64 encoding of binary data. If an attacker is able to supply very large
28 amounts of input data then a length check can overflow resulting in a heap
31 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
32 the PEM_write_bio* family of functions. These are mainly used within the
33 OpenSSL command line applications, so any application which processes data
34 from an untrusted source and outputs it as a PEM file should be considered
35 vulnerable to this issue. User applications that call these APIs directly
36 with large amounts of untrusted data may also be vulnerable.
38 This issue was reported by Guido Vranken.
42 *) Fix EVP_EncryptUpdate overflow
44 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
45 is able to supply very large amounts of input data after a previous call to
46 EVP_EncryptUpdate() with a partial block then a length check can overflow
47 resulting in a heap corruption. Following an analysis of all OpenSSL
48 internal usage of the EVP_EncryptUpdate() function all usage is one of two
49 forms. The first form is where the EVP_EncryptUpdate() call is known to be
50 the first called function after an EVP_EncryptInit(), and therefore that
51 specific call must be safe. The second form is where the length passed to
52 EVP_EncryptUpdate() can be seen from the code to be some small value and
53 therefore there is no possibility of an overflow. Since all instances are
54 one of these two forms, it is believed that there can be no overflows in
55 internal code due to this problem. It should be noted that
56 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
57 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
58 of these calls have also been analysed too and it is believed there are no
59 instances in internal usage where an overflow could occur.
61 This issue was reported by Guido Vranken.
65 *) Prevent ASN.1 BIO excessive memory allocation
67 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
68 a short invalid encoding can casuse allocation of large amounts of memory
69 potentially consuming excessive resources or exhausting memory.
71 Any application parsing untrusted data through d2i BIO functions is
72 affected. The memory based functions such as d2i_X509() are *not* affected.
73 Since the memory based functions are used by the TLS library, TLS
74 applications are not affected.
76 This issue was reported by Brian Carpenter.
82 ASN1 Strings that are over 1024 bytes can cause an overread in applications
83 using the X509_NAME_oneline() function on EBCDIC systems. This could result
84 in arbitrary stack data being returned in the buffer.
86 This issue was reported by Guido Vranken.
90 *) Modify behavior of ALPN to invoke callback after SNI/servername
91 callback, such that updates to the SSL_CTX affect ALPN.
94 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
98 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
99 methods are enabled and ssl2 is disabled the methods return NULL.
102 Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
104 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
105 Builds that are not configured with "enable-weak-ssl-ciphers" will not
106 provide any "EXPORT" or "LOW" strength ciphers.
109 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
110 is by default disabled at build-time. Builds that are not configured with
111 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
112 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
113 will need to explicitly call either of:
115 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
117 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
119 as appropriate. Even if either of those is used, or the application
120 explicitly uses the version-specific SSLv2_method() or its client and
121 server variants, SSLv2 ciphers vulnerable to exhaustive search key
122 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
123 ciphers, and SSLv2 56-bit DES are no longer available.
127 *) Fix a double-free in DSA code
129 A double free bug was discovered when OpenSSL parses malformed DSA private
130 keys and could lead to a DoS attack or memory corruption for applications
131 that receive DSA private keys from untrusted sources. This scenario is
134 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
139 *) Disable SRP fake user seed to address a server memory leak.
141 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
143 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
144 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
145 was changed to ignore the "fake user" SRP seed, even if the seed
148 Users should use SRP_VBASE_get1_by_user instead. Note that in
149 SRP_VBASE_get1_by_user, caller must free the returned value. Note
150 also that even though configuring the SRP seed attempts to hide
151 invalid usernames by continuing the handshake with fake
152 credentials, this behaviour is not constant time and no strong
153 guarantees are made that the handshake is indistinguishable from
154 that of a valid user.
158 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
160 In the BN_hex2bn function the number of hex digits is calculated using an
161 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
162 large values of |i| this can result in |bn_expand| not allocating any
163 memory because |i * 4| is negative. This can leave the internal BIGNUM data
164 field as NULL leading to a subsequent NULL ptr deref. For very large values
165 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
166 In this case memory is allocated to the internal BIGNUM data field, but it
167 is insufficiently sized leading to heap corruption. A similar issue exists
168 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
169 is ever called by user applications with very large untrusted hex/dec data.
170 This is anticipated to be a rare occurrence.
172 All OpenSSL internal usage of these functions use data that is not expected
173 to be untrusted, e.g. config file data or application command line
174 arguments. If user developed applications generate config file data based
175 on untrusted data then it is possible that this could also lead to security
176 consequences. This is also anticipated to be rare.
178 This issue was reported to OpenSSL by Guido Vranken.
182 *) Fix memory issues in BIO_*printf functions
184 The internal |fmtstr| function used in processing a "%s" format string in
185 the BIO_*printf functions could overflow while calculating the length of a
186 string and cause an OOB read when printing very long strings.
188 Additionally the internal |doapr_outch| function can attempt to write to an
189 OOB memory location (at an offset from the NULL pointer) in the event of a
190 memory allocation failure. In 1.0.2 and below this could be caused where
191 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
192 could be in processing a very long "%s" format string. Memory leaks can
195 The first issue may mask the second issue dependent on compiler behaviour.
196 These problems could enable attacks where large amounts of untrusted data
197 is passed to the BIO_*printf functions. If applications use these functions
198 in this way then they could be vulnerable. OpenSSL itself uses these
199 functions when printing out human-readable dumps of ASN.1 data. Therefore
200 applications that print this data could be vulnerable if the data is from
201 untrusted sources. OpenSSL command line applications could also be
202 vulnerable where they print out ASN.1 data, or if untrusted data is passed
203 as command line arguments.
205 Libssl is not considered directly vulnerable. Additionally certificates etc
206 received via remote connections via libssl are also unlikely to be able to
207 trigger these issues because of message size limits enforced within libssl.
209 This issue was reported to OpenSSL Guido Vranken.
213 *) Side channel attack on modular exponentiation
215 A side-channel attack was found which makes use of cache-bank conflicts on
216 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
217 of RSA keys. The ability to exploit this issue is limited as it relies on
218 an attacker who has control of code in a thread running on the same
219 hyper-threaded core as the victim thread which is performing decryptions.
221 This issue was reported to OpenSSL by Yuval Yarom, The University of
222 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
223 Nadia Heninger, University of Pennsylvania with more information at
224 http://cachebleed.info.
228 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
229 if no keysize is specified with default_bits. This fixes an
230 omission in an earlier change that changed all RSA/DSA key generation
231 apps to use 2048 bits by default.
234 Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
236 *) Protection for DH small subgroup attacks
238 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
239 switched on by default and cannot be disabled. This could have some
243 *) SSLv2 doesn't block disabled ciphers
245 A malicious client can negotiate SSLv2 ciphers that have been disabled on
246 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
247 been disabled, provided that the SSLv2 protocol was not also disabled via
250 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
251 and Sebastian Schinzel.
255 *) Reject DH handshakes with parameters shorter than 1024 bits.
258 Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
260 *) Certificate verify crash with missing PSS parameter
262 The signature verification routines will crash with a NULL pointer
263 dereference if presented with an ASN.1 signature using the RSA PSS
264 algorithm and absent mask generation function parameter. Since these
265 routines are used to verify certificate signature algorithms this can be
266 used to crash any certificate verification operation and exploited in a
267 DoS attack. Any application which performs certificate verification is
268 vulnerable including OpenSSL clients and servers which enable client
271 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
275 *) X509_ATTRIBUTE memory leak
277 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
278 memory. This structure is used by the PKCS#7 and CMS routines so any
279 application which reads PKCS#7 or CMS data from untrusted sources is
280 affected. SSL/TLS is not affected.
282 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
287 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
288 This changes the decoding behaviour for some invalid messages,
289 though the change is mostly in the more lenient direction, and
290 legacy behaviour is preserved as much as possible.
293 *) In DSA_generate_parameters_ex, if the provided seed is too short,
294 use a random seed, as already documented.
295 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
297 Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
299 *) Alternate chains certificate forgery
301 During certificate verfification, OpenSSL will attempt to find an
302 alternative certificate chain if the first attempt to build such a chain
303 fails. An error in the implementation of this logic can mean that an
304 attacker could cause certain checks on untrusted certificates to be
305 bypassed, such as the CA flag, enabling them to use a valid leaf
306 certificate to act as a CA and "issue" an invalid certificate.
308 This issue was reported to OpenSSL by Adam Langley/David Benjamin
313 *) Race condition handling PSK identify hint
315 If PSK identity hints are received by a multi-threaded client then
316 the values are wrongly updated in the parent SSL_CTX structure. This can
317 result in a race condition potentially leading to a double free of the
322 Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
323 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
324 incompatibility in the handling of HMAC. The previous ABI has now been
327 Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
329 *) Malformed ECParameters causes infinite loop
331 When processing an ECParameters structure OpenSSL enters an infinite loop
332 if the curve specified is over a specially malformed binary polynomial
335 This can be used to perform denial of service against any
336 system which processes public keys, certificate requests or
337 certificates. This includes TLS clients and TLS servers with
338 client authentication enabled.
340 This issue was reported to OpenSSL by Joseph Barr-Pixton.
344 *) Exploitable out-of-bounds read in X509_cmp_time
346 X509_cmp_time does not properly check the length of the ASN1_TIME
347 string and can read a few bytes out of bounds. In addition,
348 X509_cmp_time accepts an arbitrary number of fractional seconds in the
351 An attacker can use this to craft malformed certificates and CRLs of
352 various sizes and potentially cause a segmentation fault, resulting in
353 a DoS on applications that verify certificates or CRLs. TLS clients
354 that verify CRLs are affected. TLS clients and servers with client
355 authentication enabled may be affected if they use custom verification
358 This issue was reported to OpenSSL by Robert Swiecki (Google), and
359 independently by Hanno Böck.
363 *) PKCS7 crash with missing EnvelopedContent
365 The PKCS#7 parsing code does not handle missing inner EncryptedContent
366 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
367 with missing content and trigger a NULL pointer dereference on parsing.
369 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
370 structures from untrusted sources are affected. OpenSSL clients and
371 servers are not affected.
373 This issue was reported to OpenSSL by Michal Zalewski (Google).
377 *) CMS verify infinite loop with unknown hash function
379 When verifying a signedData message the CMS code can enter an infinite loop
380 if presented with an unknown hash function OID. This can be used to perform
381 denial of service against any system which verifies signedData messages using
383 This issue was reported to OpenSSL by Johannes Bauer.
387 *) Race condition handling NewSessionTicket
389 If a NewSessionTicket is received by a multi-threaded client when attempting to
390 reuse a previous ticket then a race condition can occur potentially leading to
391 a double free of the ticket data.
395 *) Reject DH handshakes with parameters shorter than 768 bits.
396 [Kurt Roeckx and Emilia Kasper]
398 *) dhparam: generate 2048-bit parameters by default.
399 [Kurt Roeckx and Emilia Kasper]
401 Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
403 *) Segmentation fault in ASN1_TYPE_cmp fix
405 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
406 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
407 certificate signature algorithm consistency this can be used to crash any
408 certificate verification operation and exploited in a DoS attack. Any
409 application which performs certificate verification is vulnerable including
410 OpenSSL clients and servers which enable client authentication.
414 *) ASN.1 structure reuse memory corruption fix
416 Reusing a structure in ASN.1 parsing may allow an attacker to cause
417 memory corruption via an invalid write. Such reuse is and has been
418 strongly discouraged and is believed to be rare.
420 Applications that parse structures containing CHOICE or ANY DEFINED BY
421 components may be affected. Certificate parsing (d2i_X509 and related
422 functions) are however not affected. OpenSSL clients and servers are
427 *) PKCS7 NULL pointer dereferences fix
429 The PKCS#7 parsing code does not handle missing outer ContentInfo
430 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
431 missing content and trigger a NULL pointer dereference on parsing.
433 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
434 otherwise parse PKCS#7 structures from untrusted sources are
435 affected. OpenSSL clients and servers are not affected.
437 This issue was reported to OpenSSL by Michal Zalewski (Google).
441 *) DoS via reachable assert in SSLv2 servers fix
443 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
444 servers that both support SSLv2 and enable export cipher suites by sending
445 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
447 This issue was discovered by Sean Burford (Google) and Emilia Käsper
448 (OpenSSL development team).
452 *) Use After Free following d2i_ECPrivatekey error fix
454 A malformed EC private key file consumed via the d2i_ECPrivateKey function
455 could cause a use after free condition. This, in turn, could cause a double
456 free in several private key parsing functions (such as d2i_PrivateKey
457 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
458 for applications that receive EC private keys from untrusted
459 sources. This scenario is considered rare.
461 This issue was discovered by the BoringSSL project and fixed in their
466 *) X509_to_X509_REQ NULL pointer deref fix
468 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
469 the certificate key is invalid. This function is rarely used in practice.
471 This issue was discovered by Brian Carpenter.
475 *) Removed the export ciphers from the DEFAULT ciphers
478 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
480 *) Build fixes for the Windows and OpenVMS platforms
481 [Matt Caswell and Richard Levitte]
483 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
485 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
486 message can cause a segmentation fault in OpenSSL due to a NULL pointer
487 dereference. This could lead to a Denial Of Service attack. Thanks to
488 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
492 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
493 dtls1_buffer_record function under certain conditions. In particular this
494 could occur if an attacker sent repeated DTLS records with the same
495 sequence number but for the next epoch. The memory leak could be exploited
496 by an attacker in a Denial of Service attack through memory exhaustion.
497 Thanks to Chris Mueller for reporting this issue.
501 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
502 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
503 method would be set to NULL which could later result in a NULL pointer
504 dereference. Thanks to Frank Schmirler for reporting this issue.
508 *) Abort handshake if server key exchange message is omitted for ephemeral
511 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
512 reporting this issue.
516 *) Remove non-export ephemeral RSA code on client and server. This code
517 violated the TLS standard by allowing the use of temporary RSA keys in
518 non-export ciphersuites and could be used by a server to effectively
519 downgrade the RSA key length used to a value smaller than the server
520 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
521 INRIA or reporting this issue.
525 *) Fixed issue where DH client certificates are accepted without verification.
526 An OpenSSL server will accept a DH certificate for client authentication
527 without the certificate verify message. This effectively allows a client to
528 authenticate without the use of a private key. This only affects servers
529 which trust a client certificate authority which issues certificates
530 containing DH keys: these are extremely rare and hardly ever encountered.
531 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
536 *) Ensure that the session ID context of an SSL is updated when its
537 SSL_CTX is updated via SSL_set_SSL_CTX.
539 The session ID context is typically set from the parent SSL_CTX,
540 and can vary with the CTX.
543 *) Fix various certificate fingerprint issues.
545 By using non-DER or invalid encodings outside the signed portion of a
546 certificate the fingerprint can be changed without breaking the signature.
547 Although no details of the signed portion of the certificate can be changed
548 this can cause problems with some applications: e.g. those using the
549 certificate fingerprint for blacklists.
551 1. Reject signatures with non zero unused bits.
553 If the BIT STRING containing the signature has non zero unused bits reject
554 the signature. All current signature algorithms require zero unused bits.
556 2. Check certificate algorithm consistency.
558 Check the AlgorithmIdentifier inside TBS matches the one in the
559 certificate signature. NB: this will result in signature failure
560 errors for some broken certificates.
562 Thanks to Konrad Kraszewski from Google for reporting this issue.
564 3. Check DSA/ECDSA signatures use DER.
566 Reencode DSA/ECDSA signatures and compare with the original received
567 signature. Return an error if there is a mismatch.
569 This will reject various cases including garbage after signature
570 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
571 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
572 (negative or with leading zeroes).
574 Further analysis was conducted and fixes were developed by Stephen Henson
575 of the OpenSSL core team.
580 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
581 results on some platforms, including x86_64. This bug occurs at random
582 with a very low probability, and is not known to be exploitable in any
583 way, though its exact impact is difficult to determine. Thanks to Pieter
584 Wuille (Blockstream) who reported this issue and also suggested an initial
585 fix. Further analysis was conducted by the OpenSSL development team and
586 Adam Langley of Google. The final fix was developed by Andy Polyakov of
587 the OpenSSL core team.
591 *) Do not resume sessions on the server if the negotiated protocol
592 version does not match the session's version. Resuming with a different
593 version, while not strictly forbidden by the RFC, is of questionable
594 sanity and breaks all known clients.
595 [David Benjamin, Emilia Käsper]
597 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
598 early CCS messages during renegotiation. (Note that because
599 renegotiation is encrypted, this early CCS was not exploitable.)
602 *) Tighten client-side session ticket handling during renegotiation:
603 ensure that the client only accepts a session ticket if the server sends
604 the extension anew in the ServerHello. Previously, a TLS client would
605 reuse the old extension state and thus accept a session ticket if one was
606 announced in the initial ServerHello.
608 Similarly, ensure that the client requires a session ticket if one
609 was advertised in the ServerHello. Previously, a TLS client would
610 ignore a missing NewSessionTicket message.
613 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
617 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
618 sends a carefully crafted handshake message, to cause OpenSSL to fail
619 to free up to 64k of memory causing a memory leak. This could be
620 exploited in a Denial Of Service attack. This issue affects OpenSSL
621 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
622 whether SRTP is used or configured. Implementations of OpenSSL that
623 have been compiled with OPENSSL_NO_SRTP defined are not affected.
625 The fix was developed by the OpenSSL team.
629 *) Session Ticket Memory Leak.
631 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
632 integrity of that ticket is first verified. In the event of a session
633 ticket integrity check failing, OpenSSL will fail to free memory
634 causing a memory leak. By sending a large number of invalid session
635 tickets an attacker could exploit this issue in a Denial Of Service
640 *) Build option no-ssl3 is incomplete.
642 When OpenSSL is configured with "no-ssl3" as a build option, servers
643 could accept and complete a SSL 3.0 handshake, and clients could be
644 configured to send them.
646 [Akamai and the OpenSSL team]
648 *) Add support for TLS_FALLBACK_SCSV.
649 Client applications doing fallback retries should call
650 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
652 [Adam Langley, Bodo Moeller]
654 *) Add additional DigestInfo checks.
656 Reencode DigestInto in DER and check against the original when
657 verifying RSA signature: this will reject any improperly encoded
658 DigestInfo structures.
660 Note: this is a precautionary measure and no attacks are currently known.
664 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
666 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
667 SRP code can be overrun an internal buffer. Add sanity check that
668 g, A, B < N to SRP code.
670 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
671 Group for discovering this issue.
675 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
676 TLS 1.0 instead of higher protocol versions when the ClientHello message
677 is badly fragmented. This allows a man-in-the-middle attacker to force a
678 downgrade to TLS 1.0 even if both the server and the client support a
679 higher protocol version, by modifying the client's TLS records.
681 Thanks to David Benjamin and Adam Langley (Google) for discovering and
682 researching this issue.
686 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
687 to a denial of service attack. A malicious server can crash the client
688 with a null pointer dereference (read) by specifying an anonymous (EC)DH
689 ciphersuite and sending carefully crafted handshake messages.
691 Thanks to Felix Gröbert (Google) for discovering and researching this
696 *) By sending carefully crafted DTLS packets an attacker could cause openssl
697 to leak memory. This can be exploited through a Denial of Service attack.
698 Thanks to Adam Langley for discovering and researching this issue.
702 *) An attacker can force openssl to consume large amounts of memory whilst
703 processing DTLS handshake messages. This can be exploited through a
704 Denial of Service attack.
705 Thanks to Adam Langley for discovering and researching this issue.
709 *) An attacker can force an error condition which causes openssl to crash
710 whilst processing DTLS packets due to memory being freed twice. This
711 can be exploited through a Denial of Service attack.
712 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
717 *) If a multithreaded client connects to a malicious server using a resumed
718 session and the server sends an ec point format extension it could write
719 up to 255 bytes to freed memory.
721 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
726 *) A malicious server can crash an OpenSSL client with a null pointer
727 dereference (read) by specifying an SRP ciphersuite even though it was not
728 properly negotiated with the client. This can be exploited through a
729 Denial of Service attack.
731 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
732 discovering and researching this issue.
736 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
737 X509_name_oneline, X509_name_print_ex et al. to leak some information
738 from the stack. Applications may be affected if they echo pretty printing
739 output to the attacker.
741 Thanks to Ivan Fratric (Google) for discovering this issue.
743 [Emilia Käsper, and Steve Henson]
745 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
746 for corner cases. (Certain input points at infinity could lead to
747 bogus results, with non-infinity inputs mapped to infinity too.)
750 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
752 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
753 handshake can force the use of weak keying material in OpenSSL
754 SSL/TLS clients and servers.
756 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
757 researching this issue. (CVE-2014-0224)
758 [KIKUCHI Masashi, Steve Henson]
760 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
761 OpenSSL DTLS client the code can be made to recurse eventually crashing
764 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
766 [Imre Rad, Steve Henson]
768 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
769 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
770 client or server. This is potentially exploitable to run arbitrary
771 code on a vulnerable client or server.
773 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
774 [Jüri Aedla, Steve Henson]
776 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
777 are subject to a denial of service attack.
779 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
780 this issue. (CVE-2014-3470)
781 [Felix Gröbert, Ivan Fratric, Steve Henson]
783 *) Harmonize version and its documentation. -f flag is used to display
785 [mancha <mancha1@zoho.com>]
787 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
789 [mancha <mancha1@zoho.com>]
791 *) Fix some double frees. These are not thought to be exploitable.
792 [mancha <mancha1@zoho.com>]
794 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
796 *) A missing bounds check in the handling of the TLS heartbeat extension
797 can be used to reveal up to 64k of memory to a connected client or
800 Thanks for Neel Mehta of Google Security for discovering this bug and to
801 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
802 preparing the fix (CVE-2014-0160)
803 [Adam Langley, Bodo Moeller]
805 *) Fix for the attack described in the paper "Recovering OpenSSL
806 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
807 by Yuval Yarom and Naomi Benger. Details can be obtained from:
808 http://eprint.iacr.org/2014/140
810 Thanks to Yuval Yarom and Naomi Benger for discovering this
811 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
812 [Yuval Yarom and Naomi Benger]
814 *) TLS pad extension: draft-agl-tls-padding-03
816 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
817 TLS client Hello record length value would otherwise be > 255 and
818 less that 512 pad with a dummy extension containing zeroes so it
819 is at least 512 bytes long.
821 [Adam Langley, Steve Henson]
823 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
825 *) Fix for TLS record tampering bug. A carefully crafted invalid
826 handshake could crash OpenSSL with a NULL pointer exception.
827 Thanks to Anton Johansson for reporting this issues.
830 *) Keep original DTLS digest and encryption contexts in retransmission
831 structures so we can use the previous session parameters if they need
832 to be resent. (CVE-2013-6450)
835 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
836 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
837 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
838 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
839 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
840 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
841 [Rob Stradling, Adam Langley]
843 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
845 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
846 supporting platforms or when small records were transferred.
847 [Andy Polyakov, Steve Henson]
849 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
851 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
853 This addresses the flaw in CBC record processing discovered by
854 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
855 at: http://www.isg.rhul.ac.uk/tls/
857 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
858 Security Group at Royal Holloway, University of London
859 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
860 Emilia Käsper for the initial patch.
862 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
864 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
865 ciphersuites which can be exploited in a denial of service attack.
866 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
867 and detecting this bug and to Wolfgang Ettlinger
868 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
872 *) Return an error when checking OCSP signatures when key is NULL.
873 This fixes a DoS attack. (CVE-2013-0166)
876 *) Make openssl verify return errors.
877 [Chris Palmer <palmer@google.com> and Ben Laurie]
879 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
880 the right response is stapled. Also change SSL_get_certificate()
881 so it returns the certificate actually sent.
882 See http://rt.openssl.org/Ticket/Display.html?id=2836.
883 [Rob Stradling <rob.stradling@comodo.com>]
885 *) Fix possible deadlock when decoding public keys.
888 *) Don't use TLS 1.0 record version number in initial client hello
892 Changes between 1.0.1b and 1.0.1c [10 May 2012]
894 *) Sanity check record length before skipping explicit IV in TLS
895 1.2, 1.1 and DTLS to fix DoS attack.
897 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
898 fuzzing as a service testing platform.
902 *) Initialise tkeylen properly when encrypting CMS messages.
903 Thanks to Solar Designer of Openwall for reporting this issue.
906 *) In FIPS mode don't try to use composite ciphers as they are not
910 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
912 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
913 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
914 mean any application compiled against OpenSSL 1.0.0 headers setting
915 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
916 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
917 0x10000000L Any application which was previously compiled against
918 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
919 will need to be recompiled as a result. Letting be results in
920 inability to disable specifically TLS 1.1 and in client context,
921 in unlike event, limit maximum offered version to TLS 1.0 [see below].
924 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
925 disable just protocol X, but all protocols above X *if* there are
926 protocols *below* X still enabled. In more practical terms it means
927 that if application wants to disable TLS1.0 in favor of TLS1.1 and
928 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
929 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
933 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
935 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
936 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
937 in CRYPTO_realloc_clean.
939 Thanks to Tavis Ormandy, Google Security Team, for discovering this
940 issue and to Adam Langley <agl@chromium.org> for fixing it.
942 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
944 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
947 *) Workarounds for some broken servers that "hang" if a client hello
948 record length exceeds 255 bytes.
950 1. Do not use record version number > TLS 1.0 in initial client
951 hello: some (but not all) hanging servers will now work.
952 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
953 the number of ciphers sent in the client hello. This should be
954 set to an even number, such as 50, for example by passing:
955 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
956 Most broken servers should now work.
957 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
958 TLS 1.2 client support entirely.
961 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
964 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
966 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
967 STRING form instead of a DigestInfo.
970 *) The format used for MDC2 RSA signatures is inconsistent between EVP
971 and the RSA_sign/RSA_verify functions. This was made more apparent when
972 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
973 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
974 the correct format in RSA_verify so both forms transparently work.
977 *) Some servers which support TLS 1.0 can choke if we initially indicate
978 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
979 encrypted premaster secret. As a workaround use the maximum pemitted
980 client version in client hello, this should keep such servers happy
981 and still work with previous versions of OpenSSL.
984 *) Add support for TLS/DTLS heartbeats.
985 [Robin Seggelmann <seggelmann@fh-muenster.de>]
987 *) Add support for SCTP.
988 [Robin Seggelmann <seggelmann@fh-muenster.de>]
990 *) Improved PRNG seeding for VOS.
991 [Paul Green <Paul.Green@stratus.com>]
993 *) Extensive assembler packs updates, most notably:
995 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
996 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
997 - x86_64: bit-sliced AES implementation;
998 - ARM: NEON support, contemporary platforms optimizations;
999 - s390x: z196 support;
1000 - *: GHASH and GF(2^m) multiplication implementations;
1004 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1005 (removal of unnecessary code)
1006 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1008 *) Add TLS key material exporter from RFC 5705.
1011 *) Add DTLS-SRTP negotiation from RFC 5764.
1014 *) Add Next Protocol Negotiation,
1015 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1016 disabled with a no-npn flag to config or Configure. Code donated
1018 [Adam Langley <agl@google.com> and Ben Laurie]
1020 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1021 NIST-P256, NIST-P521, with constant-time single point multiplication on
1022 typical inputs. Compiler support for the nonstandard type __uint128_t is
1023 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1024 Code made available under Apache License version 2.0.
1026 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1027 line to include this in your build of OpenSSL, and run "make depend" (or
1028 "make update"). This enables the following EC_METHODs:
1030 EC_GFp_nistp224_method()
1031 EC_GFp_nistp256_method()
1032 EC_GFp_nistp521_method()
1034 EC_GROUP_new_by_curve_name() will automatically use these (while
1035 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1037 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1039 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1040 all platforms. Move ssize_t definition from e_os.h to the public
1041 header file e_os2.h as it now appears in public header file cms.h
1044 *) New -sigopt option to the ca, req and x509 utilities. Additional
1045 signature parameters can be passed using this option and in
1049 *) Add RSA PSS signing function. This will generate and set the
1050 appropriate AlgorithmIdentifiers for PSS based on those in the
1051 corresponding EVP_MD_CTX structure. No application support yet.
1054 *) Support for companion algorithm specific ASN1 signing routines.
1055 New function ASN1_item_sign_ctx() signs a pre-initialised
1056 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1057 the appropriate parameters.
1060 *) Add new algorithm specific ASN1 verification initialisation function
1061 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1062 handling will be the same no matter what EVP_PKEY_METHOD is used.
1063 Add a PSS handler to support verification of PSS signatures: checked
1064 against a number of sample certificates.
1067 *) Add signature printing for PSS. Add PSS OIDs.
1068 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1070 *) Add algorithm specific signature printing. An individual ASN1 method
1071 can now print out signatures instead of the standard hex dump.
1073 More complex signatures (e.g. PSS) can print out more meaningful
1074 information. Include DSA version that prints out the signature
1078 *) Password based recipient info support for CMS library: implementing
1082 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1083 neatly separates the code into cipher and PBE sections and is required
1084 for some algorithms that split PBES2 into separate pieces (such as
1085 password based CMS).
1088 *) Session-handling fixes:
1089 - Fix handling of connections that are resuming with a session ID,
1090 but also support Session Tickets.
1091 - Fix a bug that suppressed issuing of a new ticket if the client
1092 presented a ticket with an expired session.
1093 - Try to set the ticket lifetime hint to something reasonable.
1094 - Make tickets shorter by excluding irrelevant information.
1095 - On the client side, don't ignore renewed tickets.
1096 [Adam Langley, Bodo Moeller (Google)]
1098 *) Fix PSK session representation.
1101 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1103 This work was sponsored by Intel.
1106 *) Add GCM support to TLS library. Some custom code is needed to split
1107 the IV between the fixed (from PRF) and explicit (from TLS record)
1108 portions. This adds all GCM ciphersuites supported by RFC5288 and
1109 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
1110 add a special AESGCM string for GCM only.
1113 *) Expand range of ctrls for AES GCM. Permit setting invocation
1114 field on decrypt and retrieval of invocation field only on encrypt.
1117 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1118 As required by RFC5289 these ciphersuites cannot be used if for
1119 versions of TLS earlier than 1.2.
1122 *) For FIPS capable OpenSSL interpret a NULL default public key method
1123 as unset and return the appopriate default but do *not* set the default.
1124 This means we can return the appopriate method in applications that
1125 swicth between FIPS and non-FIPS modes.
1128 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1129 ENGINE is used then we cannot handle that in the FIPS module so we
1130 keep original code iff non-FIPS operations are allowed.
1133 *) Add -attime option to openssl utilities.
1134 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1136 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1139 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1140 FIPS EC methods unconditionally for now.
1143 *) New build option no-ec2m to disable characteristic 2 code.
1146 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1147 all cases can be covered as some introduce binary incompatibilities.
1150 *) Redirect RSA operations to FIPS module including keygen,
1151 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1154 *) Add similar low level API blocking to ciphers.
1157 *) Low level digest APIs are not approved in FIPS mode: any attempt
1158 to use these will cause a fatal error. Applications that *really* want
1159 to use them can use the private_* version instead.
1162 *) Redirect cipher operations to FIPS module for FIPS builds.
1165 *) Redirect digest operations to FIPS module for FIPS builds.
1168 *) Update build system to add "fips" flag which will link in fipscanister.o
1169 for static and shared library builds embedding a signature if needed.
1172 *) Output TLS supported curves in preference order instead of numerical
1173 order. This is currently hardcoded for the highest order curves first.
1174 This should be configurable so applications can judge speed vs strength.
1177 *) Add TLS v1.2 server support for client authentication.
1180 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1184 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1185 FIPS modules versions.
1188 *) Add TLS v1.2 client side support for client authentication. Keep cache
1189 of handshake records longer as we don't know the hash algorithm to use
1190 until after the certificate request message is received.
1193 *) Initial TLS v1.2 client support. Add a default signature algorithms
1194 extension including all the algorithms we support. Parse new signature
1195 format in client key exchange. Relax some ECC signing restrictions for
1196 TLS v1.2 as indicated in RFC5246.
1199 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1200 to new signature format when needed using client digest preference.
1201 All server ciphersuites should now work correctly in TLS v1.2. No client
1202 support yet and no support for client certificates.
1205 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1206 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1207 ciphersuites. At present only RSA key exchange ciphersuites work with
1208 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1209 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1210 and version checking.
1213 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1214 with this defined it will not be affected by any changes to ssl internal
1215 structures. Add several utility functions to allow openssl application
1216 to work with OPENSSL_NO_SSL_INTERN defined.
1220 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1222 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1225 *) Permit abbreviated handshakes when renegotiating using the function
1226 SSL_renegotiate_abbreviated().
1227 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1229 *) Add call to ENGINE_register_all_complete() to
1230 ENGINE_load_builtin_engines(), so some implementations get used
1231 automatically instead of needing explicit application support.
1234 *) Add support for TLS key exporter as described in RFC5705.
1235 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1237 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1238 a few changes are required:
1240 Add SSL_OP_NO_TLSv1_1 flag.
1241 Add TLSv1_1 methods.
1242 Update version checking logic to handle version 1.1.
1243 Add explicit IV handling (ported from DTLS code).
1244 Add command line options to s_client/s_server.
1247 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1249 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1250 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1251 content decryption and always return the same error. Note: this attack
1252 needs on average 2^20 messages so it only affects automated senders. The
1253 old behaviour can be reenabled in the CMS code by setting the
1254 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1255 an MMA defence is not necessary.
1256 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1257 this issue. (CVE-2012-0884)
1260 *) Fix CVE-2011-4619: make sure we really are receiving a
1261 client hello before rejecting multiple SGC restarts. Thanks to
1262 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1265 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1267 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1268 Thanks to Antonio Martin, Enterprise Secure Access Research and
1269 Development, Cisco Systems, Inc. for discovering this bug and
1270 preparing a fix. (CVE-2012-0050)
1273 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1275 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1276 of the Vaudenay padding oracle attack on CBC mode encryption
1277 which enables an efficient plaintext recovery attack against
1278 the OpenSSL implementation of DTLS. Their attack exploits timing
1279 differences arising during decryption processing. A research
1280 paper describing this attack can be found at:
1281 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1282 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1283 Security Group at Royal Holloway, University of London
1284 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1285 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1286 for preparing the fix. (CVE-2011-4108)
1287 [Robin Seggelmann, Michael Tuexen]
1289 *) Clear bytes used for block padding of SSL 3.0 records.
1291 [Adam Langley (Google)]
1293 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1294 Kadianakis <desnacked@gmail.com> for discovering this issue and
1295 Adam Langley for preparing the fix. (CVE-2011-4619)
1296 [Adam Langley (Google)]
1298 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1299 [Andrey Kulikov <amdeich@gmail.com>]
1301 *) Prevent malformed RFC3779 data triggering an assertion failure.
1302 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1303 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1304 [Rob Austein <sra@hactrn.net>]
1306 *) Improved PRNG seeding for VOS.
1307 [Paul Green <Paul.Green@stratus.com>]
1309 *) Fix ssl_ciph.c set-up race.
1310 [Adam Langley (Google)]
1312 *) Fix spurious failures in ecdsatest.c.
1313 [Emilia Käsper (Google)]
1315 *) Fix the BIO_f_buffer() implementation (which was mixing different
1316 interpretations of the '..._len' fields).
1317 [Adam Langley (Google)]
1319 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1320 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1321 threads won't reuse the same blinding coefficients.
1323 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1324 lock to call BN_BLINDING_invert_ex, and avoids one use of
1325 BN_BLINDING_update for each BN_BLINDING structure (previously,
1326 the last update always remained unused).
1327 [Emilia Käsper (Google)]
1329 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1330 [Bob Buckholz (Google)]
1332 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1334 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1335 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1336 [Kaspar Brand <ossl@velox.ch>]
1338 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1339 for multi-threaded use of ECDH. (CVE-2011-3210)
1340 [Adam Langley (Google)]
1342 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1345 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1346 signature public key algorithm by using OID xref utilities instead.
1347 Before this you could only use some ECC ciphersuites with SHA1 only.
1350 *) Add protection against ECDSA timing attacks as mentioned in the paper
1351 by Billy Bob Brumley and Nicola Tuveri, see:
1353 http://eprint.iacr.org/2011/232.pdf
1355 [Billy Bob Brumley and Nicola Tuveri]
1357 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1359 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1360 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1362 *) Fix bug in string printing code: if *any* escaping is enabled we must
1363 escape the escape character (backslash) or the resulting string is
1367 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1369 *) Disable code workaround for ancient and obsolete Netscape browsers
1370 and servers: an attacker can use it in a ciphersuite downgrade attack.
1371 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1374 *) Fixed J-PAKE implementation error, originally discovered by
1375 Sebastien Martini, further info and confirmation from Stefan
1376 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1379 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1381 *) Fix extension code to avoid race conditions which can result in a buffer
1382 overrun vulnerability: resumed sessions must not be modified as they can
1383 be shared by multiple threads. CVE-2010-3864
1386 *) Fix WIN32 build system to correctly link an ENGINE directory into
1390 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1392 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1394 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1396 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1398 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1399 context. The operation can be customised via the ctrl mechanism in
1400 case ENGINEs want to include additional functionality.
1403 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1406 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1407 output hashes compatible with older versions of OpenSSL.
1408 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1410 *) Fix compression algorithm handling: if resuming a session use the
1411 compression algorithm of the resumed session instead of determining
1412 it from client hello again. Don't allow server to change algorithm.
1415 *) Add load_crls() function to apps tidying load_certs() too. Add option
1416 to verify utility to allow additional CRLs to be included.
1419 *) Update OCSP request code to permit adding custom headers to the request:
1420 some responders need this.
1423 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1425 [Julia Lawall <julia@diku.dk>]
1427 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1428 needlessly dereferenced structures, used obsolete functions and
1429 didn't handle all updated verify codes correctly.
1432 *) Disable MD2 in the default configuration.
1435 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1436 indicate the initial BIO being pushed or popped. This makes it possible
1437 to determine whether the BIO is the one explicitly called or as a result
1438 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1439 it handles reference counts correctly and doesn't zero out the I/O bio
1440 when it is not being explicitly popped. WARNING: applications which
1441 included workarounds for the old buggy behaviour will need to be modified
1442 or they could free up already freed BIOs.
1445 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1446 renaming to all platforms (within the 0.9.8 branch, this was
1447 done conditionally on Netware platforms to avoid a name clash).
1448 [Guenter <lists@gknw.net>]
1450 *) Add ECDHE and PSK support to DTLS.
1451 [Michael Tuexen <tuexen@fh-muenster.de>]
1453 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1457 *) Add "missing" function EVP_MD_flags() (without this the only way to
1458 retrieve a digest flags is by accessing the structure directly. Update
1459 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1460 or cipher is registered as in the "from" argument. Print out all
1461 registered digests in the dgst usage message instead of manually
1462 attempting to work them out.
1465 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1466 this allows the use of compression and extensions. Change default cipher
1467 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1468 by default unless an application cipher string requests it.
1471 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1472 key ids to find matching certificates and keys but some PKCS#12 files
1473 don't follow the (somewhat unwritten) rules and this strategy fails.
1474 Now just gather all certificates together and the first private key
1475 then look for the first certificate that matches the key.
1478 *) Support use of registered digest and cipher names for dgst and cipher
1479 commands instead of having to add each one as a special case. So now
1486 openssl dgst -sha256 foo
1488 and this works for ENGINE based algorithms too.
1492 *) Update Gost ENGINE to support parameter files.
1493 [Victor B. Wagner <vitus@cryptocom.ru>]
1495 *) Support GeneralizedTime in ca utility.
1496 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1498 *) Enhance the hash format used for certificate directory links. The new
1499 form uses the canonical encoding (meaning equivalent names will work
1500 even if they aren't identical) and uses SHA1 instead of MD5. This form
1501 is incompatible with the older format and as a result c_rehash should
1502 be used to rebuild symbolic links.
1505 *) Make PKCS#8 the default write format for private keys, replacing the
1506 traditional format. This form is standardised, more secure and doesn't
1507 include an implicit MD5 dependency.
1510 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1511 committed to OpenSSL should pass this lot as a minimum.
1514 *) Add session ticket override functionality for use by EAP-FAST.
1515 [Jouni Malinen <j@w1.fi>]
1517 *) Modify HMAC functions to return a value. Since these can be implemented
1518 in an ENGINE errors can occur.
1521 *) Type-checked OBJ_bsearch_ex.
1524 *) Type-checked OBJ_bsearch. Also some constification necessitated
1525 by type-checking. Still to come: TXT_DB, bsearch(?),
1526 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1530 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1531 seconds to a tm structure directly, instead of going through OS
1532 specific date routines. This avoids any issues with OS routines such
1533 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1534 and X509_time_adj_ex() to cover the extended range. The existing
1535 X509_time_adj() is still usable and will no longer have any date issues.
1538 *) Delta CRL support. New use deltas option which will attempt to locate
1539 and search any appropriate delta CRLs available.
1541 This work was sponsored by Google.
1544 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1545 code and add additional score elements. Validate alternate CRL paths
1546 as part of the CRL checking and indicate a new error "CRL path validation
1547 error" in this case. Applications wanting additional details can use
1548 the verify callback and check the new "parent" field. If this is not
1549 NULL CRL path validation is taking place. Existing applications wont
1550 see this because it requires extended CRL support which is off by
1553 This work was sponsored by Google.
1556 *) Support for freshest CRL extension.
1558 This work was sponsored by Google.
1561 *) Initial indirect CRL support. Currently only supported in the CRLs
1562 passed directly and not via lookup. Process certificate issuer
1563 CRL entry extension and lookup CRL entries by bother issuer name
1564 and serial number. Check and process CRL issuer entry in IDP extension.
1566 This work was sponsored by Google.
1569 *) Add support for distinct certificate and CRL paths. The CRL issuer
1570 certificate is validated separately in this case. Only enabled if
1571 an extended CRL support flag is set: this flag will enable additional
1572 CRL functionality in future.
1574 This work was sponsored by Google.
1577 *) Add support for policy mappings extension.
1579 This work was sponsored by Google.
1582 *) Fixes to pathlength constraint, self issued certificate handling,
1583 policy processing to align with RFC3280 and PKITS tests.
1585 This work was sponsored by Google.
1588 *) Support for name constraints certificate extension. DN, email, DNS
1589 and URI types are currently supported.
1591 This work was sponsored by Google.
1594 *) To cater for systems that provide a pointer-based thread ID rather
1595 than numeric, deprecate the current numeric thread ID mechanism and
1596 replace it with a structure and associated callback type. This
1597 mechanism allows a numeric "hash" to be extracted from a thread ID in
1598 either case, and on platforms where pointers are larger than 'long',
1599 mixing is done to help ensure the numeric 'hash' is usable even if it
1600 can't be guaranteed unique. The default mechanism is to use "&errno"
1601 as a pointer-based thread ID to distinguish between threads.
1603 Applications that want to provide their own thread IDs should now use
1604 CRYPTO_THREADID_set_callback() to register a callback that will call
1605 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1607 Note that ERR_remove_state() is now deprecated, because it is tied
1608 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1609 to free the current thread's error state should be replaced by
1610 ERR_remove_thread_state(NULL).
1612 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1613 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1614 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1615 application was previously providing a numeric thread callback that
1616 was inappropriate for distinguishing threads, then uniqueness might
1617 have been obtained with &errno that happened immediately in the
1618 intermediate development versions of OpenSSL; this is no longer the
1619 case, the numeric thread callback will now override the automatic use
1621 [Geoff Thorpe, with help from Bodo Moeller]
1623 *) Initial support for different CRL issuing certificates. This covers a
1624 simple case where the self issued certificates in the chain exist and
1625 the real CRL issuer is higher in the existing chain.
1627 This work was sponsored by Google.
1630 *) Removed effectively defunct crypto/store from the build.
1633 *) Revamp of STACK to provide stronger type-checking. Still to come:
1634 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1635 ASN1_STRING, CONF_VALUE.
1638 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1639 RAM on SSL connections. This option can save about 34k per idle SSL.
1642 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1643 STACK, TXT_DB, bsearch, qsort.
1646 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1647 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1648 support for data, signedData, compressedData, digestedData and
1649 encryptedData, envelopedData types included. Scripts to check against
1650 RFC4134 examples draft and interop and consistency checks of many
1651 content types and variants.
1654 *) Add options to enc utility to support use of zlib compression BIO.
1657 *) Extend mk1mf to support importing of options and assembly language
1658 files from Configure script, currently only included in VC-WIN32.
1659 The assembly language rules can now optionally generate the source
1660 files from the associated perl scripts.
1663 *) Implement remaining functionality needed to support GOST ciphersuites.
1664 Interop testing has been performed using CryptoPro implementations.
1665 [Victor B. Wagner <vitus@cryptocom.ru>]
1667 *) s390x assembler pack.
1670 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1674 *) Implement Opaque PRF Input TLS extension as specified in
1675 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1676 official specification yet and no extension type assignment by
1677 IANA exists, this extension (for now) will have to be explicitly
1678 enabled when building OpenSSL by providing the extension number
1679 to use. For example, specify an option
1681 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1683 to the "config" or "Configure" script to enable the extension,
1684 assuming extension number 0x9527 (which is a completely arbitrary
1685 and unofficial assignment based on the MD5 hash of the Internet
1686 Draft). Note that by doing so, you potentially lose
1687 interoperability with other TLS implementations since these might
1688 be using the same extension number for other purposes.
1690 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1691 opaque PRF input value to use in the handshake. This will create
1692 an interal copy of the length-'len' string at 'src', and will
1693 return non-zero for success.
1695 To get more control and flexibility, provide a callback function
1698 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1699 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1703 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1706 Callback function 'cb' will be called in handshakes, and is
1707 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1708 Argument 'arg' is for application purposes (the value as given to
1709 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1710 be provided to the callback function). The callback function
1711 has to return non-zero to report success: usually 1 to use opaque
1712 PRF input just if possible, or 2 to enforce use of the opaque PRF
1713 input. In the latter case, the library will abort the handshake
1714 if opaque PRF input is not successfully negotiated.
1716 Arguments 'peerinput' and 'len' given to the callback function
1717 will always be NULL and 0 in the case of a client. A server will
1718 see the client's opaque PRF input through these variables if
1719 available (NULL and 0 otherwise). Note that if the server
1720 provides an opaque PRF input, the length must be the same as the
1721 length of the client's opaque PRF input.
1723 Note that the callback function will only be called when creating
1724 a new session (session resumption can resume whatever was
1725 previously negotiated), and will not be called in SSL 2.0
1726 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1727 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1728 for applications that need to enforce opaque PRF input.
1732 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1735 [Victor B. Wagner <vitus@cryptocom.ru>]
1737 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1738 RFC4507bis. The encrypted ticket format is an encrypted encoded
1739 SSL_SESSION structure, that way new session features are automatically
1742 If a client application caches session in an SSL_SESSION structure
1743 support is transparent because tickets are now stored in the encoded
1746 The SSL_CTX structure automatically generates keys for ticket
1747 protection in servers so again support should be possible
1748 with no application modification.
1750 If a client or server wishes to disable RFC4507 support then the option
1751 SSL_OP_NO_TICKET can be set.
1753 Add a TLS extension debugging callback to allow the contents of any client
1754 or server extensions to be examined.
1756 This work was sponsored by Google.
1759 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1760 OpenSSL should now compile cleanly on gcc 4.2
1761 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1763 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1764 support including streaming MAC support: this is required for GOST
1765 ciphersuite support.
1766 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1768 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1769 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1770 to output in BER and PEM format.
1773 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1774 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1775 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1776 ENGINE support for HMAC keys which are unextractable. New -mac and
1777 -macopt options to dgst utility.
1780 *) New option -sigopt to dgst utility. Update dgst to use
1781 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1782 alternative signing paramaters such as X9.31 or PSS in the dgst
1786 *) Change ssl_cipher_apply_rule(), the internal function that does
1787 the work each time a ciphersuite string requests enabling
1788 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1789 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1790 the order of disabled ciphersuites such that those ciphersuites
1791 that most recently went from enabled to disabled not only stay
1792 in order with respect to each other, but also have higher priority
1793 than other disabled ciphersuites the next time ciphersuites are
1796 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1797 the same ciphersuites as with "HIGH" alone, but in a specific
1798 order where the PSK ciphersuites come first (since they are the
1799 most recently disabled ciphersuites when "HIGH" is parsed).
1801 Also, change ssl_create_cipher_list() (using this new
1802 funcionality) such that between otherwise identical
1803 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1807 *) Change ssl_create_cipher_list() so that it automatically
1808 arranges the ciphersuites in reasonable order before starting
1809 to process the rule string. Thus, the definition for "DEFAULT"
1810 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1811 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1812 This makes it much easier to arrive at a reasonable default order
1813 in applications for which anonymous ciphers are OK (meaning
1814 that you can't actually use DEFAULT).
1815 [Bodo Moeller; suggested by Victor Duchovni]
1817 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1818 processing) into multiple integers instead of setting
1819 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1820 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1821 (These masks as well as the individual bit definitions are hidden
1822 away into the non-exported interface ssl/ssl_locl.h, so this
1823 change to the definition of the SSL_CIPHER structure shouldn't
1824 affect applications.) This give us more bits for each of these
1825 categories, so there is no longer a need to coagulate AES128 and
1826 AES256 into a single algorithm bit, and to coagulate Camellia128
1827 and Camellia256 into a single algorithm bit, which has led to all
1830 Thus, among other things, the kludge introduced in 0.9.7m and
1831 0.9.8e for masking out AES256 independently of AES128 or masking
1832 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1834 With the change, we also introduce new ciphersuite aliases that
1835 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1839 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1840 Use the leftmost N bytes of the signature input if the input is
1841 larger than the prime q (with N being the size in bytes of q).
1844 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1845 it yet and it is largely untested.
1848 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1851 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1852 some compilers (gcc 4.2 and later) reject their use. Safestack is
1853 reimplemented. Update ASN1 to avoid use of legacy functions.
1856 *) Win32/64 targets are linked with Winsock2.
1859 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1860 to external functions. This can be used to increase CRL handling
1861 efficiency especially when CRLs are very large by (for example) storing
1862 the CRL revoked certificates in a database.
1865 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1866 new CRLs added to a directory can be used. New command line option
1867 -verify_return_error to s_client and s_server. This causes real errors
1868 to be returned by the verify callback instead of carrying on no matter
1869 what. This reflects the way a "real world" verify callback would behave.
1872 *) GOST engine, supporting several GOST algorithms and public key formats.
1873 Kindly donated by Cryptocom.
1876 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1877 partitioned by DP are handled but no indirect CRL or reason partitioning
1878 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1879 selected via a scoring technique which handles IDP and AKID in CRLs.
1882 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1883 will ultimately be used for all verify operations: this will remove the
1884 X509_STORE dependency on certificate verification and allow alternative
1885 lookup methods. X509_STORE based implementations of these two callbacks.
1888 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1889 Modify get_crl() to find a valid (unexpired) CRL if possible.
1892 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1893 this would be called X509_CRL_cmp() but that name is already used by
1894 a function that just compares CRL issuer names. Cache several CRL
1895 extensions in X509_CRL structure and cache CRLDP in X509.
1898 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1899 this maps equivalent X509_NAME structures into a consistent structure.
1900 Name comparison can then be performed rapidly using memcmp().
1903 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1907 *) Allow digests to supply their own micalg string for S/MIME type using
1908 the ctrl EVP_MD_CTRL_MICALG.
1911 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1912 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1913 ctrl. It can then customise the structure before and/or after signing
1917 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1918 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1919 to free up any added signature OIDs.
1922 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1923 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1924 digest and cipher tables. New options added to openssl utility:
1925 list-message-digest-algorithms and list-cipher-algorithms.
1928 *) Change the array representation of binary polynomials: the list
1929 of degrees of non-zero coefficients is now terminated with -1.
1930 Previously it was terminated with 0, which was also part of the
1931 value; thus, the array representation was not applicable to
1932 polynomials where t^0 has coefficient zero. This change makes
1933 the array representation useful in a more general context.
1936 *) Various modifications and fixes to SSL/TLS cipher string
1937 handling. For ECC, the code now distinguishes between fixed ECDH
1938 with RSA certificates on the one hand and with ECDSA certificates
1939 on the other hand, since these are separate ciphersuites. The
1940 unused code for Fortezza ciphersuites has been removed.
1942 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1943 (not "ECDHE"). For consistency with the code for DH
1944 certificates, use of ECDH certificates is now considered ECDH
1945 authentication, not RSA or ECDSA authentication (the latter is
1946 merely the CA's signing algorithm and not actively used in the
1949 The temporary ciphersuite alias "ECCdraft" is no longer
1950 available, and ECC ciphersuites are no longer excluded from "ALL"
1951 and "DEFAULT". The following aliases now exist for RFC 4492
1952 ciphersuites, most of these by analogy with the DH case:
1954 kECDHr - ECDH cert, signed with RSA
1955 kECDHe - ECDH cert, signed with ECDSA
1956 kECDH - ECDH cert (signed with either RSA or ECDSA)
1957 kEECDH - ephemeral ECDH
1958 ECDH - ECDH cert or ephemeral ECDH
1964 AECDH - anonymous ECDH
1965 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1969 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1970 Use correct micalg parameters depending on digest(s) in signed message.
1973 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1974 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1977 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1978 an engine to register a method. Add ENGINE lookups for methods and
1979 functional reference processing.
1982 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1983 EVP_{Sign,Verify}* which allow an application to customise the signature
1987 *) New -resign option to smime utility. This adds one or more signers
1988 to an existing PKCS#7 signedData structure. Also -md option to use an
1989 alternative message digest algorithm for signing.
1992 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1993 create PKCS7 structures containing multiple signers. Update smime
1994 application to support multiple signers.
1997 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2001 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2002 Reorganize PBE internals to lookup from a static table using NIDs,
2003 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2004 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2005 PRF which will be automatically used with PBES2.
2008 *) Replace the algorithm specific calls to generate keys in "req" with the
2012 *) Update PKCS#7 enveloped data routines to use new API. This is now
2013 supported by any public key method supporting the encrypt operation. A
2014 ctrl is added to allow the public key algorithm to examine or modify
2015 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2019 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2020 a default digest type to use. In most cases this will be SHA1 but some
2021 algorithms (such as GOST) need to specify an alternative digest. The
2022 return value indicates how strong the prefernce is 1 means optional and
2023 2 is mandatory (that is it is the only supported type). Modify
2024 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2025 use the default md. Update openssl utilities to use the default digest
2026 type for signing if it is not explicitly indicated.
2029 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2030 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2031 signing method from the key type. This effectively removes the link
2032 between digests and public key types.
2035 *) Add an OID cross reference table and utility functions. Its purpose is to
2036 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2037 rsaEncryption. This will allow some of the algorithm specific hackery
2038 needed to use the correct OID to be removed.
2041 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2042 structures for PKCS7_sign(). They are now set up by the relevant public
2046 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2049 *) Add support for key derivation (agreement) in the API, DH method and
2053 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2054 public and private key formats. As a side effect these add additional
2055 command line functionality not previously available: DSA signatures can be
2056 generated and verified using pkeyutl and DH key support and generation in
2061 [Oliver Tappe <zooey@hirschkaefer.de>]
2063 *) New make target "install_html_docs" installs HTML renditions of the
2065 [Oliver Tappe <zooey@hirschkaefer.de>]
2067 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
2068 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2069 support key and parameter generation and add initial key generation
2070 functionality for RSA.
2073 *) Add functions for main EVP_PKEY_method operations. The undocumented
2074 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2075 EVP_PKEY_{encrypt,decrypt}_old.
2078 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2079 key API, doesn't do much yet.
2082 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2083 public key algorithms. New option to openssl utility:
2084 "list-public-key-algorithms" to print out info.
2087 *) Implement the Supported Elliptic Curves Extension for
2088 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2091 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2092 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2095 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2096 utilities such as rsa, dsa, dsaparam etc except they process any key
2100 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2101 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2102 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2106 *) Initial support for pluggable public key ASN1.
2107 De-spaghettify the public key ASN1 handling. Move public and private
2108 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2109 algorithm specific handling to a single module within the relevant
2110 algorithm directory. Add functions to allow (near) opaque processing
2111 of public and private key structures.
2114 *) Implement the Supported Point Formats Extension for
2115 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2118 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2119 for the psk identity [hint] and the psk callback functions to the
2120 SSL_SESSION, SSL and SSL_CTX structure.
2123 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2127 SSL_CTX_use_psk_identity_hint
2128 SSL_get_psk_identity_hint
2129 SSL_get_psk_identity
2130 SSL_use_psk_identity_hint
2132 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2134 *) Add RFC 3161 compliant time stamp request creation, response generation
2135 and response verification functionality.
2136 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2138 *) Add initial support for TLS extensions, specifically for the server_name
2139 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2140 have new members for a host name. The SSL data structure has an
2141 additional member SSL_CTX *initial_ctx so that new sessions can be
2142 stored in that context to allow for session resumption, even after the
2143 SSL has been switched to a new SSL_CTX in reaction to a client's
2144 server_name extension.
2146 New functions (subject to change):
2148 SSL_get_servername()
2149 SSL_get_servername_type()
2152 New CTRL codes and macros (subject to change):
2154 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2155 - SSL_CTX_set_tlsext_servername_callback()
2156 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2157 - SSL_CTX_set_tlsext_servername_arg()
2158 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2160 openssl s_client has a new '-servername ...' option.
2162 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2163 '-key2 ...', '-servername_fatal' (subject to change). This allows
2164 testing the HostName extension for a specific single host name ('-cert'
2165 and '-key' remain fallbacks for handshakes without HostName
2166 negotiation). If the unrecogninzed_name alert has to be sent, this by
2167 default is a warning; it becomes fatal with the '-servername_fatal'
2170 [Peter Sylvester, Remy Allais, Christophe Renou]
2172 *) Whirlpool hash implementation is added.
2175 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2176 bn(64,32). Because of instruction set limitations it doesn't have
2177 any negative impact on performance. This was done mostly in order
2178 to make it possible to share assembler modules, such as bn_mul_mont
2179 implementations, between 32- and 64-bit builds without hassle.
2182 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2183 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2187 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2188 dedicated Montgomery multiplication procedure, is introduced.
2189 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2190 "64-bit" performance on certain 32-bit targets.
2193 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2194 in SSL structures. New SSL ctrl to set maximum send fragment size.
2195 Save memory by seeting the I/O buffer sizes dynamically instead of
2196 using the maximum available value.
2199 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2200 in addition to the text details.
2203 *) Very, very preliminary EXPERIMENTAL support for printing of general
2204 ASN1 structures. This currently produces rather ugly output and doesn't
2205 handle several customised structures at all.
2208 *) Integrated support for PVK file format and some related formats such
2209 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2210 these in the 'rsa' and 'dsa' utilities.
2213 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2216 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2217 place for the (very old) "NETSCAPE" format certificates which are now
2218 handled using new ASN1 code equivalents.
2221 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2222 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2223 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2226 *) Modify CRL distribution points extension code to print out previously
2227 unsupported fields. Enhance extension setting code to allow setting of
2231 *) Add print and set support for Issuing Distribution Point CRL extension.
2234 *) Change 'Configure' script to enable Camellia by default.
2237 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2239 *) When rejecting SSL/TLS records due to an incorrect version number, never
2240 update s->server with a new major version number. As of
2241 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2242 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2243 the previous behavior could result in a read attempt at NULL when
2244 receiving specific incorrect SSL/TLS records once record payload
2245 protection is active. (CVE-2010-0740)
2246 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2248 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2249 could be crashed if the relevant tables were not present (e.g. chrooted).
2250 [Tomas Hoger <thoger@redhat.com>]
2252 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2254 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2255 [Martin Olsson, Neel Mehta]
2257 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2258 accommodate for stack sorting, always a write lock!).
2261 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2262 excessive delays in the RAND_poll(): over a minute. As a workaround
2263 include a time check in the inner Heap32Next loop too.
2266 *) The code that handled flushing of data in SSL/TLS originally used the
2267 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2268 the problem outlined in PR#1949. The fix suggested there however can
2269 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2270 of Apache). So instead simplify the code to flush unconditionally.
2271 This should be fine since flushing with no data to flush is a no op.
2274 *) Handle TLS versions 2.0 and later properly and correctly use the
2275 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2276 off ancient servers have a habit of sticking around for a while...
2279 *) Modify compression code so it frees up structures without using the
2280 ex_data callbacks. This works around a problem where some applications
2281 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2282 restarting) then use compression (e.g. SSL with compression) later.
2283 This results in significant per-connection memory leaks and
2284 has caused some security issues including CVE-2008-1678 and
2288 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2289 change when encrypting or decrypting.
2292 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2293 connect and renegotiate with servers which do not support RI.
2294 Until RI is more widely deployed this option is enabled by default.
2297 *) Add "missing" ssl ctrls to clear options and mode.
2300 *) If client attempts to renegotiate and doesn't support RI respond with
2301 a no_renegotiation alert as required by RFC5746. Some renegotiating
2302 TLS clients will continue a connection gracefully when they receive
2303 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2304 waiting for a server hello which it will never receive. Now we treat a
2305 received no_renegotiation alert as a fatal error. This is because
2306 applications requesting a renegotiation might well expect it to succeed
2307 and would have no code in place to handle the server denying it so the
2308 only safe thing to do is to terminate the connection.
2311 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2312 peer supports secure renegotiation and 0 otherwise. Print out peer
2313 renegotiation support in s_client/s_server.
2316 *) Replace the highly broken and deprecated SPKAC certification method with
2317 the updated NID creation version. This should correctly handle UTF8.
2320 *) Implement RFC5746. Re-enable renegotiation but require the extension
2321 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2322 turns out to be a bad idea. It has been replaced by
2323 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2324 SSL_CTX_set_options(). This is really not recommended unless you
2325 know what you are doing.
2326 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2328 *) Fixes to stateless session resumption handling. Use initial_ctx when
2329 issuing and attempting to decrypt tickets in case it has changed during
2330 servername handling. Use a non-zero length session ID when attempting
2331 stateless session resumption: this makes it possible to determine if
2332 a resumption has occurred immediately after receiving server hello
2333 (several places in OpenSSL subtly assume this) instead of later in
2337 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2338 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2339 fixes for a few places where the return code is not checked
2341 [Julia Lawall <julia@diku.dk>]
2343 *) Add --strict-warnings option to Configure script to include devteam
2344 warnings in other configurations.
2347 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2348 makes it possible to install openssl libraries in locations which
2349 have names other than "lib", for example "/usr/lib64" which some
2351 [Steve Henson, based on patch from Jeremy Utley]
2353 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2354 X690 8.9.12 and can produce some misleading textual output of OIDs.
2355 [Steve Henson, reported by Dan Kaminsky]
2357 *) Delete MD2 from algorithm tables. This follows the recommendation in
2358 several standards that it is not used in new applications due to
2359 several cryptographic weaknesses. For binary compatibility reasons
2360 the MD2 API is still compiled in by default.
2363 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2367 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2368 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2370 [Guenter <lists@gknw.net>]
2372 *) Fix the server certificate chain building code to use X509_verify_cert(),
2373 it used to have an ad-hoc builder which was unable to cope with anything
2374 other than a simple chain.
2375 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2377 *) Don't check self signed certificate signatures in X509_verify_cert()
2378 by default (a flag can override this): it just wastes time without
2379 adding any security. As a useful side effect self signed root CAs
2380 with non-FIPS digests are now usable in FIPS mode.
2383 *) In dtls1_process_out_of_seq_message() the check if the current message
2384 is already buffered was missing. For every new message was memory
2385 allocated, allowing an attacker to perform an denial of service attack
2386 with sending out of seq handshake messages until there is no memory
2387 left. Additionally every future messege was buffered, even if the
2388 sequence number made no sense and would be part of another handshake.
2389 So only messages with sequence numbers less than 10 in advance will be
2390 buffered. (CVE-2009-1378)
2391 [Robin Seggelmann, discovered by Daniel Mentz]
2393 *) Records are buffered if they arrive with a future epoch to be
2394 processed after finishing the corresponding handshake. There is
2395 currently no limitation to this buffer allowing an attacker to perform
2396 a DOS attack with sending records with future epochs until there is no
2397 memory left. This patch adds the pqueue_size() function to detemine
2398 the size of a buffer and limits the record buffer to 100 entries.
2400 [Robin Seggelmann, discovered by Daniel Mentz]
2402 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2403 parent structure is freed. (CVE-2009-1379)
2406 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2407 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2410 [Ilya O. <vrghost@gmail.com>]
2412 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2414 *) Disable renegotiation completely - this fixes a severe security
2415 problem (CVE-2009-3555) at the cost of breaking all
2416 renegotiation. Renegotiation can be re-enabled by setting
2417 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2418 run-time. This is really not recommended unless you know what
2422 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2424 *) Don't set val to NULL when freeing up structures, it is freed up by
2425 underlying code. If sizeof(void *) > sizeof(long) this can result in
2426 zeroing past the valid field. (CVE-2009-0789)
2427 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2429 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2430 checked correctly. This would allow some invalid signed attributes to
2431 appear to verify correctly. (CVE-2009-0591)
2432 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2434 *) Reject UniversalString and BMPString types with invalid lengths. This
2435 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2436 a legal length. (CVE-2009-0590)
2439 *) Set S/MIME signing as the default purpose rather than setting it
2440 unconditionally. This allows applications to override it at the store
2444 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2445 to handle some structures.
2448 *) Improve efficiency of mem_gets: don't search whole buffer each time
2450 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2452 *) New -hex option for openssl rand.
2455 *) Print out UTF8String and NumericString when parsing ASN1.
2458 *) Support NumericString type for name components.
2461 *) Allow CC in the environment to override the automatically chosen
2462 compiler. Note that nothing is done to ensure flags work with the
2466 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2468 *) Properly check EVP_VerifyFinal() and similar return values
2470 [Ben Laurie, Bodo Moeller, Google Security Team]
2472 *) Enable TLS extensions by default.
2475 *) Allow the CHIL engine to be loaded, whether the application is
2476 multithreaded or not. (This does not release the developer from the
2477 obligation to set up the dynamic locking callbacks.)
2478 [Sander Temme <sander@temme.net>]
2480 *) Use correct exit code if there is an error in dgst command.
2481 [Steve Henson; problem pointed out by Roland Dirlewanger]
2483 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2484 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2487 *) Add experimental JPAKE support, including demo authentication in
2488 s_client and s_server.
2491 *) Set the comparison function in v3_addr_canonize().
2492 [Rob Austein <sra@hactrn.net>]
2494 *) Add support for XMPP STARTTLS in s_client.
2495 [Philip Paeps <philip@freebsd.org>]
2497 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2498 to ensure that even with this option, only ciphersuites in the
2499 server's preference list will be accepted. (Note that the option
2500 applies only when resuming a session, so the earlier behavior was
2501 just about the algorithm choice for symmetric cryptography.)
2504 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2506 *) Fix NULL pointer dereference if a DTLS server received
2507 ChangeCipherSpec as first record (CVE-2009-1386).
2510 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2511 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2514 *) The fix in 0.9.8c that supposedly got rid of unsafe
2515 double-checked locking was incomplete for RSA blinding,
2516 addressing just one layer of what turns out to have been
2517 doubly unsafe triple-checked locking.
2519 So now fix this for real by retiring the MONT_HELPER macro
2520 in crypto/rsa/rsa_eay.c.
2522 [Bodo Moeller; problem pointed out by Marius Schilder]
2524 *) Various precautionary measures:
2526 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2528 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2529 (NB: This would require knowledge of the secret session ticket key
2530 to exploit, in which case you'd be SOL either way.)
2532 - Change bn_nist.c so that it will properly handle input BIGNUMs
2533 outside the expected range.
2535 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2538 [Neel Mehta, Bodo Moeller]
2540 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2541 the load fails. Useful for distros.
2542 [Ben Laurie and the FreeBSD team]
2544 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2547 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2550 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2552 This work was sponsored by Logica.
2555 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2556 keystores. Support for SSL/TLS client authentication too.
2557 Not compiled unless enable-capieng specified to Configure.
2559 This work was sponsored by Logica.
2562 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2563 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2564 attribute creation routines such as certifcate requests and PKCS#12
2568 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2570 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2571 handshake which could lead to a cilent crash as found using the
2572 Codenomicon TLS test suite (CVE-2008-1672)
2573 [Steve Henson, Mark Cox]
2575 *) Fix double free in TLS server name extensions which could lead to
2576 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2579 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2581 Clear the error queue to ensure that error entries left from
2582 older function calls do not interfere with the correct operation.
2583 [Lutz Jaenicke, Erik de Castro Lopo]
2585 *) Remove root CA certificates of commercial CAs:
2587 The OpenSSL project does not recommend any specific CA and does not
2588 have any policy with respect to including or excluding any CA.
2589 Therefore it does not make any sense to ship an arbitrary selection
2590 of root CA certificates with the OpenSSL software.
2593 *) RSA OAEP patches to fix two separate invalid memory reads.
2594 The first one involves inputs when 'lzero' is greater than
2595 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2596 before the beginning of from). The second one involves inputs where
2597 the 'db' section contains nothing but zeroes (there is a one-byte
2598 invalid read after the end of 'db').
2599 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2601 *) Partial backport from 0.9.9-dev:
2603 Introduce bn_mul_mont (dedicated Montgomery multiplication
2604 procedure) as a candidate for BIGNUM assembler implementation.
2605 While 0.9.9-dev uses assembler for various architectures, only
2606 x86_64 is available by default here in the 0.9.8 branch, and
2607 32-bit x86 is available through a compile-time setting.
2609 To try the 32-bit x86 assembler implementation, use Configure
2610 option "enable-montasm" (which exists only for this backport).
2612 As "enable-montasm" for 32-bit x86 disclaims code stability
2613 anyway, in this constellation we activate additional code
2614 backported from 0.9.9-dev for further performance improvements,
2615 namely BN_from_montgomery_word. (To enable this otherwise,
2616 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2618 [Andy Polyakov (backport partially by Bodo Moeller)]
2620 *) Add TLS session ticket callback. This allows an application to set
2621 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2622 values. This is useful for key rollover for example where several key
2623 sets may exist with different names.
2626 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2627 This was broken until now in 0.9.8 releases, such that the only way
2628 a registered ENGINE could be used (assuming it initialises
2629 successfully on the host) was to explicitly set it as the default
2630 for the relevant algorithms. This is in contradiction with 0.9.7
2631 behaviour and the documentation. With this fix, when an ENGINE is
2632 registered into a given algorithm's table of implementations, the
2633 'uptodate' flag is reset so that auto-discovery will be used next
2634 time a new context for that algorithm attempts to select an
2636 [Ian Lister (tweaked by Geoff Thorpe)]
2638 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2639 implemention in the following ways:
2641 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2644 Lack of BER streaming support means one pass streaming processing is
2645 only supported if data is detached: setting the streaming flag is
2646 ignored for embedded content.
2648 CMS support is disabled by default and must be explicitly enabled
2649 with the enable-cms configuration option.
2652 *) Update the GMP engine glue to do direct copies between BIGNUM and
2653 mpz_t when openssl and GMP use the same limb size. Otherwise the
2654 existing "conversion via a text string export" trick is still used.
2655 [Paul Sheer <paulsheer@gmail.com>]
2657 *) Zlib compression BIO. This is a filter BIO which compressed and
2658 uncompresses any data passed through it.
2661 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2662 RFC3394 compatible AES key wrapping.
2665 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2666 sets string data without copying. X509_ALGOR_set0() and
2667 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2668 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2669 from an X509_ATTRIBUTE structure optionally checking it occurs only
2670 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2674 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2675 to get the expected BN_FLG_CONSTTIME behavior.
2676 [Bodo Moeller (Google)]
2680 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2681 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2682 - added some more tests to do_tests.pl
2683 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2684 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2685 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2686 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2687 - various changes to netware.pl to enable gcc-cross builds on Win32
2689 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2690 - various changes to fix missing prototype warnings
2691 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2692 - added AES, WHIRLPOOL and CPUID assembler code to build files
2693 - added missing AES assembler make rules to mk1mf.pl
2694 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2695 [Guenter Knauf <eflash@gmx.net>]
2697 *) Implement certificate status request TLS extension defined in RFC3546.
2698 A client can set the appropriate parameters and receive the encoded
2699 OCSP response via a callback. A server can query the supplied parameters
2700 and set the encoded OCSP response in the callback. Add simplified examples
2701 to s_client and s_server.
2704 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2706 *) Fix various bugs:
2707 + Binary incompatibility of ssl_ctx_st structure
2708 + DTLS interoperation with non-compliant servers
2709 + Don't call get_session_cb() without proposed session
2710 + Fix ia64 assembler code
2711 [Andy Polyakov, Steve Henson]
2713 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2715 *) DTLS Handshake overhaul. There were longstanding issues with
2716 OpenSSL DTLS implementation, which were making it impossible for
2717 RFC 4347 compliant client to communicate with OpenSSL server.
2718 Unfortunately just fixing these incompatibilities would "cut off"
2719 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2720 server keeps tolerating non RFC compliant syntax. The opposite is
2721 not true, 0.9.8f client can not communicate with earlier server.
2722 This update even addresses CVE-2007-4995.
2725 *) Changes to avoid need for function casts in OpenSSL: some compilers
2726 (gcc 4.2 and later) reject their use.
2727 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2730 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2731 RFC4507bis. The encrypted ticket format is an encrypted encoded
2732 SSL_SESSION structure, that way new session features are automatically
2735 If a client application caches session in an SSL_SESSION structure
2736 support is transparent because tickets are now stored in the encoded
2739 The SSL_CTX structure automatically generates keys for ticket
2740 protection in servers so again support should be possible
2741 with no application modification.
2743 If a client or server wishes to disable RFC4507 support then the option
2744 SSL_OP_NO_TICKET can be set.
2746 Add a TLS extension debugging callback to allow the contents of any client
2747 or server extensions to be examined.
2749 This work was sponsored by Google.
2752 *) Add initial support for TLS extensions, specifically for the server_name
2753 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2754 have new members for a host name. The SSL data structure has an
2755 additional member SSL_CTX *initial_ctx so that new sessions can be
2756 stored in that context to allow for session resumption, even after the
2757 SSL has been switched to a new SSL_CTX in reaction to a client's
2758 server_name extension.
2760 New functions (subject to change):
2762 SSL_get_servername()
2763 SSL_get_servername_type()
2766 New CTRL codes and macros (subject to change):
2768 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2769 - SSL_CTX_set_tlsext_servername_callback()
2770 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2771 - SSL_CTX_set_tlsext_servername_arg()
2772 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2774 openssl s_client has a new '-servername ...' option.
2776 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2777 '-key2 ...', '-servername_fatal' (subject to change). This allows
2778 testing the HostName extension for a specific single host name ('-cert'
2779 and '-key' remain fallbacks for handshakes without HostName
2780 negotiation). If the unrecogninzed_name alert has to be sent, this by
2781 default is a warning; it becomes fatal with the '-servername_fatal'
2784 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2786 *) Add AES and SSE2 assembly language support to VC++ build.
2789 *) Mitigate attack on final subtraction in Montgomery reduction.
2792 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2793 (which previously caused an internal error).
2796 *) Squeeze another 10% out of IGE mode when in != out.
2799 *) AES IGE mode speedup.
2800 [Dean Gaudet (Google)]
2802 *) Add the Korean symmetric 128-bit cipher SEED (see
2803 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2804 add SEED ciphersuites from RFC 4162:
2806 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2807 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2808 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2809 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2811 To minimize changes between patchlevels in the OpenSSL 0.9.8
2812 series, SEED remains excluded from compilation unless OpenSSL
2813 is configured with 'enable-seed'.
2814 [KISA, Bodo Moeller]
2816 *) Mitigate branch prediction attacks, which can be practical if a
2817 single processor is shared, allowing a spy process to extract
2818 information. For detailed background information, see
2819 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2820 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2821 and Necessary Software Countermeasures"). The core of the change
2822 are new versions BN_div_no_branch() and
2823 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2824 respectively, which are slower, but avoid the security-relevant
2825 conditional branches. These are automatically called by BN_div()
2826 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2827 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2828 remove a conditional branch.
2830 BN_FLG_CONSTTIME is the new name for the previous
2831 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2832 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2833 in the exponent causes BN_mod_exp_mont() to use the alternative
2834 implementation in BN_mod_exp_mont_consttime().) The old name
2835 remains as a deprecated alias.
2837 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2838 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2839 constant-time implementations for more than just exponentiation.
2840 Here too the old name is kept as a deprecated alias.
2842 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2843 the BN_BLINDING structure gets an independent copy of the
2844 modulus. This means that the previous "BIGNUM *m" argument to
2845 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2846 essentially becomes "const BIGNUM *m", although we can't actually
2847 change this in the header file before 0.9.9. It allows
2848 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2849 enable BN_FLG_CONSTTIME.
2851 [Matthew D Wood (Intel Corp)]
2853 *) In the SSL/TLS server implementation, be strict about session ID
2854 context matching (which matters if an application uses a single
2855 external cache for different purposes). Previously,
2856 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2857 set. This did ensure strict client verification, but meant that,
2858 with applications using a single external cache for quite
2859 different requirements, clients could circumvent ciphersuite
2860 restrictions for a given session ID context by starting a session
2861 in a different context.
2864 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2865 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2866 authentication-only ciphersuites.
2869 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2870 not complete and could lead to a possible single byte overflow
2871 (CVE-2007-5135) [Ben Laurie]
2873 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2875 *) Since AES128 and AES256 (and similarly Camellia128 and
2876 Camellia256) share a single mask bit in the logic of
2877 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2878 kludge to work properly if AES128 is available and AES256 isn't
2879 (or if Camellia128 is available and Camellia256 isn't).
2882 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2883 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2884 When a point or a seed is encoded in a BIT STRING, we need to
2885 prevent the removal of trailing zero bits to get the proper DER
2886 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2887 of a NamedBitList, for which trailing 0 bits need to be removed.)
2890 *) Have SSL/TLS server implementation tolerate "mismatched" record
2891 protocol version while receiving ClientHello even if the
2892 ClientHello is fragmented. (The server can't insist on the
2893 particular protocol version it has chosen before the ServerHello
2894 message has informed the client about his choice.)
2897 *) Add RFC 3779 support.
2898 [Rob Austein for ARIN, Ben Laurie]
2900 *) Load error codes if they are not already present instead of using a
2901 static variable. This allows them to be cleanly unloaded and reloaded.
2902 Improve header file function name parsing.
2905 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2906 or CAPABILITY handshake as required by RFCs.
2909 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2911 *) Introduce limits to prevent malicious keys being able to
2912 cause a denial of service. (CVE-2006-2940)
2913 [Steve Henson, Bodo Moeller]
2915 *) Fix ASN.1 parsing of certain invalid structures that can result
2916 in a denial of service. (CVE-2006-2937) [Steve Henson]
2918 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2919 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2921 *) Fix SSL client code which could crash if connecting to a
2922 malicious SSLv2 server. (CVE-2006-4343)
2923 [Tavis Ormandy and Will Drewry, Google Security Team]
2925 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2926 match only those. Before that, "AES256-SHA" would be interpreted
2927 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2928 the same strength classification in 0.9.7h) as we currently only
2929 have a single AES bit in the ciphersuite description bitmap.
2930 That change, however, also applied to ciphersuite strings such as
2931 "RC4-MD5" that intentionally matched multiple ciphersuites --
2932 namely, SSL 2.0 ciphersuites in addition to the more common ones
2933 from SSL 3.0/TLS 1.0.
2935 So we change the selection algorithm again: Naming an explicit
2936 ciphersuite selects this one ciphersuite, and any other similar
2937 ciphersuite (same bitmap) from *other* protocol versions.
2938 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2939 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2941 Since SSL 2.0 does not have any ciphersuites for which the
2942 128/256 bit distinction would be relevant, this works for now.
2943 The proper fix will be to use different bits for AES128 and
2944 AES256, which would have avoided the problems from the beginning;
2945 however, bits are scarce, so we can only do this in a new release
2946 (not just a patchlevel) when we can change the SSL_CIPHER
2947 definition to split the single 'unsigned long mask' bitmap into
2948 multiple values to extend the available space.
2952 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2954 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2955 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2957 *) Add AES IGE and biIGE modes.
2960 *) Change the Unix randomness entropy gathering to use poll() when
2961 possible instead of select(), since the latter has some
2962 undesirable limitations.
2963 [Darryl Miles via Richard Levitte and Bodo Moeller]
2965 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2966 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2967 cannot be implicitly activated as part of, e.g., the "AES" alias.
2968 However, please upgrade to OpenSSL 0.9.9[-dev] for
2969 non-experimental use of the ECC ciphersuites to get TLS extension
2970 support, which is required for curve and point format negotiation
2971 to avoid potential handshake problems.
2974 *) Disable rogue ciphersuites:
2976 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2977 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2978 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2980 The latter two were purportedly from
2981 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2984 Also deactivate the remaining ciphersuites from
2985 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2986 unofficial, and the ID has long expired.
2989 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2990 dual-core machines) and other potential thread-safety issues.
2993 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2994 versions), which is now available for royalty-free use
2995 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2996 Also, add Camellia TLS ciphersuites from RFC 4132.
2998 To minimize changes between patchlevels in the OpenSSL 0.9.8
2999 series, Camellia remains excluded from compilation unless OpenSSL
3000 is configured with 'enable-camellia'.
3003 *) Disable the padding bug check when compression is in use. The padding
3004 bug check assumes the first packet is of even length, this is not
3005 necessarily true if compresssion is enabled and can result in false
3006 positives causing handshake failure. The actual bug test is ancient
3007 code so it is hoped that implementations will either have fixed it by
3008 now or any which still have the bug do not support compression.
3011 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3013 *) When applying a cipher rule check to see if string match is an explicit
3014 cipher suite and only match that one cipher suite if it is.
3017 *) Link in manifests for VC++ if needed.
3018 [Austin Ziegler <halostatue@gmail.com>]
3020 *) Update support for ECC-based TLS ciphersuites according to
3021 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3022 TLS extensions, which are supported starting with the 0.9.9
3023 branch, not in the OpenSSL 0.9.8 branch).
3026 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3027 opaque EVP_CIPHER_CTX handling.
3030 *) Fixes and enhancements to zlib compression code. We now only use
3031 "zlib1.dll" and use the default __cdecl calling convention on Win32
3032 to conform with the standards mentioned here:
3033 http://www.zlib.net/DLL_FAQ.txt
3034 Static zlib linking now works on Windows and the new --with-zlib-include
3035 --with-zlib-lib options to Configure can be used to supply the location
3036 of the headers and library. Gracefully handle case where zlib library
3040 *) Several fixes and enhancements to the OID generation code. The old code
3041 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3042 handle numbers larger than ULONG_MAX, truncated printing and had a
3043 non standard OBJ_obj2txt() behaviour.
3046 *) Add support for building of engines under engine/ as shared libraries
3047 under VC++ build system.
3050 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3051 Hopefully, we will not see any false combination of paths any more.
3054 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3056 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3057 (part of SSL_OP_ALL). This option used to disable the
3058 countermeasure against man-in-the-middle protocol-version
3059 rollback in the SSL 2.0 server implementation, which is a bad
3060 idea. (CVE-2005-2969)
3062 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3063 for Information Security, National Institute of Advanced Industrial
3064 Science and Technology [AIST], Japan)]
3066 *) Add two function to clear and return the verify parameter flags.
3069 *) Keep cipherlists sorted in the source instead of sorting them at
3070 runtime, thus removing the need for a lock.
3073 *) Avoid some small subgroup attacks in Diffie-Hellman.
3074 [Nick Mathewson and Ben Laurie]
3076 *) Add functions for well-known primes.
3079 *) Extended Windows CE support.
3080 [Satoshi Nakamura and Andy Polyakov]
3082 *) Initialize SSL_METHOD structures at compile time instead of during
3083 runtime, thus removing the need for a lock.
3086 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3087 attempting to decrypt each encrypted key in turn. Add support to
3091 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3093 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3096 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3099 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3100 key into the same file any more.
3103 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3106 *) Add -utf8 command line and config file option to 'ca'.
3107 [Stefan <stf@udoma.org]
3109 *) Removed the macro des_crypt(), as it seems to conflict with some
3110 libraries. Use DES_crypt().
3113 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3114 involves renaming the source and generated shared-libs for
3115 both. The engines will accept the corrected or legacy ids
3116 ('ncipher' and '4758_cca' respectively) when binding. NB,
3117 this only applies when building 'shared'.
3118 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3120 *) Add attribute functions to EVP_PKEY structure. Modify
3121 PKCS12_create() to recognize a CSP name attribute and
3122 use it. Make -CSP option work again in pkcs12 utility.
3125 *) Add new functionality to the bn blinding code:
3126 - automatic re-creation of the BN_BLINDING parameters after
3127 a fixed number of uses (currently 32)
3128 - add new function for parameter creation
3129 - introduce flags to control the update behaviour of the
3130 BN_BLINDING parameters
3131 - hide BN_BLINDING structure
3132 Add a second BN_BLINDING slot to the RSA structure to improve
3133 performance when a single RSA object is shared among several
3137 *) Add support for DTLS.
3138 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3140 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3141 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3144 *) Remove buggy and incompletet DH cert support from
3145 ssl/ssl_rsa.c and ssl/s3_both.c
3148 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3149 the apps/openssl applications.
3152 *) Compile clean with "-Wall -Wmissing-prototypes
3153 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3154 DEBUG_SAFESTACK must also be set.
3157 *) Change ./Configure so that certain algorithms can be disabled by default.
3158 The new counterpiece to "no-xxx" is "enable-xxx".
3160 The patented RC5 and MDC2 algorithms will now be disabled unless
3161 "enable-rc5" and "enable-mdc2", respectively, are specified.
3163 (IDEA remains enabled despite being patented. This is because IDEA
3164 is frequently required for interoperability, and there is no license
3165 fee for non-commercial use. As before, "no-idea" can be used to
3166 avoid this algorithm.)
3170 *) Add processing of proxy certificates (see RFC 3820). This work was
3171 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3172 EGEE (Enabling Grids for E-science in Europe).
3175 *) RC4 performance overhaul on modern architectures/implementations, such
3176 as Intel P4, IA-64 and AMD64.
3179 *) New utility extract-section.pl. This can be used specify an alternative
3180 section number in a pod file instead of having to treat each file as
3181 a separate case in Makefile. This can be done by adding two lines to the
3184 =for comment openssl_section:XXX
3186 The blank line is mandatory.
3190 *) New arguments -certform, -keyform and -pass for s_client and s_server
3191 to allow alternative format key and certificate files and passphrase
3195 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3196 update associated structures and add various utility functions.
3198 Add new policy related verify parameters, include policy checking in
3199 standard verify code. Enhance 'smime' application with extra parameters
3200 to support policy checking and print out.
3203 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3204 Nehemiah processors. These extensions support AES encryption in hardware
3205 as well as RNG (though RNG support is currently disabled).
3206 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3208 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3211 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3212 [Andy Polyakov and a number of other people]
3214 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3215 implementation contributed by IBM.
3216 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3218 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3219 exponent rather than 'unsigned long'. There is a corresponding change to
3220 the new 'rsa_keygen' element of the RSA_METHOD structure.
3221 [Jelte Jansen, Geoff Thorpe]
3223 *) Functionality for creating the initial serial number file is now
3224 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3226 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3227 number file to 1, which is bound to cause problems. To avoid
3228 the problems while respecting compatibility between different 0.9.7
3229 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3230 CA.pl for serial number initialization. With the new release 0.9.8,
3231 we can fix the problem directly in the 'ca' utility.)
3234 *) Reduced header interdepencies by declaring more opaque objects in
3235 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3236 give fewer recursive includes, which could break lazy source code - so
3237 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3238 developers should define this symbol when building and using openssl to
3239 ensure they track the recommended behaviour, interfaces, [etc], but
3240 backwards-compatible behaviour prevails when this isn't defined.
3243 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3246 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3247 This will generate a random key of the appropriate length based on the
3248 cipher context. The EVP_CIPHER can provide its own random key generation
3249 routine to support keys of a specific form. This is used in the des and
3250 3des routines to generate a key of the correct parity. Update S/MIME
3251 code to use new functions and hence generate correct parity DES keys.
3252 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3253 valid (weak or incorrect parity).
3256 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3257 as looking them up. This is useful when the verified structure may contain
3258 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3259 present unless the new PKCS7_NO_CRL flag is asserted.
3262 *) Extend ASN1 oid configuration module. It now additionally accepts the
3265 shortName = some long name, 1.2.3.4
3268 *) Reimplemented the BN_CTX implementation. There is now no more static
3269 limitation on the number of variables it can handle nor the depth of the
3270 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3271 information can now expand as required, and rather than having a single
3272 static array of bignums, BN_CTX now uses a linked-list of such arrays
3273 allowing it to expand on demand whilst maintaining the usefulness of
3274 BN_CTX's "bundling".
3277 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3278 to allow all RSA operations to function using a single BN_CTX.
3281 *) Preliminary support for certificate policy evaluation and checking. This
3282 is initially intended to pass the tests outlined in "Conformance Testing
3283 of Relying Party Client Certificate Path Processing Logic" v1.07.
3286 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3287 remained unused and not that useful. A variety of other little bignum
3288 tweaks and fixes have also been made continuing on from the audit (see
3292 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3293 associated ASN1, EVP and SSL functions and old ASN1 macros.
3296 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3297 and this should never fail. So the return value from the use of
3298 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3299 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3302 *) BN_CTX_get() should return zero-valued bignums, providing the same
3303 initialised value as BN_new().
3304 [Geoff Thorpe, suggested by Ulf Möller]
3306 *) Support for inhibitAnyPolicy certificate extension.
3309 *) An audit of the BIGNUM code is underway, for which debugging code is
3310 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3311 is considered valid when processing BIGNUMs, and causes execution to
3312 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3313 further steps are taken to deliberately pollute unused data in BIGNUM
3314 structures to try and expose faulty code further on. For now, openssl will
3315 (in its default mode of operation) continue to tolerate the inconsistent
3316 forms that it has tolerated in the past, but authors and packagers should
3317 consider trying openssl and their own applications when compiled with
3318 these debugging symbols defined. It will help highlight potential bugs in
3319 their own code, and will improve the test coverage for OpenSSL itself. At
3320 some point, these tighter rules will become openssl's default to improve
3321 maintainability, though the assert()s and other overheads will remain only
3322 in debugging configurations. See bn.h for more details.
3323 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3325 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3326 that can only be obtained through BN_CTX_new() (which implicitly
3327 initialises it). The presence of this function only made it possible
3328 to overwrite an existing structure (and cause memory leaks).
3331 *) Because of the callback-based approach for implementing LHASH as a
3332 template type, lh_insert() adds opaque objects to hash-tables and
3333 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3334 to clean up those corresponding objects before destroying the hash table
3335 (and losing the object pointers). So some over-zealous constifications in
3336 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3337 objects as "const" and the lh_doall[_arg] callback wrappers are not
3338 prototyped to have "const" restrictions on the object pointers they are
3339 given (and so aren't required to cast them away any more).
3342 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3343 (speed) prefers to use its own implementation. The two implementations
3344 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3345 its object type properly exposed (MS_TM) instead of casting to/from "char
3346 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3347 aren't necessarily the greatest nomenclatures - but this is what was used
3348 internally to the implementation so I've used that for now.
3351 *) Ensure that deprecated functions do not get compiled when
3352 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3353 the self-tests were still using deprecated key-generation functions so
3354 these have been updated also.
3357 *) Reorganise PKCS#7 code to separate the digest location functionality
3358 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3359 New function PKCS7_set_digest() to set the digest type for PKCS#7
3360 digestedData type. Add additional code to correctly generate the
3361 digestedData type and add support for this type in PKCS7 initialization
3365 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3366 structure of type "other".
3369 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3370 sure the loop does correctly stop and breaking ("division by zero")
3371 modulus operations are not performed. The (pre-generated) prime
3372 table crypto/bn/bn_prime.h was already correct, but it could not be
3373 re-generated on some platforms because of the "division by zero"
3374 situation in the script.
3375 [Ralf S. Engelschall]
3377 *) Update support for ECC-based TLS ciphersuites according to
3378 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3379 SHA-1 now is only used for "small" curves (where the
3380 representation of a field element takes up to 24 bytes); for
3381 larger curves, the field element resulting from ECDH is directly
3382 used as premaster secret.
3383 [Douglas Stebila (Sun Microsystems Laboratories)]
3385 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3386 curve secp160r1 to the tests.
3387 [Douglas Stebila (Sun Microsystems Laboratories)]
3389 *) Add the possibility to load symbols globally with DSO.
3390 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3392 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3393 control of the error stack.
3396 *) Add support for STORE in ENGINE.
3399 *) Add the STORE type. The intention is to provide a common interface
3400 to certificate and key stores, be they simple file-based stores, or
3401 HSM-type store, or LDAP stores, or...
3402 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3405 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3406 pass a list of arguments to any function as well as provide a way
3407 for a function to pass data back to the caller.
3410 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3411 works like BUF_strdup() but can be used to duplicate a portion of
3412 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3416 *) Add the function sk_find_ex() which works like sk_find(), but will
3417 return an index to an element even if an exact match couldn't be
3418 found. The index is guaranteed to point at the element where the
3419 searched-for key would be inserted to preserve sorting order.
3422 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3423 takes an extra flags argument for optional functionality. Currently,
3424 the following flags are defined:
3426 OBJ_BSEARCH_VALUE_ON_NOMATCH
3427 This one gets OBJ_bsearch_ex() to return a pointer to the first
3428 element where the comparing function returns a negative or zero
3431 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3432 This one gets OBJ_bsearch_ex() to return a pointer to the first
3433 element where the comparing function returns zero. This is useful
3434 if there are more than one element where the comparing function
3438 *) Make it possible to create self-signed certificates with 'openssl ca'
3439 in such a way that the self-signed certificate becomes part of the
3440 CA database and uses the same mechanisms for serial number generation
3441 as all other certificate signing. The new flag '-selfsign' enables
3442 this functionality. Adapt CA.sh and CA.pl.in.
3445 *) Add functionality to check the public key of a certificate request
3446 against a given private. This is useful to check that a certificate
3447 request can be signed by that key (self-signing).
3450 *) Make it possible to have multiple active certificates with the same
3451 subject in the CA index file. This is done only if the keyword
3452 'unique_subject' is set to 'no' in the main CA section (default
3453 if 'CA_default') of the configuration file. The value is saved
3454 with the database itself in a separate index attribute file,
3455 named like the index file with '.attr' appended to the name.
3458 *) Generate muti valued AVAs using '+' notation in config files for
3462 *) Support for nameConstraints certificate extension.
3465 *) Support for policyConstraints certificate extension.
3468 *) Support for policyMappings certificate extension.
3471 *) Make sure the default DSA_METHOD implementation only uses its
3472 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3473 and change its own handlers to be NULL so as to remove unnecessary
3474 indirection. This lets alternative implementations fallback to the
3475 default implementation more easily.
3478 *) Support for directoryName in GeneralName related extensions
3482 *) Make it possible to link applications using Makefile.shared.
3483 Make that possible even when linking against static libraries!
3486 *) Support for single pass processing for S/MIME signing. This now
3487 means that S/MIME signing can be done from a pipe, in addition
3488 cleartext signing (multipart/signed type) is effectively streaming
3489 and the signed data does not need to be all held in memory.
3491 This is done with a new flag PKCS7_STREAM. When this flag is set
3492 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3493 is done after the data is output (and digests calculated) in
3494 SMIME_write_PKCS7().
3497 *) Add full support for -rpath/-R, both in shared libraries and
3498 applications, at least on the platforms where it's known how
3502 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3503 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3504 will now compute a table of multiples of the generator that
3505 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3506 faster (notably in the case of a single point multiplication,
3507 scalar * generator).
3508 [Nils Larsch, Bodo Moeller]
3510 *) IPv6 support for certificate extensions. The various extensions
3511 which use the IP:a.b.c.d can now take IPv6 addresses using the
3512 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3516 *) Added an ENGINE that implements RSA by performing private key
3517 exponentiations with the GMP library. The conversions to and from
3518 GMP's mpz_t format aren't optimised nor are any montgomery forms
3519 cached, and on x86 it appears OpenSSL's own performance has caught up.
3520 However there are likely to be other architectures where GMP could
3521 provide a boost. This ENGINE is not built in by default, but it can be
3522 specified at Configure time and should be accompanied by the necessary
3523 linker additions, eg;
3524 ./config -DOPENSSL_USE_GMP -lgmp
3527 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3528 testing availability of engines with "-t" - the old behaviour is
3529 produced by increasing the feature's verbosity with "-tt".
3532 *) ECDSA routines: under certain error conditions uninitialized BN objects
3533 could be freed. Solution: make sure initialization is performed early
3534 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3538 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3539 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3540 software implementations. For DSA and DH, parameter generation can
3541 also be overriden by providing the appropriate method callbacks.
3544 *) Change the "progress" mechanism used in key-generation and
3545 primality testing to functions that take a new BN_GENCB pointer in
3546 place of callback/argument pairs. The new API functions have "_ex"
3547 postfixes and the older functions are reimplemented as wrappers for
3548 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3549 declarations of the old functions to help (graceful) attempts to
3550 migrate to the new functions. Also, the new key-generation API
3551 functions operate on a caller-supplied key-structure and return
3552 success/failure rather than returning a key or NULL - this is to
3553 help make "keygen" another member function of RSA_METHOD etc.
3555 Example for using the new callback interface:
3557 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3561 BN_GENCB_set(&my_cb, my_callback, my_arg);
3563 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3564 /* For the meaning of a, b in calls to my_callback(), see the
3565 * documentation of the function that calls the callback.
3566 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3567 * my_callback should return 1 if it wants BN_is_prime_ex()
3568 * to continue, or 0 to stop.
3573 *) Change the ZLIB compression method to be stateful, and make it
3574 available to TLS with the number defined in
3575 draft-ietf-tls-compression-04.txt.
3578 *) Add the ASN.1 structures and functions for CertificatePair, which
3579 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3581 CertificatePair ::= SEQUENCE {
3582 forward [0] Certificate OPTIONAL,
3583 reverse [1] Certificate OPTIONAL,
3584 -- at least one of the pair shall be present -- }
3586 Also implement the PEM functions to read and write certificate
3587 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3589 This needed to be defined, mostly for the sake of the LDAP
3590 attribute crossCertificatePair, but may prove useful elsewhere as
3594 *) Make it possible to inhibit symlinking of shared libraries in
3595 Makefile.shared, for Cygwin's sake.
3598 *) Extend the BIGNUM API by creating a function
3599 void BN_set_negative(BIGNUM *a, int neg);
3600 and a macro that behave like
3601 int BN_is_negative(const BIGNUM *a);
3603 to avoid the need to access 'a->neg' directly in applications.
3606 *) Implement fast modular reduction for pseudo-Mersenne primes
3607 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3608 EC_GROUP_new_curve_GFp() will now automatically use this
3610 [Nils Larsch <nla@trustcenter.de>]
3612 *) Add new lock type (CRYPTO_LOCK_BN).
3615 *) Change the ENGINE framework to automatically load engines
3616 dynamically from specific directories unless they could be
3617 found to already be built in or loaded. Move all the
3618 current engines except for the cryptodev one to a new
3620 The engines in engines/ are built as shared libraries if
3621 the "shared" options was given to ./Configure or ./config.
3622 Otherwise, they are inserted in libcrypto.a.
3623 /usr/local/ssl/engines is the default directory for dynamic
3624 engines, but that can be overriden at configure time through
3625 the usual use of --prefix and/or --openssldir, and at run
3626 time with the environment variable OPENSSL_ENGINES.
3627 [Geoff Thorpe and Richard Levitte]
3629 *) Add Makefile.shared, a helper makefile to build shared
3630 libraries. Addapt Makefile.org.
3633 *) Add version info to Win32 DLLs.
3634 [Peter 'Luna' Runestig" <peter@runestig.com>]
3636 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3637 can be added using this API to created arbitrary PKCS#12
3638 files while avoiding the low level API.
3640 New options to PKCS12_create(), key or cert can be NULL and
3641 will then be omitted from the output file. The encryption
3642 algorithm NIDs can be set to -1 for no encryption, the mac
3643 iteration count can be set to 0 to omit the mac.
3645 Enhance pkcs12 utility by making the -nokeys and -nocerts
3646 options work when creating a PKCS#12 file. New option -nomac
3647 to omit the mac, NONE can be set for an encryption algorithm.
3648 New code is modified to use the enhanced PKCS12_create()
3649 instead of the low level API.
3652 *) Extend ASN1 encoder to support indefinite length constructed
3653 encoding. This can output sequences tags and octet strings in
3654 this form. Modify pk7_asn1.c to support indefinite length
3655 encoding. This is experimental and needs additional code to
3656 be useful, such as an ASN1 bio and some enhanced streaming
3659 Extend template encode functionality so that tagging is passed
3660 down to the template encoder.
3663 *) Let 'openssl req' fail if an argument to '-newkey' is not
3664 recognized instead of using RSA as a default.
3667 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3668 As these are not official, they are not included in "ALL";
3669 the "ECCdraft" ciphersuite group alias can be used to select them.
3670 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3672 *) Add ECDH engine support.
3673 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3675 *) Add ECDH in new directory crypto/ecdh/.
3676 [Douglas Stebila (Sun Microsystems Laboratories)]
3678 *) Let BN_rand_range() abort with an error after 100 iterations
3679 without success (which indicates a broken PRNG).
3682 *) Change BN_mod_sqrt() so that it verifies that the input value
3683 is really the square of the return value. (Previously,
3684 BN_mod_sqrt would show GIGO behaviour.)
3687 *) Add named elliptic curves over binary fields from X9.62, SECG,
3688 and WAP/WTLS; add OIDs that were still missing.
3690 [Sheueling Chang Shantz and Douglas Stebila
3691 (Sun Microsystems Laboratories)]
3693 *) Extend the EC library for elliptic curves over binary fields
3694 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3697 EC_GF2m_simple_method
3701 EC_GROUP_new_curve_GF2m
3702 EC_GROUP_set_curve_GF2m
3703 EC_GROUP_get_curve_GF2m
3704 EC_POINT_set_affine_coordinates_GF2m
3705 EC_POINT_get_affine_coordinates_GF2m
3706 EC_POINT_set_compressed_coordinates_GF2m
3708 Point compression for binary fields is disabled by default for
3709 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3712 As binary polynomials are represented as BIGNUMs, various members
3713 of the EC_GROUP and EC_POINT data structures can be shared
3714 between the implementations for prime fields and binary fields;
3715 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3716 are essentially identical to their ..._GFp counterparts.
3717 (For simplicity, the '..._GFp' prefix has been dropped from
3718 various internal method names.)
3720 An internal 'field_div' method (similar to 'field_mul' and
3721 'field_sqr') has been added; this is used only for binary fields.
3723 [Sheueling Chang Shantz and Douglas Stebila
3724 (Sun Microsystems Laboratories)]
3726 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3727 through methods ('mul', 'precompute_mult').
3729 The generic implementations (now internally called 'ec_wNAF_mul'
3730 and 'ec_wNAF_precomputed_mult') remain the default if these
3731 methods are undefined.
3733 [Sheueling Chang Shantz and Douglas Stebila
3734 (Sun Microsystems Laboratories)]
3736 *) New function EC_GROUP_get_degree, which is defined through
3737 EC_METHOD. For curves over prime fields, this returns the bit
3738 length of the modulus.
3740 [Sheueling Chang Shantz and Douglas Stebila
3741 (Sun Microsystems Laboratories)]
3743 *) New functions EC_GROUP_dup, EC_POINT_dup.
3744 (These simply call ..._new and ..._copy).
3746 [Sheueling Chang Shantz and Douglas Stebila
3747 (Sun Microsystems Laboratories)]
3749 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3750 Polynomials are represented as BIGNUMs (where the sign bit is not
3751 used) in the following functions [macros]:
3754 BN_GF2m_sub [= BN_GF2m_add]
3755 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3756 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3757 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3759 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3760 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3761 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3762 BN_GF2m_cmp [= BN_ucmp]
3764 (Note that only the 'mod' functions are actually for fields GF(2^m).
3765 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3767 For some functions, an the irreducible polynomial defining a
3768 field can be given as an 'unsigned int[]' with strictly
3769 decreasing elements giving the indices of those bits that are set;
3770 i.e., p[] represents the polynomial
3771 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3773 p[0] > p[1] > ... > p[k] = 0.
3774 This applies to the following functions:
3779 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3780 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3782 BN_GF2m_mod_sqrt_arr
3783 BN_GF2m_mod_solve_quad_arr
3787 Conversion can be performed by the following functions:
3792 bntest.c has additional tests for binary polynomial arithmetic.
3794 Two implementations for BN_GF2m_mod_div() are available.
3795 The default algorithm simply uses BN_GF2m_mod_inv() and
3796 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3797 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3798 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3800 [Sheueling Chang Shantz and Douglas Stebila
3801 (Sun Microsystems Laboratories)]
3803 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3804 functionality is disabled at compile-time.
3805 [Douglas Stebila <douglas.stebila@sun.com>]
3807 *) Change default behaviour of 'openssl asn1parse' so that more
3808 information is visible when viewing, e.g., a certificate:
3810 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3811 mode the content of non-printable OCTET STRINGs is output in a
3812 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3813 avoid the appearance of a printable string.
3814 [Nils Larsch <nla@trustcenter.de>]
3816 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3818 EC_GROUP_set_asn1_flag()
3819 EC_GROUP_get_asn1_flag()
3820 EC_GROUP_set_point_conversion_form()
3821 EC_GROUP_get_point_conversion_form()
3822 These control ASN1 encoding details:
3823 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3824 has been set to OPENSSL_EC_NAMED_CURVE.
3825 - Points are encoded in uncompressed form by default; options for
3826 asn1_for are as for point2oct, namely
3827 POINT_CONVERSION_COMPRESSED
3828 POINT_CONVERSION_UNCOMPRESSED
3829 POINT_CONVERSION_HYBRID
3831 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3834 EC_GROUP_get0_seed()
3835 EC_GROUP_get_seed_len()
3836 This is used only for ASN1 purposes (so far).
3837 [Nils Larsch <nla@trustcenter.de>]
3839 *) Add 'field_type' member to EC_METHOD, which holds the NID
3840 of the appropriate field type OID. The new function
3841 EC_METHOD_get_field_type() returns this value.
3842 [Nils Larsch <nla@trustcenter.de>]
3847 EC_POINT_point2hex()
3848 EC_POINT_hex2point()
3849 providing useful interfaces to EC_POINT_point2oct() and
3850 EC_POINT_oct2point().
3851 [Nils Larsch <nla@trustcenter.de>]
3853 *) Change internals of the EC library so that the functions
3854 EC_GROUP_set_generator()
3855 EC_GROUP_get_generator()
3856 EC_GROUP_get_order()
3857 EC_GROUP_get_cofactor()
3858 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3859 to methods, which would lead to unnecessary code duplication when
3860 adding different types of curves.
3861 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3863 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3864 arithmetic, and such that modified wNAFs are generated
3865 (which avoid length expansion in many cases).
3868 *) Add a function EC_GROUP_check_discriminant() (defined via
3869 EC_METHOD) that verifies that the curve discriminant is non-zero.
3871 Add a function EC_GROUP_check() that makes some sanity tests
3872 on a EC_GROUP, its generator and order. This includes
3873 EC_GROUP_check_discriminant().
3874 [Nils Larsch <nla@trustcenter.de>]
3876 *) Add ECDSA in new directory crypto/ecdsa/.
3878 Add applications 'openssl ecparam' and 'openssl ecdsa'
3879 (these are based on 'openssl dsaparam' and 'openssl dsa').
3881 ECDSA support is also included in various other files across the
3882 library. Most notably,
3883 - 'openssl req' now has a '-newkey ecdsa:file' option;
3884 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3885 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3886 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3887 them suitable for ECDSA where domain parameters must be
3888 extracted before the specific public key;
3889 - ECDSA engine support has been added.
3890 [Nils Larsch <nla@trustcenter.de>]
3892 *) Include some named elliptic curves, and add OIDs from X9.62,
3893 SECG, and WAP/WTLS. Each curve can be obtained from the new
3895 EC_GROUP_new_by_curve_name(),
3896 and the list of available named curves can be obtained with
3897 EC_get_builtin_curves().
3898 Also add a 'curve_name' member to EC_GROUP objects, which can be
3900 EC_GROUP_set_curve_name()
3901 EC_GROUP_get_curve_name()
3902 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3904 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3905 was actually never needed) and in BN_mul(). The removal in BN_mul()
3906 required a small change in bn_mul_part_recursive() and the addition
3907 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3908 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3909 bn_sub_words() and bn_add_words() except they take arrays with
3913 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3915 *) Cleanse PEM buffers before freeing them since they may contain
3917 [Benjamin Bennett <ben@psc.edu>]
3919 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3920 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3921 authentication-only ciphersuites.
3924 *) Since AES128 and AES256 share a single mask bit in the logic of
3925 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3926 kludge to work properly if AES128 is available and AES256 isn't.
3929 *) Expand security boundary to match 1.1.1 module.
3932 *) Remove redundant features: hash file source, editing of test vectors
3933 modify fipsld to use external fips_premain.c signature.
3936 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3937 run algorithm test programs.
3940 *) Make algorithm test programs more tolerant of whitespace.
3943 *) Have SSL/TLS server implementation tolerate "mismatched" record
3944 protocol version while receiving ClientHello even if the
3945 ClientHello is fragmented. (The server can't insist on the
3946 particular protocol version it has chosen before the ServerHello
3947 message has informed the client about his choice.)
3950 *) Load error codes if they are not already present instead of using a
3951 static variable. This allows them to be cleanly unloaded and reloaded.
3954 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3956 *) Introduce limits to prevent malicious keys being able to
3957 cause a denial of service. (CVE-2006-2940)
3958 [Steve Henson, Bodo Moeller]
3960 *) Fix ASN.1 parsing of certain invalid structures that can result
3961 in a denial of service. (CVE-2006-2937) [Steve Henson]
3963 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3964 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3966 *) Fix SSL client code which could crash if connecting to a
3967 malicious SSLv2 server. (CVE-2006-4343)
3968 [Tavis Ormandy and Will Drewry, Google Security Team]
3970 *) Change ciphersuite string processing so that an explicit
3971 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3972 will no longer include "AES128-SHA"), and any other similar
3973 ciphersuite (same bitmap) from *other* protocol versions (so that
3974 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3975 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3976 changes from 0.9.8b and 0.9.8d.
3979 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3981 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3982 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3984 *) Change the Unix randomness entropy gathering to use poll() when
3985 possible instead of select(), since the latter has some
3986 undesirable limitations.
3987 [Darryl Miles via Richard Levitte and Bodo Moeller]
3989 *) Disable rogue ciphersuites:
3991 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3992 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3993 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3995 The latter two were purportedly from
3996 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3999 Also deactive the remaining ciphersuites from
4000 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4001 unofficial, and the ID has long expired.
4004 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4005 dual-core machines) and other potential thread-safety issues.
4008 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4010 *) Adapt fipsld and the build system to link against the validated FIPS
4011 module in FIPS mode.
4014 *) Fixes for VC++ 2005 build under Windows.
4017 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4018 from a Windows bash shell such as MSYS. It is autodetected from the
4019 "config" script when run from a VC++ environment. Modify standard VC++
4020 build to use fipscanister.o from the GNU make build.
4023 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4025 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4026 The value now differs depending on if you build for FIPS or not.
4027 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4028 safely run with a non-FIPSed libcrypto, as it may crash because of
4029 the difference induced by this change.
4032 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4034 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4035 (part of SSL_OP_ALL). This option used to disable the
4036 countermeasure against man-in-the-middle protocol-version
4037 rollback in the SSL 2.0 server implementation, which is a bad
4038 idea. (CVE-2005-2969)
4040 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4041 for Information Security, National Institute of Advanced Industrial
4042 Science and Technology [AIST], Japan)]
4044 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4045 mainly for FIPS compliance and not fully integrated at this stage.
4048 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4049 the exponentiation using a fixed-length exponent. (Otherwise,
4050 the information leaked through timing could expose the secret key
4051 after many signatures; cf. Bleichenbacher's attack on DSA with
4055 *) Make a new fixed-window mod_exp implementation the default for
4056 RSA, DSA, and DH private-key operations so that the sequence of
4057 squares and multiplies and the memory access pattern are
4058 independent of the particular secret key. This will mitigate
4059 cache-timing and potential related attacks.
4061 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4062 and this is automatically used by BN_mod_exp_mont() if the new flag
4063 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4064 will use this BN flag for private exponents unless the flag
4065 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4066 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4068 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4070 *) Change the client implementation for SSLv23_method() and
4071 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4072 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4073 (Previously, the SSL 2.0 backwards compatible Client Hello
4074 message format would be used even with SSL_OP_NO_SSLv2.)
4077 *) Add support for smime-type MIME parameter in S/MIME messages which some
4081 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4082 a threadsafe manner. Modify rsa code to use new function and add calls
4083 to dsa and dh code (which had race conditions before).
4086 *) Include the fixed error library code in the C error file definitions
4087 instead of fixing them up at runtime. This keeps the error code
4088 structures constant.
4091 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4093 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4096 *) Fixes for newer kerberos headers. NB: the casts are needed because
4097 the 'length' field is signed on one version and unsigned on another
4098 with no (?) obvious way to tell the difference, without these VC++
4099 complains. Also the "definition" of FAR (blank) is no longer included
4100 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4101 some needed definitions.
4104 *) Undo Cygwin change.
4107 *) Added support for proxy certificates according to RFC 3820.
4108 Because they may be a security thread to unaware applications,
4109 they must be explicitely allowed in run-time. See
4110 docs/HOWTO/proxy_certificates.txt for further information.
4113 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4115 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4116 server and client random values. Previously
4117 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4118 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4120 This change has negligible security impact because:
4122 1. Server and client random values still have 24 bytes of pseudo random
4125 2. Server and client random values are sent in the clear in the initial
4128 3. The master secret is derived using the premaster secret (48 bytes in
4129 size for static RSA ciphersuites) as well as client server and random
4132 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4135 [Stephen Henson, reported by UK NISCC]
4137 *) Use Windows randomness collection on Cygwin.
4140 *) Fix hang in EGD/PRNGD query when communication socket is closed
4141 prematurely by EGD/PRNGD.
4142 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4144 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4147 *) Back-port of selected performance improvements from development
4148 branch, as well as improved support for PowerPC platforms.
4151 *) Add lots of checks for memory allocation failure, error codes to indicate
4152 failure and freeing up memory if a failure occurs.
4153 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4155 *) Add new -passin argument to dgst.
4158 *) Perform some character comparisons of different types in X509_NAME_cmp:
4159 this is needed for some certificates that reencode DNs into UTF8Strings
4160 (in violation of RFC3280) and can't or wont issue name rollover
4164 *) Make an explicit check during certificate validation to see that
4165 the CA setting in each certificate on the chain is correct. As a
4166 side effect always do the following basic checks on extensions,
4167 not just when there's an associated purpose to the check:
4169 - if there is an unhandled critical extension (unless the user
4170 has chosen to ignore this fault)
4171 - if the path length has been exceeded (if one is set at all)
4172 - that certain extensions fit the associated purpose (if one has
4176 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4178 *) Avoid a race condition when CRLs are checked in a multi threaded
4179 environment. This would happen due to the reordering of the revoked
4180 entries during signature checking and serial number lookup. Now the
4181 encoding is cached and the serial number sort performed under a lock.
4182 Add new STACK function sk_is_sorted().
4185 *) Add Delta CRL to the extension code.
4188 *) Various fixes to s3_pkt.c so alerts are sent properly.
4189 [David Holmes <d.holmes@f5.com>]
4191 *) Reduce the chances of duplicate issuer name and serial numbers (in
4192 violation of RFC3280) using the OpenSSL certificate creation utilities.
4193 This is done by creating a random 64 bit value for the initial serial
4194 number when a serial number file is created or when a self signed
4195 certificate is created using 'openssl req -x509'. The initial serial
4196 number file is created using 'openssl x509 -next_serial' in CA.pl
4197 rather than being initialized to 1.
4200 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4202 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4203 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4204 [Joe Orton, Steve Henson]
4206 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4208 [Joe Orton, Steve Henson]
4210 *) Make it possible to have multiple active certificates with the same
4211 subject in the CA index file. This is done only if the keyword
4212 'unique_subject' is set to 'no' in the main CA section (default
4213 if 'CA_default') of the configuration file. The value is saved
4214 with the database itself in a separate index attribute file,
4215 named like the index file with '.attr' appended to the name.
4218 *) X509 verify fixes. Disable broken certificate workarounds when
4219 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4220 keyUsage extension present. Don't accept CRLs with unhandled critical
4221 extensions: since verify currently doesn't process CRL extensions this
4222 rejects a CRL with *any* critical extensions. Add new verify error codes
4226 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4227 A clarification of RFC2560 will require the use of OCTET STRINGs and
4228 some implementations cannot handle the current raw format. Since OpenSSL
4229 copies and compares OCSP nonces as opaque blobs without any attempt at
4230 parsing them this should not create any compatibility issues.
4233 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4234 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4235 this HMAC (and other) operations are several times slower than OpenSSL
4239 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4240 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4242 *) Use the correct content when signing type "other".
4245 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4247 *) Fix various bugs revealed by running the NISCC test suite:
4249 Stop out of bounds reads in the ASN1 code when presented with
4250 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4252 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4254 If verify callback ignores invalid public key errors don't try to check
4255 certificate signature with the NULL public key.
4259 *) New -ignore_err option in ocsp application to stop the server
4260 exiting on the first error in a request.
4263 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4264 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4268 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4269 extra data after the compression methods not only for TLS 1.0
4270 but also for SSL 3.0 (as required by the specification).
4271 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4273 *) Change X509_certificate_type() to mark the key as exported/exportable
4274 when it's 512 *bits* long, not 512 bytes.
4277 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4278 blocks during encryption.
4281 *) Various fixes to base64 BIO and non blocking I/O. On write
4282 flushes were not handled properly if the BIO retried. On read
4283 data was not being buffered properly and had various logic bugs.
4284 This also affects blocking I/O when the data being decoded is a
4288 *) Various S/MIME bugfixes and compatibility changes:
4289 output correct application/pkcs7 MIME type if
4290 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4291 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4292 of files as .eml work). Correctly handle very long lines in MIME
4296 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4298 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4299 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4300 a protocol version number mismatch like a decryption error
4301 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4304 *) Turn on RSA blinding by default in the default implementation
4305 to avoid a timing attack. Applications that don't want it can call
4306 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4307 They would be ill-advised to do so in most cases.
4308 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4310 *) Change RSA blinding code so that it works when the PRNG is not
4311 seeded (in this case, the secret RSA exponent is abused as
4312 an unpredictable seed -- if it is not unpredictable, there
4313 is no point in blinding anyway). Make RSA blinding thread-safe
4314 by remembering the creator's thread ID in rsa->blinding and
4315 having all other threads use local one-time blinding factors
4316 (this requires more computation than sharing rsa->blinding, but
4317 avoids excessive locking; and if an RSA object is not shared
4318 between threads, blinding will still be very fast).
4321 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4322 ENGINE as defaults for all supported algorithms irrespective of
4323 the 'flags' parameter. 'flags' is now honoured, so applications
4324 should make sure they are passing it correctly.
4327 *) Target "mingw" now allows native Windows code to be generated in
4328 the Cygwin environment as well as with the MinGW compiler.
4331 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4333 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4334 via timing by performing a MAC computation even if incorrrect
4335 block cipher padding has been found. This is a countermeasure
4336 against active attacks where the attacker has to distinguish
4337 between bad padding and a MAC verification error. (CVE-2003-0078)
4339 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4340 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4341 Martin Vuagnoux (EPFL, Ilion)]
4343 *) Make the no-err option work as intended. The intention with no-err
4344 is not to have the whole error stack handling routines removed from
4345 libcrypto, it's only intended to remove all the function name and
4346 reason texts, thereby removing some of the footprint that may not
4347 be interesting if those errors aren't displayed anyway.
4349 NOTE: it's still possible for any application or module to have it's
4350 own set of error texts inserted. The routines are there, just not
4351 used by default when no-err is given.
4354 *) Add support for FreeBSD on IA64.
4355 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4357 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4358 Kerberos function mit_des_cbc_cksum(). Before this change,
4359 the value returned by DES_cbc_cksum() was like the one from
4360 mit_des_cbc_cksum(), except the bytes were swapped.
4361 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4363 *) Allow an application to disable the automatic SSL chain building.
4364 Before this a rather primitive chain build was always performed in
4365 ssl3_output_cert_chain(): an application had no way to send the
4366 correct chain if the automatic operation produced an incorrect result.
4368 Now the chain builder is disabled if either:
4370 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4372 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4374 The reasoning behind this is that an application would not want the
4375 auto chain building to take place if extra chain certificates are
4376 present and it might also want a means of sending no additional
4377 certificates (for example the chain has two certificates and the
4381 *) Add the possibility to build without the ENGINE framework.
4382 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4384 *) Under Win32 gmtime() can return NULL: check return value in
4385 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4388 *) DSA routines: under certain error conditions uninitialized BN objects
4389 could be freed. Solution: make sure initialization is performed early
4390 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4391 Nils Larsch <nla@trustcenter.de> via PR#459)
4394 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4395 checked on reconnect on the client side, therefore session resumption
4396 could still fail with a "ssl session id is different" error. This
4397 behaviour is masked when SSL_OP_ALL is used due to
4398 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4399 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4400 followup to PR #377.
4403 *) IA-32 assembler support enhancements: unified ELF targets, support
4404 for SCO/Caldera platforms, fix for Cygwin shared build.
4407 *) Add support for FreeBSD on sparc64. As a consequence, support for
4408 FreeBSD on non-x86 processors is separate from x86 processors on
4409 the config script, much like the NetBSD support.
4410 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4412 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4414 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4417 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4418 code (06) was taken as the first octet of the session ID and the last
4419 octet was ignored consequently. As a result SSLv2 client side session
4420 caching could not have worked due to the session ID mismatch between
4422 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4426 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4427 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4431 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4432 seems that in spite of existing for more than a year, many application
4433 author have done nothing to provide the necessary callbacks, which
4434 means that this particular engine will not work properly anywhere.
4435 This is a very unfortunate situation which forces us, in the name
4436 of usability, to give the hw_ncipher.c a static lock, which is part
4438 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4439 appear in 0.9.8 or later. We EXPECT application authors to have
4440 dealt properly with this when 0.9.8 is released (unless we actually
4441 make such changes in the libcrypto locking code that changes will
4442 have to be made anyway).
4445 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4446 octets have been read, EOF or an error occurs. Without this change
4447 some truncated ASN1 structures will not produce an error.
4450 *) Disable Heimdal support, since it hasn't been fully implemented.
4451 Still give the possibility to force the use of Heimdal, but with
4452 warnings and a request that patches get sent to openssl-dev.
4455 *) Add the VC-CE target, introduce the WINCE sysname, and add
4456 INSTALL.WCE and appropriate conditionals to make it build.
4457 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4459 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4460 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4461 edit numbers of the version.
4462 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4464 *) Introduce safe string copy and catenation functions
4465 (BUF_strlcpy() and BUF_strlcat()).
4466 [Ben Laurie (CHATS) and Richard Levitte]
4468 *) Avoid using fixed-size buffers for one-line DNs.
4469 [Ben Laurie (CHATS)]
4471 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4472 resizing buffers containing secrets, and use where appropriate.
4473 [Ben Laurie (CHATS)]
4475 *) Avoid using fixed size buffers for configuration file location.
4476 [Ben Laurie (CHATS)]
4478 *) Avoid filename truncation for various CA files.
4479 [Ben Laurie (CHATS)]
4481 *) Use sizeof in preference to magic numbers.
4482 [Ben Laurie (CHATS)]
4484 *) Avoid filename truncation in cert requests.
4485 [Ben Laurie (CHATS)]
4487 *) Add assertions to check for (supposedly impossible) buffer
4489 [Ben Laurie (CHATS)]
4491 *) Don't cache truncated DNS entries in the local cache (this could
4492 potentially lead to a spoofing attack).
4493 [Ben Laurie (CHATS)]
4495 *) Fix various buffers to be large enough for hex/decimal
4496 representations in a platform independent manner.
4497 [Ben Laurie (CHATS)]
4499 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4500 resizing buffers containing secrets, and use where appropriate.
4501 [Ben Laurie (CHATS)]
4503 *) Add BIO_indent() to avoid much slightly worrying code to do
4505 [Ben Laurie (CHATS)]
4507 *) Convert sprintf()/BIO_puts() to BIO_printf().
4508 [Ben Laurie (CHATS)]
4510 *) buffer_gets() could terminate with the buffer only half
4512 [Ben Laurie (CHATS)]
4514 *) Add assertions to prevent user-supplied crypto functions from
4515 overflowing internal buffers by having large block sizes, etc.
4516 [Ben Laurie (CHATS)]
4518 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4520 [Ben Laurie (CHATS)]
4522 *) Eliminate unused copy of key in RC4.
4523 [Ben Laurie (CHATS)]
4525 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4526 [Ben Laurie (CHATS)]
4528 *) Fix off-by-one error in EGD path.
4529 [Ben Laurie (CHATS)]
4531 *) If RANDFILE path is too long, ignore instead of truncating.
4532 [Ben Laurie (CHATS)]
4534 *) Eliminate unused and incorrectly sized X.509 structure
4536 [Ben Laurie (CHATS)]
4538 *) Eliminate unused and dangerous function knumber().
4539 [Ben Laurie (CHATS)]
4541 *) Eliminate unused and dangerous structure, KSSL_ERR.
4542 [Ben Laurie (CHATS)]
4544 *) Protect against overlong session ID context length in an encoded
4545 session object. Since these are local, this does not appear to be
4547 [Ben Laurie (CHATS)]
4549 *) Change from security patch (see 0.9.6e below) that did not affect
4550 the 0.9.6 release series:
4552 Remote buffer overflow in SSL3 protocol - an attacker could
4553 supply an oversized master key in Kerberos-enabled versions.
4555 [Ben Laurie (CHATS)]
4557 *) Change the SSL kerb5 codes to match RFC 2712.
4560 *) Make -nameopt work fully for req and add -reqopt switch.
4561 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4563 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4564 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4566 *) Make sure tests can be performed even if the corresponding algorithms
4567 have been removed entirely. This was also the last step to make
4568 OpenSSL compilable with DJGPP under all reasonable conditions.
4569 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4571 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4572 to allow version independent disabling of normally unselected ciphers,
4573 which may be activated as a side-effect of selecting a single cipher.
4575 (E.g., cipher list string "RSA" enables ciphersuites that are left
4576 out of "ALL" because they do not provide symmetric encryption.
4577 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4578 [Lutz Jaenicke, Bodo Moeller]
4580 *) Add appropriate support for separate platform-dependent build
4581 directories. The recommended way to make a platform-dependent
4582 build directory is the following (tested on Linux), maybe with
4585 # Place yourself outside of the OpenSSL source tree. In
4586 # this example, the environment variable OPENSSL_SOURCE
4587 # is assumed to contain the absolute OpenSSL source directory.
4588 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4589 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4590 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4591 mkdir -p `dirname $F`
4592 ln -s $OPENSSL_SOURCE/$F $F
4595 To be absolutely sure not to disturb the source tree, a "make clean"
4596 is a good thing. If it isn't successfull, don't worry about it,
4597 it probably means the source directory is very clean.
4600 *) Make sure any ENGINE control commands make local copies of string
4601 pointers passed to them whenever necessary. Otherwise it is possible
4602 the caller may have overwritten (or deallocated) the original string
4603 data when a later ENGINE operation tries to use the stored values.
4604 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4606 *) Improve diagnostics in file reading and command-line digests.
4607 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4609 *) Add AES modes CFB and OFB to the object database. Correct an
4610 error in AES-CFB decryption.
4613 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4614 allows existing EVP_CIPHER_CTX structures to be reused after
4615 calling EVP_*Final(). This behaviour is used by encryption
4616 BIOs and some applications. This has the side effect that
4617 applications must explicitly clean up cipher contexts with
4618 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4621 *) Check the values of dna and dnb in bn_mul_recursive before calling
4622 bn_mul_comba (a non zero value means the a or b arrays do not contain
4623 n2 elements) and fallback to bn_mul_normal if either is not zero.
4626 *) Fix escaping of non-ASCII characters when using the -subj option
4627 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4630 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4631 form for "surname", serialNumber has no short form.
4632 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4633 therefore remove "mail" short name for "internet 7".
4634 The OID for unique identifiers in X509 certificates is
4635 x500UniqueIdentifier, not uniqueIdentifier.
4636 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4639 *) Add an "init" command to the ENGINE config module and auto initialize
4640 ENGINEs. Without any "init" command the ENGINE will be initialized
4641 after all ctrl commands have been executed on it. If init=1 the
4642 ENGINE is initailized at that point (ctrls before that point are run
4643 on the uninitialized ENGINE and after on the initialized one). If
4644 init=0 then the ENGINE will not be iniatialized at all.
4647 *) Fix the 'app_verify_callback' interface so that the user-defined
4648 argument is actually passed to the callback: In the
4649 SSL_CTX_set_cert_verify_callback() prototype, the callback
4650 declaration has been changed from
4653 int (*cb)(X509_STORE_CTX *,void *);
4654 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4655 i=s->ctx->app_verify_callback(&ctx)
4656 has been changed into
4657 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4659 To update applications using SSL_CTX_set_cert_verify_callback(),
4660 a dummy argument can be added to their callback functions.
4661 [D. K. Smetters <smetters@parc.xerox.com>]
4663 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4664 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4666 *) Add and OPENSSL_LOAD_CONF define which will cause
4667 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4668 This allows older applications to transparently support certain
4669 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4670 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4671 load the config file and OPENSSL_add_all_algorithms_conf() which will
4672 always load it have also been added.
4675 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4676 Adjust NIDs and EVP layer.
4677 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4679 *) Config modules support in openssl utility.
4681 Most commands now load modules from the config file,
4682 though in a few (such as version) this isn't done
4683 because it couldn't be used for anything.
4685 In the case of ca and req the config file used is
4686 the same as the utility itself: that is the -config
4687 command line option can be used to specify an
4691 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4692 use "openssl_conf" if filename is NULL use default openssl config file.
4695 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4696 config section name. Add a new flag to tolerate a missing config file
4697 and move code to CONF_modules_load_file().
4700 *) Support for crypto accelerator cards from Accelerated Encryption
4701 Processing, www.aep.ie. (Use engine 'aep')
4702 The support was copied from 0.9.6c [engine] and adapted/corrected
4703 to work with the new engine framework.
4704 [AEP Inc. and Richard Levitte]
4706 *) Support for SureWare crypto accelerator cards from Baltimore
4707 Technologies. (Use engine 'sureware')
4708 The support was copied from 0.9.6c [engine] and adapted
4709 to work with the new engine framework.
4712 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4713 make the newer ENGINE framework commands for the CHIL engine work.
4714 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4716 *) Make it possible to produce shared libraries on ReliantUNIX.
4717 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4719 *) Add the configuration target debug-linux-ppro.
4720 Make 'openssl rsa' use the general key loading routines
4721 implemented in apps.c, and make those routines able to
4722 handle the key format FORMAT_NETSCAPE and the variant
4724 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4726 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4727 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4729 *) Add -keyform to rsautl, and document -engine.
4730 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4732 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4733 BIO_R_NO_SUCH_FILE error code rather than the generic
4734 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4737 *) Add new functions
4739 ERR_peek_last_error_line
4740 ERR_peek_last_error_line_data.
4741 These are similar to
4744 ERR_peek_error_line_data,
4745 but report on the latest error recorded rather than the first one
4746 still in the error queue.
4747 [Ben Laurie, Bodo Moeller]
4749 *) default_algorithms option in ENGINE config module. This allows things
4751 default_algorithms = ALL
4752 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4755 *) Prelminary ENGINE config module.
4758 *) New experimental application configuration code.
4761 *) Change the AES code to follow the same name structure as all other
4762 symmetric ciphers, and behave the same way. Move everything to
4763 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4764 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4766 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4767 [Ben Laurie and Theo de Raadt]
4769 *) Add option to output public keys in req command.
4770 [Massimiliano Pala madwolf@openca.org]
4772 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4773 (up to about 10% better than before for P-192 and P-224).
4776 *) New functions/macros
4778 SSL_CTX_set_msg_callback(ctx, cb)
4779 SSL_CTX_set_msg_callback_arg(ctx, arg)
4780 SSL_set_msg_callback(ssl, cb)
4781 SSL_set_msg_callback_arg(ssl, arg)
4783 to request calling a callback function
4785 void cb(int write_p, int version, int content_type,
4786 const void *buf, size_t len, SSL *ssl, void *arg)
4788 whenever a protocol message has been completely received
4789 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4790 protocol version according to which the SSL library interprets
4791 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4792 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4793 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4794 specification (change_cipher_spec(20), alert(21), handshake(22)).
4795 'buf' and 'len' point to the actual message, 'ssl' to the
4796 SSL object, and 'arg' is the application-defined value set by
4797 SSL[_CTX]_set_msg_callback_arg().
4799 'openssl s_client' and 'openssl s_server' have new '-msg' options
4800 to enable a callback that displays all protocol messages.
4803 *) Change the shared library support so shared libraries are built as
4804 soon as the corresponding static library is finished, and thereby get
4805 openssl and the test programs linked against the shared library.
4806 This still only happens when the keyword "shard" has been given to
4807 the configuration scripts.
4809 NOTE: shared library support is still an experimental thing, and
4810 backward binary compatibility is still not guaranteed.
4811 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4813 *) Add support for Subject Information Access extension.
4814 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4816 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4817 additional bytes when new memory had to be allocated, not just
4818 when reusing an existing buffer.
4821 *) New command line and configuration option 'utf8' for the req command.
4822 This allows field values to be specified as UTF8 strings.
4825 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4826 runs for the former and machine-readable output for the latter.
4829 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4830 of the e-mail address in the DN (i.e., it will go into a certificate
4831 extension only). The new configuration file option 'email_in_dn = no'
4832 has the same effect.
4833 [Massimiliano Pala madwolf@openca.org]
4835 *) Change all functions with names starting with des_ to be starting
4836 with DES_ instead. Add wrappers that are compatible with libdes,
4837 but are named _ossl_old_des_*. Finally, add macros that map the
4838 des_* symbols to the corresponding _ossl_old_des_* if libdes
4839 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4840 desired, the des_* symbols will be mapped to DES_*, with one
4843 Since we provide two compatibility mappings, the user needs to
4844 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4845 compatibility is desired. The default (i.e., when that macro
4846 isn't defined) is OpenSSL 0.9.6c compatibility.
4848 There are also macros that enable and disable the support of old
4849 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4850 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4851 are defined, the default will apply: to support the old des routines.
4853 In either case, one must include openssl/des.h to get the correct
4854 definitions. Do not try to just include openssl/des_old.h, that
4857 NOTE: This is a major break of an old API into a new one. Software
4858 authors are encouraged to switch to the DES_ style functions. Some
4859 time in the future, des_old.h and the libdes compatibility functions
4860 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4861 default), and then completely removed.
4864 *) Test for certificates which contain unsupported critical extensions.
4865 If such a certificate is found during a verify operation it is
4866 rejected by default: this behaviour can be overridden by either
4867 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4868 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4869 X509_supported_extension() has also been added which returns 1 if a
4870 particular extension is supported.
4873 *) Modify the behaviour of EVP cipher functions in similar way to digests
4874 to retain compatibility with existing code.
4877 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4878 compatibility with existing code. In particular the 'ctx' parameter does
4879 not have to be to be initialized before the call to EVP_DigestInit() and
4880 it is tidied up after a call to EVP_DigestFinal(). New function
4881 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4882 EVP_MD_CTX_copy() changed to not require the destination to be
4883 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4884 requires the destination to be valid.
4886 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4887 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4890 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4891 so that complete 'Handshake' protocol structures are kept in memory
4892 instead of overwriting 'msg_type' and 'length' with 'body' data.
4895 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4896 [Massimo Santin via Richard Levitte]
4898 *) Major restructuring to the underlying ENGINE code. This includes
4899 reduction of linker bloat, separation of pure "ENGINE" manipulation
4900 (initialisation, etc) from functionality dealing with implementations
4901 of specific crypto iterfaces. This change also introduces integrated
4902 support for symmetric ciphers and digest implementations - so ENGINEs
4903 can now accelerate these by providing EVP_CIPHER and EVP_MD
4904 implementations of their own. This is detailed in crypto/engine/README
4905 as it couldn't be adequately described here. However, there are a few
4906 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4907 were changed in the original introduction of ENGINE code have now
4908 reverted back - the hooking from this code to ENGINE is now a good
4909 deal more passive and at run-time, operations deal directly with
4910 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4911 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4912 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4913 they were not being used by the framework as there is no concept of a
4914 BIGNUM_METHOD and they could not be generalised to the new
4915 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4916 ENGINE_cpy() has been removed as it cannot be consistently defined in
4920 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4923 *) Change mkdef.pl to sort symbols that get the same entry number,
4924 and make sure the automatically generated functions ERR_load_*
4925 become part of libeay.num as well.
4928 *) New function SSL_renegotiate_pending(). This returns true once
4929 renegotiation has been requested (either SSL_renegotiate() call
4930 or HelloRequest/ClientHello receveived from the peer) and becomes
4931 false once a handshake has been completed.
4932 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4933 sends a HelloRequest, but does not ensure that a handshake takes
4934 place. SSL_renegotiate_pending() is useful for checking if the
4935 client has followed the request.)
4938 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4939 By default, clients may request session resumption even during
4940 renegotiation (if session ID contexts permit); with this option,
4941 session resumption is possible only in the first handshake.
4943 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4944 more bits available for options that should not be part of
4945 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4948 *) Add some demos for certificate and certificate request creation.
4951 *) Make maximum certificate chain size accepted from the peer application
4952 settable (SSL*_get/set_max_cert_list()), as proposed by
4953 "Douglas E. Engert" <deengert@anl.gov>.
4956 *) Add support for shared libraries for Unixware-7
4957 (Boyd Lynn Gerber <gerberb@zenez.com>).
4960 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4961 be done prior to destruction. Use this to unload error strings from
4962 ENGINEs that load their own error strings. NB: This adds two new API
4963 functions to "get" and "set" this destroy handler in an ENGINE.
4966 *) Alter all existing ENGINE implementations (except "openssl" and
4967 "openbsd") to dynamically instantiate their own error strings. This
4968 makes them more flexible to be built both as statically-linked ENGINEs
4969 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4970 Also, add stub code to each that makes building them as self-contained
4971 shared-libraries easier (see README.ENGINE).
4974 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4975 implementations into applications that are completely implemented in
4976 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4977 commands that can be used to configure what shared-library to load and
4978 to control aspects of the way it is handled. Also, made an update to
4979 the README.ENGINE file that brings its information up-to-date and
4980 provides some information and instructions on the "dynamic" ENGINE
4981 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4984 *) Make it possible to unload ranges of ERR strings with a new
4985 "ERR_unload_strings" function.
4988 *) Add a copy() function to EVP_MD.
4991 *) Make EVP_MD routines take a context pointer instead of just the
4992 md_data void pointer.
4995 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4996 that the digest can only process a single chunk of data
4997 (typically because it is provided by a piece of
4998 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4999 is only going to provide a single chunk of data, and hence the
5000 framework needn't accumulate the data for oneshot drivers.
5003 *) As with "ERR", make it possible to replace the underlying "ex_data"
5004 functions. This change also alters the storage and management of global
5005 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5006 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5007 index counters. The API functions that use this state have been changed
5008 to take a "class_index" rather than pointers to the class's local STACK
5009 and counter, and there is now an API function to dynamically create new
5010 classes. This centralisation allows us to (a) plug a lot of the
5011 thread-safety problems that existed, and (b) makes it possible to clean
5012 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5013 such data would previously have always leaked in application code and
5014 workarounds were in place to make the memory debugging turn a blind eye
5015 to it. Application code that doesn't use this new function will still
5016 leak as before, but their memory debugging output will announce it now
5017 rather than letting it slide.
5019 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5020 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5021 has a return value to indicate success or failure.
5024 *) Make it possible to replace the underlying "ERR" functions such that the
5025 global state (2 LHASH tables and 2 locks) is only used by the "default"
5026 implementation. This change also adds two functions to "get" and "set"
5027 the implementation prior to it being automatically set the first time
5028 any other ERR function takes place. Ie. an application can call "get",
5029 pass the return value to a module it has just loaded, and that module
5030 can call its own "set" function using that value. This means the
5031 module's "ERR" operations will use (and modify) the error state in the
5032 application and not in its own statically linked copy of OpenSSL code.
5035 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5036 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5037 the operation, and provides a more encapsulated way for external code
5038 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5039 to use these functions rather than manually incrementing the counts.
5041 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5044 *) Add EVP test program.
5047 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5050 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5051 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5052 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5053 These allow a CRL to be built without having to access X509_CRL fields
5054 directly. Modify 'ca' application to use new functions.
5057 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5058 bug workarounds. Rollback attack detection is a security feature.
5059 The problem will only arise on OpenSSL servers when TLSv1 is not
5060 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5061 Software authors not wanting to support TLSv1 will have special reasons
5062 for their choice and can explicitly enable this option.
5063 [Bodo Moeller, Lutz Jaenicke]
5065 *) Rationalise EVP so it can be extended: don't include a union of
5066 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5067 (similar to those existing for EVP_CIPHER_CTX).
5072 EVP_MD_CTX_init(&md); /* new function call */
5073 EVP_DigestInit(&md, EVP_sha1());
5074 EVP_DigestUpdate(&md, in, len);
5075 EVP_DigestFinal(&md, out, NULL);
5076 EVP_MD_CTX_cleanup(&md); /* new function call */
5080 *) Make DES key schedule conform to the usual scheme, as well as
5081 correcting its structure. This means that calls to DES functions
5082 now have to pass a pointer to a des_key_schedule instead of a
5083 plain des_key_schedule (which was actually always a pointer
5086 des_key_schedule ks;
5088 des_set_key_checked(..., &ks);
5089 des_ncbc_encrypt(..., &ks, ...);
5091 (Note that a later change renames 'des_...' into 'DES_...'.)
5094 *) Initial reduction of linker bloat: the use of some functions, such as
5095 PEM causes large amounts of unused functions to be linked in due to
5096 poor organisation. For example pem_all.c contains every PEM function
5097 which has a knock on effect of linking in large amounts of (unused)
5098 ASN1 code. Grouping together similar functions and splitting unrelated
5099 functions prevents this.
5102 *) Cleanup of EVP macros.
5105 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5106 correct _ecb suffix.
5109 *) Add initial OCSP responder support to ocsp application. The
5110 revocation information is handled using the text based index
5111 use by the ca application. The responder can either handle
5112 requests generated internally, supplied in files (for example
5113 via a CGI script) or using an internal minimal server.
5116 *) Add configuration choices to get zlib compression for TLS.
5119 *) Changes to Kerberos SSL for RFC 2712 compliance:
5120 1. Implemented real KerberosWrapper, instead of just using
5121 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5122 2. Implemented optional authenticator field of KerberosWrapper.
5124 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5125 and authenticator structs; see crypto/krb5/.
5127 Generalized Kerberos calls to support multiple Kerberos libraries.
5128 [Vern Staats <staatsvr@asc.hpc.mil>,
5129 Jeffrey Altman <jaltman@columbia.edu>
5130 via Richard Levitte]
5132 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5133 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5134 values for each of the key sizes rather than having just
5135 parameters (and 'speed' generating keys each time).
5138 *) Speed up EVP routines.
5141 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5142 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5143 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5144 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5146 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5147 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5148 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5151 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5153 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5156 *) Added the OS2-EMX target.
5157 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5159 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5160 to support NCONF routines in extension code. New function CONF_set_nconf()
5161 to allow functions which take an NCONF to also handle the old LHASH
5162 structure: this means that the old CONF compatible routines can be
5163 retained (in particular wrt extensions) without having to duplicate the
5164 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5167 *) Enhance the general user interface with mechanisms for inner control
5168 and with possibilities to have yes/no kind of prompts.
5171 *) Change all calls to low level digest routines in the library and
5172 applications to use EVP. Add missing calls to HMAC_cleanup() and
5173 don't assume HMAC_CTX can be copied using memcpy().
5174 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5176 *) Add the possibility to control engines through control names but with
5177 arbitrary arguments instead of just a string.
5178 Change the key loaders to take a UI_METHOD instead of a callback
5179 function pointer. NOTE: this breaks binary compatibility with earlier
5180 versions of OpenSSL [engine].
5181 Adapt the nCipher code for these new conditions and add a card insertion
5185 *) Enhance the general user interface with mechanisms to better support
5186 dialog box interfaces, application-defined prompts, the possibility
5187 to use defaults (for example default passwords from somewhere else)
5188 and interrupts/cancellations.
5191 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5192 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5195 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5196 tidy up some unnecessarily weird code in 'sk_new()').
5197 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5199 *) Change the key loading routines for ENGINEs to use the same kind
5200 callback (pem_password_cb) as all other routines that need this
5204 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5205 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5206 than this minimum value is recommended.
5209 *) New random seeder for OpenVMS, using the system process statistics
5210 that are easily reachable.
5213 *) Windows apparently can't transparently handle global
5214 variables defined in DLLs. Initialisations such as:
5216 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5218 wont compile. This is used by the any applications that need to
5219 declare their own ASN1 modules. This was fixed by adding the option
5220 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5221 needed for static libraries under Win32.
5224 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5225 setting of purpose and trust fields. New X509_STORE trust and
5226 purpose functions and tidy up setting in other SSL functions.
5229 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5230 structure. These are inherited by X509_STORE_CTX when it is
5231 initialised. This allows various defaults to be set in the
5232 X509_STORE structure (such as flags for CRL checking and custom
5233 purpose or trust settings) for functions which only use X509_STORE_CTX
5234 internally such as S/MIME.
5236 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5237 trust settings if they are not set in X509_STORE. This allows X509_STORE
5238 purposes and trust (in S/MIME for example) to override any set by default.
5240 Add command line options for CRL checking to smime, s_client and s_server
5244 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5245 are set then the CRL is looked up in the X509_STORE structure and
5246 its validity and signature checked, then if the certificate is found
5247 in the CRL the verify fails with a revoked error.
5249 Various new CRL related callbacks added to X509_STORE_CTX structure.
5251 Command line options added to 'verify' application to support this.
5253 This needs some additional work, such as being able to handle multiple
5254 CRLs with different times, extension based lookup (rather than just
5255 by subject name) and ultimately more complete V2 CRL extension
5259 *) Add a general user interface API (crypto/ui/). This is designed
5260 to replace things like des_read_password and friends (backward
5261 compatibility functions using this new API are provided).
5262 The purpose is to remove prompting functions from the DES code
5263 section as well as provide for prompting through dialog boxes in
5264 a window system and the like.
5267 *) Add "ex_data" support to ENGINE so implementations can add state at a
5268 per-structure level rather than having to store it globally.
5271 *) Make it possible for ENGINE structures to be copied when retrieved by
5272 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5273 This causes the "original" ENGINE structure to act like a template,
5274 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5275 operational state can be localised to each ENGINE structure, despite the
5276 fact they all share the same "methods". New ENGINE structures returned in
5277 this case have no functional references and the return value is the single
5278 structural reference. This matches the single structural reference returned
5279 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5283 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5284 needs to match any other type at all we need to manually clear the
5288 *) Changes to the "openssl engine" utility to include;
5289 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5290 about an ENGINE's available control commands.
5291 - executing control commands from command line arguments using the
5292 '-pre' and '-post' switches. '-post' is only used if '-t' is
5293 specified and the ENGINE is successfully initialised. The syntax for
5294 the individual commands are colon-separated, for example;
5295 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5298 *) New dynamic control command support for ENGINEs. ENGINEs can now
5299 declare their own commands (numbers), names (strings), descriptions,
5300 and input types for run-time discovery by calling applications. A
5301 subset of these commands are implicitly classed as "executable"
5302 depending on their input type, and only these can be invoked through
5303 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5304 can be based on user input, config files, etc). The distinction is
5305 that "executable" commands cannot return anything other than a boolean
5306 result and can only support numeric or string input, whereas some
5307 discoverable commands may only be for direct use through
5308 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5309 pointers, or other custom uses. The "executable" commands are to
5310 support parameterisations of ENGINE behaviour that can be
5311 unambiguously defined by ENGINEs and used consistently across any
5312 OpenSSL-based application. Commands have been added to all the
5313 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5314 control over shared-library paths without source code alterations.
5317 *) Changed all ENGINE implementations to dynamically allocate their
5318 ENGINEs rather than declaring them statically. Apart from this being
5319 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5320 this also allows the implementations to compile without using the
5321 internal engine_int.h header.
5324 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5325 'const' value. Any code that should be able to modify a RAND_METHOD
5326 should already have non-const pointers to it (ie. they should only
5327 modify their own ones).
5330 *) Made a variety of little tweaks to the ENGINE code.
5331 - "atalla" and "ubsec" string definitions were moved from header files
5332 to C code. "nuron" string definitions were placed in variables
5333 rather than hard-coded - allowing parameterisation of these values
5334 later on via ctrl() commands.
5335 - Removed unused "#if 0"'d code.
5336 - Fixed engine list iteration code so it uses ENGINE_free() to release
5337 structural references.
5338 - Constified the RAND_METHOD element of ENGINE structures.
5339 - Constified various get/set functions as appropriate and added
5340 missing functions (including a catch-all ENGINE_cpy that duplicates
5341 all ENGINE values onto a new ENGINE except reference counts/state).
5342 - Removed NULL parameter checks in get/set functions. Setting a method
5343 or function to NULL is a way of cancelling out a previously set
5344 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5345 and doesn't justify the extra error symbols and code.
5346 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5347 flags from engine_int.h to engine.h.
5348 - Changed prototypes for ENGINE handler functions (init(), finish(),
5349 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5352 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5353 to the algorithm using long division. The binary algorithm can be
5354 used only if the modulus is odd. On 32-bit systems, it is faster
5355 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5356 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5357 up to 450 bits. In 64-bit environments, the binary algorithm
5358 appears to be advantageous for much longer moduli; here we use it
5359 for moduli up to 2048 bits.
5362 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5363 could not support the combine flag in choice fields.
5366 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5367 extensions from a certificate request to the certificate.
5370 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5371 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5372 file: this allows the display of the certificate about to be
5373 signed to be customised, to allow certain fields to be included
5374 or excluded and extension details. The old system didn't display
5375 multicharacter strings properly, omitted fields not in the policy
5376 and couldn't display additional details such as extensions.
5379 *) Function EC_POINTs_mul for multiple scalar multiplication
5380 of an arbitrary number of elliptic curve points
5381 \sum scalars[i]*points[i],
5382 optionally including the generator defined for the EC_GROUP:
5383 scalar*generator + \sum scalars[i]*points[i].
5385 EC_POINT_mul is a simple wrapper function for the typical case
5386 that the point list has just one item (besides the optional
5390 *) First EC_METHODs for curves over GF(p):
5392 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5393 operations and provides various method functions that can also
5394 operate with faster implementations of modular arithmetic.
5396 EC_GFp_mont_method() reuses most functions that are part of
5397 EC_GFp_simple_method, but uses Montgomery arithmetic.
5399 [Bodo Moeller; point addition and point doubling
5400 implementation directly derived from source code provided by
5401 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5403 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5404 crypto/ec/ec_lib.c):
5406 Curves are EC_GROUP objects (with an optional group generator)
5407 based on EC_METHODs that are built into the library.
5409 Points are EC_POINT objects based on EC_GROUP objects.
5411 Most of the framework would be able to handle curves over arbitrary
5412 finite fields, but as there are no obvious types for fields other
5413 than GF(p), some functions are limited to that for now.
5416 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5417 that the file contains a complete HTTP response.
5420 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5421 change the def and num file printf format specifier from "%-40sXXX"
5422 to "%-39s XXX". The latter will always guarantee a space after the
5423 field while the former will cause them to run together if the field
5424 is 40 of more characters long.
5427 *) Constify the cipher and digest 'method' functions and structures
5428 and modify related functions to take constant EVP_MD and EVP_CIPHER
5432 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5433 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5436 *) Modify EVP_Digest*() routines so they now return values. Although the
5437 internal software routines can never fail additional hardware versions
5441 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5443 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5444 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5447 ERR_R_NESTED_ASN1_ERROR
5449 ERR_R_MISSING_ASN1_EOS
5450 were 4 .. 9, conflicting with
5451 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5453 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5454 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5456 Add new error code 'ERR_R_INTERNAL_ERROR'.
5459 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5463 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5464 sets the subject name for a new request or supersedes the
5465 subject name in a given request. Formats that can be parsed are
5466 'CN=Some Name, OU=myOU, C=IT'
5468 'CN=Some Name/OU=myOU/C=IT'.
5470 Add options '-batch' and '-verbose' to 'openssl req'.
5471 [Massimiliano Pala <madwolf@hackmasters.net>]
5473 *) Introduce the possibility to access global variables through
5474 functions on platform were that's the best way to handle exporting
5475 global variables in shared libraries. To enable this functionality,
5476 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5477 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5478 is normally done by Configure or something similar).
5480 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5481 in the source file (foo.c) like this:
5483 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5484 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5486 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5487 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5489 OPENSSL_DECLARE_GLOBAL(int,foo);
5490 #define foo OPENSSL_GLOBAL_REF(foo)
5491 OPENSSL_DECLARE_GLOBAL(double,bar);
5492 #define bar OPENSSL_GLOBAL_REF(bar)
5494 The #defines are very important, and therefore so is including the
5495 header file everywhere where the defined globals are used.
5497 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5498 of ASN.1 items, but that structure is a bit different.
5500 The largest change is in util/mkdef.pl which has been enhanced with
5501 better and easier to understand logic to choose which symbols should
5502 go into the Windows .def files as well as a number of fixes and code
5503 cleanup (among others, algorithm keywords are now sorted
5504 lexicographically to avoid constant rewrites).
5507 *) In BN_div() keep a copy of the sign of 'num' before writing the
5508 result to 'rm' because if rm==num the value will be overwritten
5509 and produce the wrong result if 'num' is negative: this caused
5510 problems with BN_mod() and BN_nnmod().
5513 *) Function OCSP_request_verify(). This checks the signature on an
5514 OCSP request and verifies the signer certificate. The signer
5515 certificate is just checked for a generic purpose and OCSP request
5519 *) Add OCSP_check_validity() function to check the validity of OCSP
5520 responses. OCSP responses are prepared in real time and may only
5521 be a few seconds old. Simply checking that the current time lies
5522 between thisUpdate and nextUpdate max reject otherwise valid responses
5523 caused by either OCSP responder or client clock inaccuracy. Instead
5524 we allow thisUpdate and nextUpdate to fall within a certain period of
5525 the current time. The age of the response can also optionally be
5526 checked. Two new options -validity_period and -status_age added to
5530 *) If signature or public key algorithm is unrecognized print out its
5531 OID rather that just UNKNOWN.
5534 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5535 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5536 ID to be generated from the issuer certificate alone which can then be
5537 passed to OCSP_id_issuer_cmp().
5540 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5541 ASN1 modules to export functions returning ASN1_ITEM pointers
5542 instead of the ASN1_ITEM structures themselves. This adds several
5543 new macros which allow the underlying ASN1 function/structure to
5544 be accessed transparently. As a result code should not use ASN1_ITEM
5545 references directly (such as &X509_it) but instead use the relevant
5546 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5547 use of the new ASN1 code on platforms where exporting structures
5548 is problematical (for example in shared libraries) but exporting
5549 functions returning pointers to structures is not.
5552 *) Add support for overriding the generation of SSL/TLS session IDs.
5553 These callbacks can be registered either in an SSL_CTX or per SSL.
5554 The purpose of this is to allow applications to control, if they wish,
5555 the arbitrary values chosen for use as session IDs, particularly as it
5556 can be useful for session caching in multiple-server environments. A
5557 command-line switch for testing this (and any client code that wishes
5558 to use such a feature) has been added to "s_server".
5559 [Geoff Thorpe, Lutz Jaenicke]
5561 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5562 of the form '#if defined(...) || defined(...) || ...' and
5563 '#if !defined(...) && !defined(...) && ...'. This also avoids
5564 the growing number of special cases it was previously handling.
5567 *) Make all configuration macros available for application by making
5568 sure they are available in opensslconf.h, by giving them names starting
5569 with "OPENSSL_" to avoid conflicts with other packages and by making
5570 sure e_os2.h will cover all platform-specific cases together with
5572 Additionally, it is now possible to define configuration/platform-
5573 specific names (called "system identities"). In the C code, these
5574 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5575 macro with the name beginning with "OPENSSL_SYS_", which is determined
5576 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5580 *) New option -set_serial to 'req' and 'x509' this allows the serial
5581 number to use to be specified on the command line. Previously self
5582 signed certificates were hard coded with serial number 0 and the
5583 CA options of 'x509' had to use a serial number in a file which was
5587 *) New options to 'ca' utility to support V2 CRL entry extensions.
5588 Currently CRL reason, invalidity date and hold instruction are
5589 supported. Add new CRL extensions to V3 code and some new objects.
5592 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5593 disable standard block padding (aka PKCS#5 padding) in the EVP
5594 API, which was previously mandatory. This means that the data is
5595 not padded in any way and so the total length much be a multiple
5596 of the block size, otherwise an error occurs.
5599 *) Initial (incomplete) OCSP SSL support.
5602 *) New function OCSP_parse_url(). This splits up a URL into its host,
5603 port and path components: primarily to parse OCSP URLs. New -url
5604 option to ocsp utility.
5607 *) New nonce behavior. The return value of OCSP_check_nonce() now
5608 reflects the various checks performed. Applications can decide
5609 whether to tolerate certain situations such as an absent nonce
5610 in a response when one was present in a request: the ocsp application
5611 just prints out a warning. New function OCSP_add1_basic_nonce()
5612 this is to allow responders to include a nonce in a response even if
5613 the request is nonce-less.
5616 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5617 skipped when using openssl x509 multiple times on a single input file,
5618 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5621 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5622 set string type: to handle setting ASN1_TIME structures. Fix ca
5623 utility to correctly initialize revocation date of CRLs.
5626 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5627 the clients preferred ciphersuites and rather use its own preferences.
5628 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5629 Internet Explorer by ensuring unchanged hash method during stepup.
5630 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5633 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5634 to aes and add a new 'exist' option to print out symbols that don't
5638 *) Additional options to ocsp utility to allow flags to be set and
5639 additional certificates supplied.
5642 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5643 OCSP client a number of certificate to only verify the response
5647 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5648 handle the new API. Currently only ECB, CBC modes supported. Add new
5651 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5652 Encryption Standard (AES) Ciphersuites for Transport Layer
5653 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5654 not enabled by default and were not part of the "ALL" ciphersuite
5655 alias because they were not yet official; they could be
5656 explicitly requested by specifying the "AESdraft" ciphersuite
5657 group alias. In the final release of OpenSSL 0.9.7, the group
5658 alias is called "AES" and is part of "ALL".)
5659 [Ben Laurie, Steve Henson, Bodo Moeller]
5661 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5662 request to response.
5665 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5666 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5667 extract information from a certificate request. OCSP_response_create()
5668 creates a response and optionally adds a basic response structure.
5669 OCSP_basic_add1_status() adds a complete single response to a basic
5670 response and returns the OCSP_SINGLERESP structure just added (to allow
5671 extensions to be included for example). OCSP_basic_add1_cert() adds a
5672 certificate to a basic response and OCSP_basic_sign() signs a basic
5673 response with various flags. New helper functions ASN1_TIME_check()
5674 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5675 (converts ASN1_TIME to GeneralizedTime).
5678 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5679 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5680 structure from a certificate. X509_pubkey_digest() digests the public_key
5681 contents: this is used in various key identifiers.
5684 *) Make sk_sort() tolerate a NULL argument.
5685 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5687 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5688 passed by the function are trusted implicitly. If any of them signed the
5689 response then it is assumed to be valid and is not verified.
5692 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5693 to data. This was previously part of the PKCS7 ASN1 code. This
5694 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5695 [Steve Henson, reported by Kenneth R. Robinette
5696 <support@securenetterm.com>]
5698 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5699 routines: without these tracing memory leaks is very painful.
5700 Fix leaks in PKCS12 and PKCS7 routines.
5703 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5704 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5705 effectively meant GeneralizedTime would never be used. Now it
5706 is initialised to -1 but X509_time_adj() now has to check the value
5707 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5708 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5709 [Steve Henson, reported by Kenneth R. Robinette
5710 <support@securenetterm.com>]
5712 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5713 result in a zero length in the ASN1_INTEGER structure which was
5714 not consistent with the structure when d2i_ASN1_INTEGER() was used
5715 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5716 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5717 where it did not print out a minus for negative ASN1_INTEGER.
5720 *) Add summary printout to ocsp utility. The various functions which
5721 convert status values to strings have been renamed to:
5722 OCSP_response_status_str(), OCSP_cert_status_str() and
5723 OCSP_crl_reason_str() and are no longer static. New options
5724 to verify nonce values and to disable verification. OCSP response
5725 printout format cleaned up.
5728 *) Add additional OCSP certificate checks. These are those specified
5729 in RFC2560. This consists of two separate checks: the CA of the
5730 certificate being checked must either be the OCSP signer certificate
5731 or the issuer of the OCSP signer certificate. In the latter case the
5732 OCSP signer certificate must contain the OCSP signing extended key
5733 usage. This check is performed by attempting to match the OCSP
5734 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5735 in the OCSP_CERTID structures of the response.
5738 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5739 and related routines. This uses the standard OpenSSL certificate
5740 verify routines to perform initial checks (just CA validity) and
5741 to obtain the certificate chain. Then additional checks will be
5742 performed on the chain. Currently the root CA is checked to see
5743 if it is explicitly trusted for OCSP signing. This is used to set
5744 a root CA as a global signing root: that is any certificate that
5745 chains to that CA is an acceptable OCSP signing certificate.
5748 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5749 extensions from a separate configuration file.
5750 As when reading extensions from the main configuration file,
5751 the '-extensions ...' option may be used for specifying the
5753 [Massimiliano Pala <madwolf@comune.modena.it>]
5755 *) New OCSP utility. Allows OCSP requests to be generated or
5756 read. The request can be sent to a responder and the output
5757 parsed, outputed or printed in text form. Not complete yet:
5758 still needs to check the OCSP response validity.
5761 *) New subcommands for 'openssl ca':
5762 'openssl ca -status <serial>' prints the status of the cert with
5763 the given serial number (according to the index file).
5764 'openssl ca -updatedb' updates the expiry status of certificates
5766 [Massimiliano Pala <madwolf@comune.modena.it>]
5768 *) New '-newreq-nodes' command option to CA.pl. This is like
5769 '-newreq', but calls 'openssl req' with the '-nodes' option
5770 so that the resulting key is not encrypted.
5771 [Damien Miller <djm@mindrot.org>]
5773 *) New configuration for the GNU Hurd.
5774 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5776 *) Initial code to implement OCSP basic response verify. This
5777 is currently incomplete. Currently just finds the signer's
5778 certificate and verifies the signature on the response.
5781 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5782 value of OPENSSLDIR. This is available via the new '-d' option
5783 to 'openssl version', and is also included in 'openssl version -a'.
5786 *) Allowing defining memory allocation callbacks that will be given
5787 file name and line number information in additional arguments
5788 (a const char* and an int). The basic functionality remains, as
5789 well as the original possibility to just replace malloc(),
5790 realloc() and free() by functions that do not know about these
5791 additional arguments. To register and find out the current
5792 settings for extended allocation functions, the following
5793 functions are provided:
5795 CRYPTO_set_mem_ex_functions
5796 CRYPTO_set_locked_mem_ex_functions
5797 CRYPTO_get_mem_ex_functions
5798 CRYPTO_get_locked_mem_ex_functions
5800 These work the same way as CRYPTO_set_mem_functions and friends.
5801 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5802 extended allocation function is enabled.
5803 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5804 a conventional allocation function is enabled.
5805 [Richard Levitte, Bodo Moeller]
5807 *) Finish off removing the remaining LHASH function pointer casts.
5808 There should no longer be any prototype-casting required when using
5809 the LHASH abstraction, and any casts that remain are "bugs". See
5810 the callback types and macros at the head of lhash.h for details
5811 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5814 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5815 If /dev/[u]random devices are not available or do not return enough
5816 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5818 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5819 /etc/entropy will be queried once each in this sequence, quering stops
5820 when enough entropy was collected without querying more sockets.
5823 *) Change the Unix RAND_poll() variant to be able to poll several
5824 random devices, as specified by DEVRANDOM, until a sufficient amount
5825 of data has been collected. We spend at most 10 ms on each file
5826 (select timeout) and read in non-blocking mode. DEVRANDOM now
5827 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5828 (previously it was just the string "/dev/urandom"), so on typical
5829 platforms the 10 ms delay will never occur.
5830 Also separate out the Unix variant to its own file, rand_unix.c.
5831 For VMS, there's a currently-empty rand_vms.c.
5834 *) Move OCSP client related routines to ocsp_cl.c. These
5835 provide utility functions which an application needing
5836 to issue a request to an OCSP responder and analyse the
5837 response will typically need: as opposed to those which an
5838 OCSP responder itself would need which will be added later.
5840 OCSP_request_sign() signs an OCSP request with an API similar
5841 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5842 response. OCSP_response_get1_basic() extracts basic response
5843 from response. OCSP_resp_find_status(): finds and extracts status
5844 information from an OCSP_CERTID structure (which will be created
5845 when the request structure is built). These are built from lower
5846 level functions which work on OCSP_SINGLERESP structures but
5847 wont normally be used unless the application wishes to examine
5848 extensions in the OCSP response for example.
5850 Replace nonce routines with a pair of functions.
5851 OCSP_request_add1_nonce() adds a nonce value and optionally
5852 generates a random value. OCSP_check_nonce() checks the
5853 validity of the nonce in an OCSP response.
5856 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5857 This doesn't copy the supplied OCSP_CERTID and avoids the
5858 need to free up the newly created id. Change return type
5859 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5860 This can then be used to add extensions to the request.
5861 Deleted OCSP_request_new(), since most of its functionality
5862 is now in OCSP_REQUEST_new() (and the case insensitive name
5863 clash) apart from the ability to set the request name which
5864 will be added elsewhere.
5867 *) Update OCSP API. Remove obsolete extensions argument from
5868 various functions. Extensions are now handled using the new
5869 OCSP extension code. New simple OCSP HTTP function which
5870 can be used to send requests and parse the response.
5873 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5874 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5875 uses the special reorder version of SET OF to sort the attributes
5876 and reorder them to match the encoded order. This resolves a long
5877 standing problem: a verify on a PKCS7 structure just after signing
5878 it used to fail because the attribute order did not match the
5879 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5880 it uses the received order. This is necessary to tolerate some broken
5881 software that does not order SET OF. This is handled by encoding
5882 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5883 to produce the required SET OF.
5886 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5887 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5888 files to get correct declarations of the ASN.1 item variables.
5891 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5892 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5893 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5894 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5895 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5896 ASN1_ITEM and no wrapper functions.
5899 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5900 replace the old function pointer based I/O routines. Change most of
5901 the *_d2i_bio() and *_d2i_fp() functions to use these.
5904 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5905 lines, recognice more "algorithms" that can be deselected, and make
5906 it complain about algorithm deselection that isn't recognised.
5909 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5910 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5911 to use new functions. Add NO_ASN1_OLD which can be set to remove
5912 some old style ASN1 functions: this can be used to determine if old
5913 code will still work when these eventually go away.
5916 *) New extension functions for OCSP structures, these follow the
5917 same conventions as certificates and CRLs.
5920 *) New function X509V3_add1_i2d(). This automatically encodes and
5921 adds an extension. Its behaviour can be customised with various
5922 flags to append, replace or delete. Various wrappers added for
5923 certifcates and CRLs.
5926 *) Fix to avoid calling the underlying ASN1 print routine when
5927 an extension cannot be parsed. Correct a typo in the
5928 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5931 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5932 entries for variables.
5935 *) Add functionality to apps/openssl.c for detecting locking
5936 problems: As the program is single-threaded, all we have
5937 to do is register a locking callback using an array for
5938 storing which locks are currently held by the program.
5941 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5942 SSL_get_ex_data_X509_STORE_idx(), which is used in
5943 ssl_verify_cert_chain() and thus can be called at any time
5944 during TLS/SSL handshakes so that thread-safety is essential.
5945 Unfortunately, the ex_data design is not at all suited
5946 for multi-threaded use, so it probably should be abolished.
5949 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5950 [Broadcom, tweaked and integrated by Geoff Thorpe]
5952 *) Move common extension printing code to new function
5953 X509V3_print_extensions(). Reorganise OCSP print routines and
5954 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5957 *) New function X509_signature_print() to remove duplication in some
5961 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5962 set (this was treated exactly the same as SET OF previously). This
5963 is used to reorder the STACK representing the structure to match the
5964 encoding. This will be used to get round a problem where a PKCS7
5965 structure which was signed could not be verified because the STACK
5966 order did not reflect the encoded order.
5969 *) Reimplement the OCSP ASN1 module using the new code.
5972 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5973 for its ASN1 operations. The old style function pointers still exist
5974 for now but they will eventually go away.
5977 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5978 completely replaces the old ASN1 functionality with a table driven
5979 encoder and decoder which interprets an ASN1_ITEM structure describing
5980 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5981 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5982 has also been converted to the new form.
5985 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5986 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5987 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5988 for negative moduli.
5991 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5992 of not touching the result's sign bit.
5995 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5999 *) Changed the LHASH code to use prototypes for callbacks, and created
6000 macros to declare and implement thin (optionally static) functions
6001 that provide type-safety and avoid function pointer casting for the
6002 type-specific callbacks.
6005 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6007 [Veers Staats <staatsvr@asc.hpc.mil>,
6008 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6010 *) Reformat the FAQ so the different questions and answers can be divided
6011 in sections depending on the subject.
6014 *) Have the zlib compression code load ZLIB.DLL dynamically under
6018 *) New function BN_mod_sqrt for computing square roots modulo a prime
6019 (using the probabilistic Tonelli-Shanks algorithm unless
6020 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6021 be handled deterministically).
6022 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6024 *) Make BN_mod_inverse faster by explicitly handling small quotients
6025 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6026 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6029 *) New function BN_kronecker.
6032 *) Fix BN_gcd so that it works on negative inputs; the result is
6033 positive unless both parameters are zero.
6034 Previously something reasonably close to an infinite loop was
6035 possible because numbers could be growing instead of shrinking
6036 in the implementation of Euclid's algorithm.
6039 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6040 sign of the number in question.
6042 Fix BN_is_word(a,w) to work correctly for w == 0.
6044 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6045 because its test if the absolute value of 'a' equals 'w'.
6046 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6047 it exists mostly for use in the implementations of BN_is_zero(),
6048 BN_is_one(), and BN_is_word().
6051 *) New function BN_swap.
6054 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6055 the exponentiation functions are more likely to produce reasonable
6056 results on negative inputs.
6059 *) Change BN_mod_mul so that the result is always non-negative.
6060 Previously, it could be negative if one of the factors was negative;
6061 I don't think anyone really wanted that behaviour.
6064 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6065 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6066 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6067 and add new functions:
6076 BN_mod_lshift1_quick
6080 These functions always generate non-negative results.
6082 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6083 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6085 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6086 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6087 be reduced modulo m.
6088 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6091 The following entry accidentily appeared in the CHANGES file
6092 distributed with OpenSSL 0.9.7. The modifications described in
6093 it do *not* apply to OpenSSL 0.9.7.
6095 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6096 was actually never needed) and in BN_mul(). The removal in BN_mul()
6097 required a small change in bn_mul_part_recursive() and the addition
6098 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6099 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6100 bn_sub_words() and bn_add_words() except they take arrays with
6105 *) In 'openssl passwd', verify passwords read from the terminal
6106 unless the '-salt' option is used (which usually means that
6107 verification would just waste user's time since the resulting
6108 hash is going to be compared with some given password hash)
6109 or the new '-noverify' option is used.
6111 This is an incompatible change, but it does not affect
6112 non-interactive use of 'openssl passwd' (passwords on the command
6113 line, '-stdin' option, '-in ...' option) and thus should not
6117 *) Remove all references to RSAref, since there's no more need for it.
6120 *) Make DSO load along a path given through an environment variable
6121 (SHLIB_PATH) with shl_load().
6124 *) Constify the ENGINE code as a result of BIGNUM constification.
6125 Also constify the RSA code and most things related to it. In a
6126 few places, most notable in the depth of the ASN.1 code, ugly
6127 casts back to non-const were required (to be solved at a later
6131 *) Make it so the openssl application has all engines loaded by default.
6134 *) Constify the BIGNUM routines a little more.
6137 *) Add the following functions:
6139 ENGINE_load_cswift()
6141 ENGINE_load_atalla()
6143 ENGINE_load_builtin_engines()
6145 That way, an application can itself choose if external engines that
6146 are built-in in OpenSSL shall ever be used or not. The benefit is
6147 that applications won't have to be linked with libdl or other dso
6148 libraries unless it's really needed.
6150 Changed 'openssl engine' to load all engines on demand.
6151 Changed the engine header files to avoid the duplication of some
6152 declarations (they differed!).
6155 *) 'openssl engine' can now list capabilities.
6158 *) Better error reporting in 'openssl engine'.
6161 *) Never call load_dh_param(NULL) in s_server.
6164 *) Add engine application. It can currently list engines by name and
6165 identity, and test if they are actually available.
6168 *) Improve RPM specification file by forcing symbolic linking and making
6169 sure the installed documentation is also owned by root.root.
6170 [Damien Miller <djm@mindrot.org>]
6172 *) Give the OpenSSL applications more possibilities to make use of
6173 keys (public as well as private) handled by engines.
6176 *) Add OCSP code that comes from CertCo.
6179 *) Add VMS support for the Rijndael code.
6182 *) Added untested support for Nuron crypto accelerator.
6185 *) Add support for external cryptographic devices. This code was
6186 previously distributed separately as the "engine" branch.
6187 [Geoff Thorpe, Richard Levitte]
6189 *) Rework the filename-translation in the DSO code. It is now possible to
6190 have far greater control over how a "name" is turned into a filename
6191 depending on the operating environment and any oddities about the
6192 different shared library filenames on each system.
6195 *) Support threads on FreeBSD-elf in Configure.
6198 *) Fix for SHA1 assembly problem with MASM: it produces
6199 warnings about corrupt line number information when assembling
6200 with debugging information. This is caused by the overlapping
6202 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6205 NCONF_get_number() has no error checking at all. As a replacement,
6206 NCONF_get_number_e() is defined (_e for "error checking") and is
6207 promoted strongly. The old NCONF_get_number is kept around for
6208 binary backward compatibility.
6209 Make it possible for methods to load from something other than a BIO,
6210 by providing a function pointer that is given a name instead of a BIO.
6211 For example, this could be used to load configuration data from an
6215 *) Fix for non blocking accept BIOs. Added new I/O special reason
6216 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6217 with non blocking I/O was not possible because no retry code was
6218 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6222 *) Added the beginnings of Rijndael support.
6225 *) Fix for bug in DirectoryString mask setting. Add support for
6226 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6227 to allow certificate printing to more controllable, additional
6228 'certopt' option to 'x509' to allow new printing options to be
6232 *) Clean old EAY MD5 hack from e_os.h.
6235 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6237 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6238 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6239 [Joe Orton, Steve Henson]
6241 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6243 *) Fix additional bug revealed by the NISCC test suite:
6245 Stop bug triggering large recursion when presented with
6246 certain ASN.1 tags (CVE-2003-0851)
6249 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6251 *) Fix various bugs revealed by running the NISCC test suite:
6253 Stop out of bounds reads in the ASN1 code when presented with
6254 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6256 If verify callback ignores invalid public key errors don't try to check
6257 certificate signature with the NULL public key.
6261 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6262 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6266 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6267 extra data after the compression methods not only for TLS 1.0
6268 but also for SSL 3.0 (as required by the specification).
6269 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6271 *) Change X509_certificate_type() to mark the key as exported/exportable
6272 when it's 512 *bits* long, not 512 bytes.
6275 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6277 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6278 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6279 a protocol version number mismatch like a decryption error
6280 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6283 *) Turn on RSA blinding by default in the default implementation
6284 to avoid a timing attack. Applications that don't want it can call
6285 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6286 They would be ill-advised to do so in most cases.
6287 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6289 *) Change RSA blinding code so that it works when the PRNG is not
6290 seeded (in this case, the secret RSA exponent is abused as
6291 an unpredictable seed -- if it is not unpredictable, there
6292 is no point in blinding anyway). Make RSA blinding thread-safe
6293 by remembering the creator's thread ID in rsa->blinding and
6294 having all other threads use local one-time blinding factors
6295 (this requires more computation than sharing rsa->blinding, but
6296 avoids excessive locking; and if an RSA object is not shared
6297 between threads, blinding will still be very fast).
6300 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6302 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6303 via timing by performing a MAC computation even if incorrrect
6304 block cipher padding has been found. This is a countermeasure
6305 against active attacks where the attacker has to distinguish
6306 between bad padding and a MAC verification error. (CVE-2003-0078)
6308 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6309 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6310 Martin Vuagnoux (EPFL, Ilion)]
6312 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6314 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6315 memory from it's contents. This is done with a counter that will
6316 place alternating values in each byte. This can be used to solve
6317 two issues: 1) the removal of calls to memset() by highly optimizing
6318 compilers, and 2) cleansing with other values than 0, since those can
6319 be read through on certain media, for example a swap space on disk.
6322 *) Bugfix: client side session caching did not work with external caching,
6323 because the session->cipher setting was not restored when reloading
6324 from the external cache. This problem was masked, when
6325 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6326 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6329 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6330 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6331 [Zeev Lieber <zeev-l@yahoo.com>]
6333 *) Undo an undocumented change introduced in 0.9.6e which caused
6334 repeated calls to OpenSSL_add_all_ciphers() and
6335 OpenSSL_add_all_digests() to be ignored, even after calling
6339 *) Change the default configuration reader to deal with last line not
6340 being properly terminated.
6343 *) Change X509_NAME_cmp() so it applies the special rules on handling
6344 DN values that are of type PrintableString, as well as RDNs of type
6345 emailAddress where the value has the type ia5String.
6346 [stefank@valicert.com via Richard Levitte]
6348 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6349 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6350 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6351 the bitwise-OR of the two for use by the majority of applications
6352 wanting this behaviour, and update the docs. The documented
6353 behaviour and actual behaviour were inconsistent and had been
6354 changing anyway, so this is more a bug-fix than a behavioural
6356 [Geoff Thorpe, diagnosed by Nadav Har'El]
6358 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6359 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6362 *) Fix initialization code race conditions in
6363 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6364 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6365 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6366 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6367 ssl2_get_cipher_by_char(),
6368 ssl3_get_cipher_by_char().
6369 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6371 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6372 the cached sessions are flushed, as the remove_cb() might use ex_data
6373 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6374 (see [openssl.org #212]).
6375 [Geoff Thorpe, Lutz Jaenicke]
6377 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6378 length, instead of the encoding length to d2i_ASN1_OBJECT.
6381 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6383 *) [In 0.9.6g-engine release:]
6384 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6385 [Lynn Gazis <lgazis@rainbow.com>]
6387 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6389 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6390 and get fix the header length calculation.
6391 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6392 Alon Kantor <alonk@checkpoint.com> (and others),
6395 *) Use proper error handling instead of 'assertions' in buffer
6396 overflow checks added in 0.9.6e. This prevents DoS (the
6397 assertions could call abort()).
6398 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6400 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6402 *) Add various sanity checks to asn1_get_length() to reject
6403 the ASN1 length bytes if they exceed sizeof(long), will appear
6404 negative or the content length exceeds the length of the
6406 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6408 *) Fix cipher selection routines: ciphers without encryption had no flags
6409 for the cipher strength set and where therefore not handled correctly
6410 by the selection routines (PR #130).
6413 *) Fix EVP_dsa_sha macro.
6417 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6418 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6419 that was added in OpenSSL 0.9.6d.
6421 As the countermeasure turned out to be incompatible with some
6422 broken SSL implementations, the new option is part of SSL_OP_ALL.
6423 SSL_OP_ALL is usually employed when compatibility with weird SSL
6424 implementations is desired (e.g. '-bugs' option to 's_client' and
6425 's_server'), so the new option is automatically set in many
6429 *) Changes in security patch:
6431 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6432 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6433 Air Force Materiel Command, USAF, under agreement number
6436 *) Add various sanity checks to asn1_get_length() to reject
6437 the ASN1 length bytes if they exceed sizeof(long), will appear
6438 negative or the content length exceeds the length of the
6439 supplied buffer. (CVE-2002-0659)
6440 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6442 *) Assertions for various potential buffer overflows, not known to
6444 [Ben Laurie (CHATS)]
6446 *) Various temporary buffers to hold ASCII versions of integers were
6447 too small for 64 bit platforms. (CVE-2002-0655)
6448 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6450 *) Remote buffer overflow in SSL3 protocol - an attacker could
6451 supply an oversized session ID to a client. (CVE-2002-0656)
6452 [Ben Laurie (CHATS)]
6454 *) Remote buffer overflow in SSL2 protocol - an attacker could
6455 supply an oversized client master key. (CVE-2002-0656)
6456 [Ben Laurie (CHATS)]
6458 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6460 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6461 encoded as NULL) with id-dsa-with-sha1.
6462 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6464 *) Check various X509_...() return values in apps/req.c.
6465 [Nils Larsch <nla@trustcenter.de>]
6467 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6468 an end-of-file condition would erronously be flagged, when the CRLF
6469 was just at the end of a processed block. The bug was discovered when
6470 processing data through a buffering memory BIO handing the data to a
6471 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6472 <ptsekov@syntrex.com> and Nedelcho Stanev.
6475 *) Implement a countermeasure against a vulnerability recently found
6476 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6477 before application data chunks to avoid the use of known IVs
6478 with data potentially chosen by the attacker.
6481 *) Fix length checks in ssl3_get_client_hello().
6484 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6485 to prevent ssl3_read_internal() from incorrectly assuming that
6486 ssl3_read_bytes() found application data while handshake
6487 processing was enabled when in fact s->s3->in_read_app_data was
6488 merely automatically cleared during the initial handshake.
6489 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6491 *) Fix object definitions for Private and Enterprise: they were not
6492 recognized in their shortname (=lowercase) representation. Extend
6493 obj_dat.pl to issue an error when using undefined keywords instead
6494 of silently ignoring the problem (Svenning Sorensen
6495 <sss@sss.dnsalias.net>).
6498 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6499 generators, i.e. generators other than 2 and 5. (Previously, the
6500 code did not properly initialise the 'add' and 'rem' values to
6501 BN_generate_prime().)
6503 In the new general case, we do not insist that 'generator' is
6504 actually a primitive root: This requirement is rather pointless;
6505 a generator of the order-q subgroup is just as good, if not
6509 *) Map new X509 verification errors to alerts. Discovered and submitted by
6510 Tom Wu <tom@arcot.com>.
6513 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6514 returning non-zero before the data has been completely received
6515 when using non-blocking I/O.
6516 [Bodo Moeller; problem pointed out by John Hughes]
6518 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6519 [Ben Laurie, Lutz Jaenicke]
6521 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6522 Yoram Zahavi <YoramZ@gilian.com>).
6525 *) Add information about CygWin 1.3 and on, and preserve proper
6526 configuration for the versions before that.
6527 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6529 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6530 check whether we deal with a copy of a session and do not delete from
6531 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6532 <izhar@checkpoint.com>.
6535 *) Do not store session data into the internal session cache, if it
6536 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6537 flag is set). Proposed by Aslam <aslam@funk.com>.
6540 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6544 *) [In 0.9.6d-engine release:]
6545 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6546 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6548 *) Add the configuration target linux-s390x.
6549 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6551 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6552 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6553 variable as an indication that a ClientHello message has been
6554 received. As the flag value will be lost between multiple
6555 invocations of ssl3_accept when using non-blocking I/O, the
6556 function may not be aware that a handshake has actually taken
6557 place, thus preventing a new session from being added to the
6560 To avoid this problem, we now set s->new_session to 2 instead of
6561 using a local variable.
6562 [Lutz Jaenicke, Bodo Moeller]
6564 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6565 if the SSL_R_LENGTH_MISMATCH error is detected.
6566 [Geoff Thorpe, Bodo Moeller]
6568 *) New 'shared_ldflag' column in Configure platform table.
6571 *) Fix EVP_CIPHER_mode macro.
6572 ["Dan S. Camper" <dan@bti.net>]
6574 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6575 type, we must throw them away by setting rr->length to 0.
6576 [D P Chang <dpc@qualys.com>]
6578 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6580 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6581 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6582 worked incorrectly for those cases where range = 10..._2 and
6583 3*range is two bits longer than range.)
6586 *) Only add signing time to PKCS7 structures if it is not already
6590 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6591 OBJ_ld_ce should be OBJ_id_ce.
6592 Also some ip-pda OIDs in crypto/objects/objects.txt were
6593 incorrect (cf. RFC 3039).
6594 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6596 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6597 returns early because it has nothing to do.
6598 [Andy Schneider <andy.schneider@bjss.co.uk>]
6600 *) [In 0.9.6c-engine release:]
6601 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6602 [Andy Schneider <andy.schneider@bjss.co.uk>]
6604 *) [In 0.9.6c-engine release:]
6605 Add support for Cryptographic Appliance's keyserver technology.
6606 (Use engine 'keyclient')
6607 [Cryptographic Appliances and Geoff Thorpe]
6609 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6610 is called via tools/c89.sh because arguments have to be
6611 rearranged (all '-L' options must appear before the first object
6613 [Richard Shapiro <rshapiro@abinitio.com>]
6615 *) [In 0.9.6c-engine release:]
6616 Add support for Broadcom crypto accelerator cards, backported
6618 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6620 *) [In 0.9.6c-engine release:]
6621 Add support for SureWare crypto accelerator cards from
6622 Baltimore Technologies. (Use engine 'sureware')
6623 [Baltimore Technologies and Mark Cox]
6625 *) [In 0.9.6c-engine release:]
6626 Add support for crypto accelerator cards from Accelerated
6627 Encryption Processing, www.aep.ie. (Use engine 'aep')
6628 [AEP Inc. and Mark Cox]
6630 *) Add a configuration entry for gcc on UnixWare.
6631 [Gary Benson <gbenson@redhat.com>]
6633 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6634 messages are stored in a single piece (fixed-length part and
6635 variable-length part combined) and fix various bugs found on the way.
6638 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6639 instead. BIO_gethostbyname() does not know what timeouts are
6640 appropriate, so entries would stay in cache even when they have
6642 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6644 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6645 faced with a pathologically small ClientHello fragment that does
6646 not contain client_version: Instead of aborting with an error,
6647 simply choose the highest available protocol version (i.e.,
6648 TLS 1.0 unless it is disabled). In practice, ClientHello
6649 messages are never sent like this, but this change gives us
6650 strictly correct behaviour at least for TLS.
6653 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6654 never resets s->method to s->ctx->method when called from within
6655 one of the SSL handshake functions.
6656 [Bodo Moeller; problem pointed out by Niko Baric]
6658 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6659 (sent using the client's version number) if client_version is
6660 smaller than the protocol version in use. Also change
6661 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6662 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6663 the client will at least see that alert.
6666 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6670 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6671 client receives HelloRequest while in a handshake.
6672 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6674 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6675 should end in 'break', not 'goto end' which circuments various
6676 cleanups done in state SSL_ST_OK. But session related stuff
6677 must be disabled for SSL_ST_OK in the case that we just sent a
6680 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6681 before just sending a HelloRequest.
6682 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6684 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6685 reveal whether illegal block cipher padding was found or a MAC
6686 verification error occured. (Neither SSLerr() codes nor alerts
6687 are directly visible to potential attackers, but the information
6688 may leak via logfiles.)
6690 Similar changes are not required for the SSL 2.0 implementation
6691 because the number of padding bytes is sent in clear for SSL 2.0,
6692 and the extra bytes are just ignored. However ssl/s2_pkt.c
6693 failed to verify that the purported number of padding bytes is in
6697 *) Add OpenUNIX-8 support including shared libraries
6698 (Boyd Lynn Gerber <gerberb@zenez.com>).
6701 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6702 'wristwatch attack' using huge encoding parameters (cf.
6703 James H. Manger's CRYPTO 2001 paper). Note that the
6704 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6705 encoding parameters and hence was not vulnerable.
6708 *) BN_sqr() bug fix.
6709 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6711 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6712 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6713 followed by modular reduction.
6714 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6716 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6717 equivalent based on BN_pseudo_rand() instead of BN_rand().
6720 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6721 This function was broken, as the check for a new client hello message
6722 to handle SGC did not allow these large messages.
6723 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6726 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6729 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6730 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6733 *) Rework the configuration and shared library support for Tru64 Unix.
6734 The configuration part makes use of modern compiler features and
6735 still retains old compiler behavior for those that run older versions
6736 of the OS. The shared library support part includes a variant that
6737 uses the RPATH feature, and is available through the special
6738 configuration target "alpha-cc-rpath", which will never be selected
6740 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6742 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6743 with the same message size as in ssl3_get_certificate_request().
6744 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6745 messages might inadvertently be reject as too long.
6746 [Petr Lampa <lampa@fee.vutbr.cz>]
6748 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6751 *) Modified SSL library such that the verify_callback that has been set
6752 specificly for an SSL object with SSL_set_verify() is actually being
6753 used. Before the change, a verify_callback set with this function was
6754 ignored and the verify_callback() set in the SSL_CTX at the time of
6755 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6756 to allow the necessary settings.
6759 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6760 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6761 done automatically (in contradiction to the requirements of the C
6762 standard). This made problems when used from OpenSSH.
6765 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6766 dh->length and always used
6768 BN_rand_range(priv_key, dh->p).
6770 BN_rand_range() is not necessary for Diffie-Hellman, and this
6771 specific range makes Diffie-Hellman unnecessarily inefficient if
6772 dh->length (recommended exponent length) is much smaller than the
6773 length of dh->p. We could use BN_rand_range() if the order of
6774 the subgroup was stored in the DH structure, but we only have
6779 BN_rand(priv_key, l, ...)
6781 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6787 RSA_eay_public_encrypt
6788 RSA_eay_private_decrypt
6789 RSA_eay_private_encrypt (signing)
6790 RSA_eay_public_decrypt (signature verification)
6792 (default implementations for RSA_public_encrypt,
6793 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6794 always reject numbers >= n.
6797 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6798 to synchronize access to 'locking_thread'. This is necessary on
6799 systems where access to 'locking_thread' (an 'unsigned long'
6800 variable) is not atomic.
6803 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6804 *before* setting the 'crypto_lock_rand' flag. The previous code had
6805 a race condition if 0 is a valid thread ID.
6806 [Travis Vitek <vitek@roguewave.com>]
6808 *) Add support for shared libraries under Irix.
6809 [Albert Chin-A-Young <china@thewrittenword.com>]
6811 *) Add configuration option to build on Linux on both big-endian and
6813 [Ralf Baechle <ralf@uni-koblenz.de>]
6815 *) Add the possibility to create shared libraries on HP-UX.
6818 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6820 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6821 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6822 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6823 PRNG state recovery was possible based on the output of
6824 one PRNG request appropriately sized to gain knowledge on
6825 'md' followed by enough consecutive 1-byte PRNG requests
6826 to traverse all of 'state'.
6828 1. When updating 'md_local' (the current thread's copy of 'md')
6829 during PRNG output generation, hash all of the previous
6830 'md_local' value, not just the half used for PRNG output.
6832 2. Make the number of bytes from 'state' included into the hash
6833 independent from the number of PRNG bytes requested.
6835 The first measure alone would be sufficient to avoid
6836 Markku-Juhani's attack. (Actually it had never occurred
6837 to me that the half of 'md_local' used for chaining was the
6838 half from which PRNG output bytes were taken -- I had always
6839 assumed that the secret half would be used.) The second
6840 measure makes sure that additional data from 'state' is never
6841 mixed into 'md_local' in small portions; this heuristically
6842 further strengthens the PRNG.
6845 *) Fix crypto/bn/asm/mips3.s.
6848 *) When only the key is given to "enc", the IV is undefined. Print out
6849 an error message in this case.
6852 *) Handle special case when X509_NAME is empty in X509 printing routines.
6855 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6856 positive and less than q.
6859 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6860 used: it isn't thread safe and the add_lock_callback should handle
6862 [Paul Rose <Paul.Rose@bridge.com>]
6864 *) Verify that incoming data obeys the block size in
6865 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6869 [Ulf Möller, Bodo Möller]
6871 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6872 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6873 when fixing the server behaviour for backwards-compatible 'client
6874 hello' messages. (Note that the attack is impractical against
6875 SSL 3.0 and TLS 1.0 anyway because length and version checking
6876 means that the probability of guessing a valid ciphertext is
6877 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6880 Before 0.9.5, the countermeasure (hide the error by generating a
6881 random 'decryption result') did not work properly because
6882 ERR_clear_error() was missing, meaning that SSL_get_error() would
6883 detect the supposedly ignored error.
6885 Both problems are now fixed.
6888 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6889 (previously it was 1024).
6892 *) Fix for compatibility mode trust settings: ignore trust settings
6893 unless some valid trust or reject settings are present.
6896 *) Fix for blowfish EVP: its a variable length cipher.
6899 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6900 parameters in DSA public key structures and return an error in the
6901 DSA routines if parameters are absent.
6904 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6905 in the current directory if neither $RANDFILE nor $HOME was set.
6906 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6907 caused some confusion to Windows users who haven't defined $HOME.
6908 Thus RAND_file_name() is changed again: e_os.h can define a
6909 DEFAULT_HOME, which will be used if $HOME is not set.
6910 For Windows, we use "C:"; on other platforms, we still require
6911 environment variables.
6913 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6914 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6915 having multiple threads call RAND_poll() concurrently.
6918 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6919 combination of a flag and a thread ID variable.
6920 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6921 flag), *other* threads can enter ssleay_add_bytes without obeying
6922 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6923 that they do not hold after the first thread unsets add_do_not_lock).
6926 *) Change bctest again: '-x' expressions are not available in all
6930 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6932 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6933 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6935 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6936 the default extension for executables, if any. Also, make the perl
6937 scripts that use symlink() to test if it really exists and use "cp"
6938 if it doesn't. All this made OpenSSL compilable and installable in
6942 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6943 If SEQUENCE is length is indefinite just set c->slen to the total
6944 amount of data available.
6945 [Steve Henson, reported by shige@FreeBSD.org]
6946 [This change does not apply to 0.9.7.]
6948 *) Change bctest to avoid here-documents inside command substitution
6949 (workaround for FreeBSD /bin/sh bug).
6950 For compatibility with Ultrix, avoid shell functions (introduced
6951 in the bctest version that searches along $PATH).
6954 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6955 with des_encrypt() defined on some operating systems, like Solaris
6959 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6960 On the Importance of Eliminating Errors in Cryptographic
6961 Computations, J. Cryptology 14 (2001) 2, 101-119,
6962 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6965 *) MIPS assembler BIGNUM division bug fix.
6968 *) Disabled incorrect Alpha assembler code.
6971 *) Fix PKCS#7 decode routines so they correctly update the length
6972 after reading an EOC for the EXPLICIT tag.
6974 [This change does not apply to 0.9.7.]
6976 *) Fix bug in PKCS#12 key generation routines. This was triggered
6977 if a 3DES key was generated with a 0 initial byte. Include
6978 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6979 (but broken) behaviour.
6982 *) Enhance bctest to search for a working bc along $PATH and print
6984 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6986 *) Fix memory leaks in err.c: free err_data string if necessary;
6987 don't write to the wrong index in ERR_set_error_data.
6990 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6994 *) Replace rdtsc with _emit statements for VC++ version 5.
6995 [Jeremy Cooper <jeremy@baymoo.org>]
6997 *) Make it possible to reuse SSLv2 sessions.
7000 *) In copy_email() check for >= 0 as a return value for
7001 X509_NAME_get_index_by_NID() since 0 is a valid index.
7002 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7004 *) Avoid coredump with unsupported or invalid public keys by checking if
7005 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7006 PKCS7_verify() fails with non detached data.
7009 *) Don't use getenv in library functions when run as setuid/setgid.
7010 New function OPENSSL_issetugid().
7013 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7014 due to incorrect handling of multi-threading:
7016 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7018 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7020 3. Count how many times MemCheck_off() has been called so that
7021 nested use can be treated correctly. This also avoids
7022 inband-signalling in the previous code (which relied on the
7023 assumption that thread ID 0 is impossible).
7026 *) Add "-rand" option also to s_client and s_server.
7029 *) Fix CPU detection on Irix 6.x.
7030 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7031 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7033 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7036 [This change does not apply to 0.9.7.]
7038 *) Use the cached encoding of an X509_NAME structure rather than
7039 copying it. This is apparently the reason for the libsafe "errors"
7040 but the code is actually correct.
7043 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7044 Bleichenbacher's DSA attack.
7045 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7046 to be set and top=0 forces the highest bit to be set; top=-1 is new
7047 and leaves the highest bit random.
7048 [Ulf Moeller, Bodo Moeller]
7050 *) In the NCONF_...-based implementations for CONF_... queries
7051 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7052 a temporary CONF structure with the data component set to NULL
7053 (which gives segmentation faults in lh_retrieve).
7054 Instead, use NULL for the CONF pointer in CONF_get_string and
7055 CONF_get_number (which may use environment variables) and directly
7056 return NULL from CONF_get_section.
7059 *) Fix potential buffer overrun for EBCDIC.
7062 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7063 keyUsage if basicConstraints absent for a CA.
7066 *) Make SMIME_write_PKCS7() write mail header values with a format that
7067 is more generally accepted (no spaces before the semicolon), since
7068 some programs can't parse those values properly otherwise. Also make
7069 sure BIO's that break lines after each write do not create invalid
7073 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7074 macros previously used would not encode an empty SEQUENCE OF
7075 and break the signature.
7077 [This change does not apply to 0.9.7.]
7079 *) Zero the premaster secret after deriving the master secret in
7083 *) Add some EVP_add_digest_alias registrations (as found in
7084 OpenSSL_add_all_digests()) to SSL_library_init()
7085 aka OpenSSL_add_ssl_algorithms(). This provides improved
7086 compatibility with peers using X.509 certificates
7087 with unconventional AlgorithmIdentifier OIDs.
7090 *) Fix for Irix with NO_ASM.
7091 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7093 *) ./config script fixes.
7094 [Ulf Moeller, Richard Levitte]
7096 *) Fix 'openssl passwd -1'.
7099 *) Change PKCS12_key_gen_asc() so it can cope with non null
7100 terminated strings whose length is passed in the passlen
7101 parameter, for example from PEM callbacks. This was done
7102 by adding an extra length parameter to asc2uni().
7103 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7105 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7106 call failed, free the DSA structure.
7109 *) Fix to uni2asc() to cope with zero length Unicode strings.
7110 These are present in some PKCS#12 files.
7113 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7114 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7115 when writing a 32767 byte record.
7116 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7118 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7119 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7121 (RSA objects have a reference count access to which is protected
7122 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7123 so they are meant to be shared between threads.)
7124 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7125 "Reddie, Steven" <Steven.Reddie@ca.com>]
7127 *) Fix a deadlock in CRYPTO_mem_leaks().
7130 *) Use better test patterns in bntest.
7133 *) rand_win.c fix for Borland C.
7136 *) BN_rshift bugfix for n == 0.
7139 *) Add a 'bctest' script that checks for some known 'bc' bugs
7140 so that 'make test' does not abort just because 'bc' is broken.
7143 *) Store verify_result within SSL_SESSION also for client side to
7144 avoid potential security hole. (Re-used sessions on the client side
7145 always resulted in verify_result==X509_V_OK, not using the original
7146 result of the server certificate verification.)
7149 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7150 SSL3_RT_APPLICATION_DATA, return 0.
7151 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7155 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7156 releases, have been re-implemented by renaming the previous
7157 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7158 and ssl3_read_internal, respectively, and adding 'peek' parameters
7159 to them. The new ssl[23]_{read,peek} functions are calls to
7160 ssl[23]_read_internal with the 'peek' flag set appropriately.
7161 A 'peek' parameter has also been added to ssl3_read_bytes, which
7162 does the actual work for ssl3_read_internal.
7165 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7166 the method-specific "init()" handler. Also clean up ex_data after
7167 calling the method-specific "finish()" handler. Previously, this was
7168 happening the other way round.
7171 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7172 The previous value, 12, was not always sufficient for BN_mod_exp().
7175 *) Make sure that shared libraries get the internal name engine with
7176 the full version number and not just 0. This should mark the
7177 shared libraries as not backward compatible. Of course, this should
7178 be changed again when we can guarantee backward binary compatibility.
7181 *) Fix typo in get_cert_by_subject() in by_dir.c
7182 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7184 *) Rework the system to generate shared libraries:
7186 - Make note of the expected extension for the shared libraries and
7187 if there is a need for symbolic links from for example libcrypto.so.0
7188 to libcrypto.so.0.9.7. There is extended info in Configure for
7191 - Make as few rebuilds of the shared libraries as possible.
7193 - Still avoid linking the OpenSSL programs with the shared libraries.
7195 - When installing, install the shared libraries separately from the
7199 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7201 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7202 and not in SSL_clear because the latter is also used by the
7203 accept/connect functions; previously, the settings made by
7204 SSL_set_read_ahead would be lost during the handshake.
7205 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7207 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7208 Previously, it would create entries for disableed algorithms no
7212 *) Added several new manual pages for SSL_* function.
7215 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7217 *) In ssl23_get_client_hello, generate an error message when faced
7218 with an initial SSL 3.0/TLS record that is too small to contain the
7219 first two bytes of the ClientHello message, i.e. client_version.
7220 (Note that this is a pathologic case that probably has never happened
7221 in real life.) The previous approach was to use the version number
7222 from the record header as a substitute; but our protocol choice
7223 should not depend on that one because it is not authenticated
7224 by the Finished messages.
7227 *) More robust randomness gathering functions for Windows.
7228 [Jeffrey Altman <jaltman@columbia.edu>]
7230 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7231 not set then we don't setup the error code for issuer check errors
7232 to avoid possibly overwriting other errors which the callback does
7233 handle. If an application does set the flag then we assume it knows
7234 what it is doing and can handle the new informational codes
7238 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7239 a general "ANY" type, as such it should be able to decode anything
7240 including tagged types. However it didn't check the class so it would
7241 wrongly interpret tagged types in the same way as their universal
7242 counterpart and unknown types were just rejected. Changed so that the
7243 tagged and unknown types are handled in the same way as a SEQUENCE:
7244 that is the encoding is stored intact. There is also a new type
7245 "V_ASN1_OTHER" which is used when the class is not universal, in this
7246 case we have no idea what the actual type is so we just lump them all
7250 *) On VMS, stdout may very well lead to a file that is written to
7251 in a record-oriented fashion. That means that every write() will
7252 write a separate record, which will be read separately by the
7253 programs trying to read from it. This can be very confusing.
7255 The solution is to put a BIO filter in the way that will buffer
7256 text until a linefeed is reached, and then write everything a
7257 line at a time, so every record written will be an actual line,
7258 not chunks of lines and not (usually doesn't happen, but I've
7259 seen it once) several lines in one record. BIO_f_linebuffer() is
7262 Currently, it's a VMS-only method, because that's where it has
7263 been tested well enough.
7266 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7267 it can return incorrect results.
7268 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7269 but it was in 0.9.6-beta[12].)
7272 *) Disable the check for content being present when verifying detached
7273 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7274 include zero length content when signing messages.
7277 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7278 BIO_ctrl (for BIO pairs).
7281 *) Add DSO method for VMS.
7284 *) Bug fix: Montgomery multiplication could produce results with the
7288 *) Add RPM specification openssl.spec and modify it to build three
7289 packages. The default package contains applications, application
7290 documentation and run-time libraries. The devel package contains
7291 include files, static libraries and function documentation. The
7292 doc package contains the contents of the doc directory. The original
7293 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7296 *) Add a large number of documentation files for many SSL routines.
7297 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7299 *) Add a configuration entry for Sony News 4.
7300 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7302 *) Don't set the two most significant bits to one when generating a
7303 random number < q in the DSA library.
7306 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7307 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7308 the underlying transport is blocking) if a handshake took place.
7309 (The default behaviour is needed by applications such as s_client
7310 and s_server that use select() to determine when to use SSL_read;
7311 but for applications that know in advance when to expect data, it
7312 just makes things more complicated.)
7315 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7319 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7320 work better on such systems.
7321 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7323 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7324 Update PKCS12_parse() so it copies the friendlyName and the
7325 keyid to the certificates aux info.
7328 *) Fix bug in PKCS7_verify() which caused an infinite loop
7329 if there was more than one signature.
7330 [Sven Uszpelkat <su@celocom.de>]
7332 *) Major change in util/mkdef.pl to include extra information
7333 about each symbol, as well as presentig variables as well
7334 as functions. This change means that there's n more need
7335 to rebuild the .num files when some algorithms are excluded.
7338 *) Allow the verify time to be set by an application,
7339 rather than always using the current time.
7342 *) Phase 2 verify code reorganisation. The certificate
7343 verify code now looks up an issuer certificate by a
7344 number of criteria: subject name, authority key id
7345 and key usage. It also verifies self signed certificates
7346 by the same criteria. The main comparison function is
7347 X509_check_issued() which performs these checks.
7349 Lot of changes were necessary in order to support this
7350 without completely rewriting the lookup code.
7352 Authority and subject key identifier are now cached.
7354 The LHASH 'certs' is X509_STORE has now been replaced
7355 by a STACK_OF(X509_OBJECT). This is mainly because an
7356 LHASH can't store or retrieve multiple objects with
7357 the same hash value.
7359 As a result various functions (which were all internal
7360 use only) have changed to handle the new X509_STORE
7361 structure. This will break anything that messed round
7362 with X509_STORE internally.
7364 The functions X509_STORE_add_cert() now checks for an
7365 exact match, rather than just subject name.
7367 The X509_STORE API doesn't directly support the retrieval
7368 of multiple certificates matching a given criteria, however
7369 this can be worked round by performing a lookup first
7370 (which will fill the cache with candidate certificates)
7371 and then examining the cache for matches. This is probably
7372 the best we can do without throwing out X509_LOOKUP
7373 entirely (maybe later...).
7375 The X509_VERIFY_CTX structure has been enhanced considerably.
7377 All certificate lookup operations now go via a get_issuer()
7378 callback. Although this currently uses an X509_STORE it
7379 can be replaced by custom lookups. This is a simple way
7380 to bypass the X509_STORE hackery necessary to make this
7381 work and makes it possible to use more efficient techniques
7382 in future. A very simple version which uses a simple
7383 STACK for its trusted certificate store is also provided
7384 using X509_STORE_CTX_trusted_stack().
7386 The verify_cb() and verify() callbacks now have equivalents
7387 in the X509_STORE_CTX structure.
7389 X509_STORE_CTX also has a 'flags' field which can be used
7390 to customise the verify behaviour.
7393 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7394 excludes S/MIME capabilities.
7397 *) When a certificate request is read in keep a copy of the
7398 original encoding of the signed data and use it when outputing
7399 again. Signatures then use the original encoding rather than
7400 a decoded, encoded version which may cause problems if the
7401 request is improperly encoded.
7404 *) For consistency with other BIO_puts implementations, call
7405 buffer_write(b, ...) directly in buffer_puts instead of calling
7408 In BIO_puts, increment b->num_write as in BIO_write.
7409 [Peter.Sylvester@EdelWeb.fr]
7411 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7412 BN_zero, we may not return a BIGNUM with an array consisting of
7416 *) Avoid calling abort() from within the library when problems are
7417 detected, except if preprocessor symbols have been defined
7418 (such as REF_CHECK, BN_DEBUG etc.).
7421 *) New openssl application 'rsautl'. This utility can be
7422 used for low level RSA operations. DER public key
7423 BIO/fp routines also added.
7426 *) New Configure entry and patches for compiling on QNX 4.
7427 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7429 *) A demo state-machine implementation was sponsored by
7430 Nuron (http://www.nuron.com/) and is now available in
7431 demos/state_machine.
7434 *) New options added to the 'dgst' utility for signature
7435 generation and verification.
7438 *) Unrecognized PKCS#7 content types are now handled via a
7439 catch all ASN1_TYPE structure. This allows unsupported
7440 types to be stored as a "blob" and an application can
7441 encode and decode it manually.
7444 *) Fix various signed/unsigned issues to make a_strex.c
7446 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7448 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7449 length if passed a buffer. ASN1_INTEGER_to_BN failed
7450 if passed a NULL BN and its argument was negative.
7451 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7453 *) Modification to PKCS#7 encoding routines to output definite
7454 length encoding. Since currently the whole structures are in
7455 memory there's not real point in using indefinite length
7456 constructed encoding. However if OpenSSL is compiled with
7457 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7460 *) Added BIO_vprintf() and BIO_vsnprintf().
7463 *) Added more prefixes to parse for in the the strings written
7464 through a logging bio, to cover all the levels that are available
7465 through syslog. The prefixes are now:
7467 PANIC, EMERG, EMR => LOG_EMERG
7468 ALERT, ALR => LOG_ALERT
7469 CRIT, CRI => LOG_CRIT
7470 ERROR, ERR => LOG_ERR
7471 WARNING, WARN, WAR => LOG_WARNING
7472 NOTICE, NOTE, NOT => LOG_NOTICE
7473 INFO, INF => LOG_INFO
7474 DEBUG, DBG => LOG_DEBUG
7476 and as before, if none of those prefixes are present at the
7477 beginning of the string, LOG_ERR is chosen.
7479 On Win32, the LOG_* levels are mapped according to this:
7481 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7482 LOG_WARNING => EVENTLOG_WARNING_TYPE
7483 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7487 *) Made it possible to reconfigure with just the configuration
7488 argument "reconf" or "reconfigure". The command line arguments
7489 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7490 and are retrieved from there when reconfiguring.
7494 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7496 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7499 *) The obj_dat.pl script was messing up the sorting of object
7500 names. The reason was that it compared the quoted version
7501 of strings as a result "OCSP" > "OCSP Signing" because
7502 " > SPACE. Changed script to store unquoted versions of
7503 names and add quotes on output. It was also omitting some
7504 names from the lookup table if they were given a default
7505 value (that is if SN is missing it is given the same
7506 value as LN and vice versa), these are now added on the
7507 grounds that if an object has a name we should be able to
7508 look it up. Finally added warning output when duplicate
7509 short or long names are found.
7512 *) Changes needed for Tandem NSK.
7513 [Scott Uroff <scott@xypro.com>]
7515 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7516 RSA_padding_check_SSLv23(), special padding was never detected
7517 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7518 version rollback attacks was not effective.
7520 In s23_clnt.c, don't use special rollback-attack detection padding
7521 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7522 client; similarly, in s23_srvr.c, don't do the rollback check if
7523 SSL 2.0 is the only protocol enabled in the server.
7526 *) Make it possible to get hexdumps of unprintable data with 'openssl
7527 asn1parse'. By implication, the functions ASN1_parse_dump() and
7528 BIO_dump_indent() are added.
7531 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7532 these print out strings and name structures based on various
7533 flags including RFC2253 support and proper handling of
7534 multibyte characters. Added options to the 'x509' utility
7535 to allow the various flags to be set.
7538 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7539 Also change the functions X509_cmp_current_time() and
7540 X509_gmtime_adj() work with an ASN1_TIME structure,
7541 this will enable certificates using GeneralizedTime in validity
7542 dates to be checked.
7545 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7546 negative public key encodings) on by default,
7547 NO_NEG_PUBKEY_BUG can be set to disable it.
7550 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7551 content octets. An i2c_ASN1_OBJECT is unnecessary because
7552 the encoding can be trivially obtained from the structure.
7555 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7556 not read locks (CRYPTO_r_[un]lock).
7559 *) A first attempt at creating official support for shared
7560 libraries through configuration. I've kept it so the
7561 default is static libraries only, and the OpenSSL programs
7562 are always statically linked for now, but there are
7563 preparations for dynamic linking in place.
7564 This has been tested on Linux and Tru64.
7567 *) Randomness polling function for Win9x, as described in:
7568 Peter Gutmann, Software Generation of Practically Strong
7572 *) Fix so PRNG is seeded in req if using an already existing
7576 *) New options to smime application. -inform and -outform
7577 allow alternative formats for the S/MIME message including
7578 PEM and DER. The -content option allows the content to be
7579 specified separately. This should allow things like Netscape
7580 form signing output easier to verify.
7583 *) Fix the ASN1 encoding of tags using the 'long form'.
7586 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7587 STRING types. These convert content octets to and from the
7588 underlying type. The actual tag and length octets are
7589 already assumed to have been read in and checked. These
7590 are needed because all other string types have virtually
7591 identical handling apart from the tag. By having versions
7592 of the ASN1 functions that just operate on content octets
7593 IMPLICIT tagging can be handled properly. It also allows
7594 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7595 and ASN1_INTEGER are identical apart from the tag.
7598 *) Change the handling of OID objects as follows:
7600 - New object identifiers are inserted in objects.txt, following
7601 the syntax given in objects.README.
7602 - objects.pl is used to process obj_mac.num and create a new
7604 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7607 This is currently kind of a hack, and the perl code in objects.pl
7608 isn't very elegant, but it works as I intended. The simplest way
7609 to check that it worked correctly is to look in obj_dat.h and
7610 check the array nid_objs and make sure the objects haven't moved
7611 around (this is important!). Additions are OK, as well as
7612 consistent name changes.
7615 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7618 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7619 The given file adds to whatever has already been seeded into the
7620 random pool through the RANDFILE configuration file option or
7621 environment variable, or the default random state file.
7624 *) mkstack.pl now sorts each macro group into lexical order.
7625 Previously the output order depended on the order the files
7626 appeared in the directory, resulting in needless rewriting
7630 *) Patches to make OpenSSL compile under Win32 again. Mostly
7631 work arounds for the VC++ problem that it treats func() as
7632 func(void). Also stripped out the parts of mkdef.pl that
7633 added extra typesafe functions: these no longer exist.
7636 *) Reorganisation of the stack code. The macros are now all
7637 collected in safestack.h . Each macro is defined in terms of
7638 a "stack macro" of the form SKM_<name>(type, a, b). The
7639 DEBUG_SAFESTACK is now handled in terms of function casts,
7640 this has the advantage of retaining type safety without the
7641 use of additional functions. If DEBUG_SAFESTACK is not defined
7642 then the non typesafe macros are used instead. Also modified the
7643 mkstack.pl script to handle the new form. Needs testing to see
7644 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7645 the default if no major problems. Similar behaviour for ASN1_SET_OF
7646 and PKCS12_STACK_OF.
7649 *) When some versions of IIS use the 'NET' form of private key the
7650 key derivation algorithm is different. Normally MD5(password) is
7651 used as a 128 bit RC4 key. In the modified case
7652 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7653 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7654 as the old Netscape_RSA functions except they have an additional
7655 'sgckey' parameter which uses the modified algorithm. Also added
7656 an -sgckey command line option to the rsa utility. Thanks to
7657 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7658 algorithm to openssl-dev.
7661 *) The evp_local.h macros were using 'c.##kname' which resulted in
7662 invalid expansion on some systems (SCO 5.0.5 for example).
7663 Corrected to 'c.kname'.
7664 [Phillip Porch <root@theporch.com>]
7666 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7667 a STACK of email addresses from a certificate or request, these look
7668 in the subject name and the subject alternative name extensions and
7669 omit any duplicate addresses.
7672 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7673 This makes DSA verification about 2 % faster.
7676 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7677 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7678 plus overhead for 1024 bit moduli).
7679 This makes exponentiations about 0.5 % faster for 1024 bit
7680 exponents (as measured by "openssl speed rsa2048").
7683 *) Rename memory handling macros to avoid conflicts with other
7685 Malloc => OPENSSL_malloc
7686 Malloc_locked => OPENSSL_malloc_locked
7687 Realloc => OPENSSL_realloc
7688 Free => OPENSSL_free
7691 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7692 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7695 *) CygWin32 support.
7696 [John Jarvie <jjarvie@newsguy.com>]
7698 *) The type-safe stack code has been rejigged. It is now only compiled
7699 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7700 by default all type-specific stack functions are "#define"d back to
7701 standard stack functions. This results in more streamlined output
7702 but retains the type-safety checking possibilities of the original
7706 *) The STACK code has been cleaned up, and certain type declarations
7707 that didn't make a lot of sense have been brought in line. This has
7708 also involved a cleanup of sorts in safestack.h to more correctly
7709 map type-safe stack functions onto their plain stack counterparts.
7710 This work has also resulted in a variety of "const"ifications of
7711 lots of the code, especially "_cmp" operations which should normally
7712 be prototyped with "const" parameters anyway.
7715 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7716 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7717 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7718 where all of 'md' is used each time the PRNG is used, but 'state'
7719 is used only indexed by a cyclic counter. As entropy may not be
7720 well distributed from the beginning, 'md' is important as a
7721 chaining variable. However, the output function chains only half
7722 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7723 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7724 in all of 'state' being rewritten, with the new values depending
7725 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7728 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7729 the handshake is continued after ssl_verify_cert_chain();
7730 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7731 can lead to 'unexplainable' connection aborts later.
7732 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7734 *) Major EVP API cipher revision.
7735 Add hooks for extra EVP features. This allows various cipher
7736 parameters to be set in the EVP interface. Support added for variable
7737 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7738 setting of RC2 and RC5 parameters.
7740 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7743 Remove lots of duplicated code from the EVP library. For example *every*
7744 cipher init() function handles the 'iv' in the same way according to the
7745 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7746 for CFB and OFB modes they zero ctx->num.
7748 New functionality allows removal of S/MIME code RC2 hack.
7750 Most of the routines have the same form and so can be declared in terms
7753 By shifting this to the top level EVP_CipherInit() it can be removed from
7754 all individual ciphers. If the cipher wants to handle IVs or keys
7755 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7758 Change lots of functions like EVP_EncryptUpdate() to now return a
7759 value: although software versions of the algorithms cannot fail
7760 any installed hardware versions can.
7763 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7764 this option is set, tolerate broken clients that send the negotiated
7765 protocol version number instead of the requested protocol version
7769 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7770 i.e. non-zero for export ciphersuites, zero otherwise.
7771 Previous versions had this flag inverted, inconsistent with
7772 rsa_tmp_cb (..._TMP_RSA_CB).
7773 [Bodo Moeller; problem reported by Amit Chopra]
7775 *) Add missing DSA library text string. Work around for some IIS
7776 key files with invalid SEQUENCE encoding.
7779 *) Add a document (doc/standards.txt) that list all kinds of standards
7780 and so on that are implemented in OpenSSL.
7783 *) Enhance c_rehash script. Old version would mishandle certificates
7784 with the same subject name hash and wouldn't handle CRLs at all.
7785 Added -fingerprint option to crl utility, to support new c_rehash
7789 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7792 *) Fix for SSL server purpose checking. Server checking was
7793 rejecting certificates which had extended key usage present
7794 but no ssl client purpose.
7795 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7797 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7798 is a little unclear about how a blank password is handled.
7799 Since the password in encoded as a BMPString with terminating
7800 double NULL a zero length password would end up as just the
7801 double NULL. However no password at all is different and is
7802 handled differently in the PKCS#12 key generation code. NS
7803 treats a blank password as zero length. MSIE treats it as no
7804 password on export: but it will try both on import. We now do
7805 the same: PKCS12_parse() tries zero length and no password if
7806 the password is set to "" or NULL (NULL is now a valid password:
7807 it wasn't before) as does the pkcs12 application.
7810 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7811 perror when PEM_read_bio_X509_REQ fails, the error message must
7812 be obtained from the error queue.
7815 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7816 it in ERR_remove_state if appropriate, and change ERR_get_state
7817 accordingly to avoid race conditions (this is necessary because
7818 thread_hash is no longer constant once set).
7821 *) Bugfix for linux-elf makefile.one.
7824 *) RSA_get_default_method() will now cause a default
7825 RSA_METHOD to be chosen if one doesn't exist already.
7826 Previously this was only set during a call to RSA_new()
7827 or RSA_new_method(NULL) meaning it was possible for
7828 RSA_get_default_method() to return NULL.
7831 *) Added native name translation to the existing DSO code
7832 that will convert (if the flag to do so is set) filenames
7833 that are sufficiently small and have no path information
7834 into a canonical native form. Eg. "blah" converted to
7835 "libblah.so" or "blah.dll" etc.
7838 *) New function ERR_error_string_n(e, buf, len) which is like
7839 ERR_error_string(e, buf), but writes at most 'len' bytes
7840 including the 0 terminator. For ERR_error_string_n, 'buf'
7842 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7844 *) CONF library reworked to become more general. A new CONF
7845 configuration file reader "class" is implemented as well as a
7846 new functions (NCONF_*, for "New CONF") to handle it. The now
7847 old CONF_* functions are still there, but are reimplemented to
7848 work in terms of the new functions. Also, a set of functions
7849 to handle the internal storage of the configuration data is
7850 provided to make it easier to write new configuration file
7851 reader "classes" (I can definitely see something reading a
7852 configuration file in XML format, for example), called _CONF_*,
7853 or "the configuration storage API"...
7855 The new configuration file reading functions are:
7857 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7858 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7860 NCONF_default, NCONF_WIN32
7862 NCONF_dump_fp, NCONF_dump_bio
7864 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7865 NCONF_new creates a new CONF object. This works in the same way
7866 as other interfaces in OpenSSL, like the BIO interface.
7867 NCONF_dump_* dump the internal storage of the configuration file,
7868 which is useful for debugging. All other functions take the same
7869 arguments as the old CONF_* functions wth the exception of the
7870 first that must be a `CONF *' instead of a `LHASH *'.
7872 To make it easer to use the new classes with the old CONF_* functions,
7873 the function CONF_set_default_method is provided.
7876 *) Add '-tls1' option to 'openssl ciphers', which was already
7877 mentioned in the documentation but had not been implemented.
7878 (This option is not yet really useful because even the additional
7879 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7882 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7883 OpenSSL-based applications) load shared libraries and bind to
7884 them in a portable way.
7885 [Geoff Thorpe, with contributions from Richard Levitte]
7887 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7889 *) Make sure _lrotl and _lrotr are only used with MSVC.
7891 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7892 (the default implementation of RAND_status).
7894 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7895 to '-clrext' (= clear extensions), as intended and documented.
7896 [Bodo Moeller; inconsistency pointed out by Michael Attili
7899 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7900 was larger than the MD block size.
7901 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7903 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7904 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7905 using the passed key: if the passed key was a private key the result
7906 of X509_print(), for example, would be to print out all the private key
7910 *) des_quad_cksum() byte order bug fix.
7911 [Ulf Möller, using the problem description in krb4-0.9.7, where
7912 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7914 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7916 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7918 *) For easily testing in shell scripts whether some command
7919 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7920 returns with exit code 0 iff no command of the given name is available.
7921 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7922 the output goes to stdout and nothing is printed to stderr.
7923 Additional arguments are always ignored.
7925 Since for each cipher there is a command of the same name,
7926 the 'no-cipher' compilation switches can be tested this way.
7928 ('openssl no-XXX' is not able to detect pseudo-commands such
7929 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7932 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7935 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7936 is set; it will be thrown away anyway because each handshake creates
7938 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7939 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7940 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7941 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7944 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7945 'Q' and 'R' lose their special meanings (quit/renegotiate).
7946 This is part of what -quiet does; unlike -quiet, -ign_eof
7947 does not suppress any output.
7950 *) Add compatibility options to the purpose and trust code. The
7951 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7952 accepts a certificate or CA, this was the previous behaviour,
7953 with all the associated security issues.
7955 X509_TRUST_COMPAT is the old trust behaviour: only and
7956 automatically trust self signed roots in certificate store. A
7957 new trust setting X509_TRUST_DEFAULT is used to specify that
7958 a purpose has no associated trust setting and it should instead
7959 use the value in the default purpose.
7962 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7963 and fix a memory leak.
7966 *) In util/mkerr.pl (which implements 'make errors'), preserve
7967 reason strings from the previous version of the .c file, as
7968 the default to have only downcase letters (and digits) in
7969 automatically generated reasons codes is not always appropriate.
7972 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7973 using strerror. Previously, ERR_reason_error_string() returned
7974 library names as reason strings for SYSerr; but SYSerr is a special
7975 case where small numbers are errno values, not library numbers.
7978 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7979 converts DSA parameters into DH parameters. (When creating parameters,
7980 DSA_generate_parameters is used.)
7983 *) Include 'length' (recommended exponent length) in C code generated
7984 by 'openssl dhparam -C'.
7987 *) The second argument to set_label in perlasm was already being used
7988 so couldn't be used as a "file scope" flag. Moved to third argument
7992 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7993 instead of RAND_bytes for encryption IVs and salts.
7996 *) Include RAND_status() into RAND_METHOD instead of implementing
7997 it only for md_rand.c Otherwise replacing the PRNG by calling
7998 RAND_set_rand_method would be impossible.
8001 *) Don't let DSA_generate_key() enter an infinite loop if the random
8002 number generation fails.
8005 *) New 'rand' application for creating pseudo-random output.
8008 *) Added configuration support for Linux/IA64
8009 [Rolf Haberrecker <rolf@suse.de>]
8011 *) Assembler module support for Mingw32.
8014 *) Shared library support for HPUX (in shlib/).
8015 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8017 *) Shared library support for Solaris gcc.
8018 [Lutz Behnke <behnke@trustcenter.de>]
8020 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8022 *) PKCS7_encrypt() was adding text MIME headers twice because they
8023 were added manually and by SMIME_crlf_copy().
8026 *) In bntest.c don't call BN_rand with zero bits argument.
8027 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8029 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8030 case was implemented. This caused BN_div_recp() to fail occasionally.
8033 *) Add an optional second argument to the set_label() in the perl
8034 assembly language builder. If this argument exists and is set
8035 to 1 it signals that the assembler should use a symbol whose
8036 scope is the entire file, not just the current function. This
8037 is needed with MASM which uses the format label:: for this scope.
8038 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8040 *) Change the ASN1 types so they are typedefs by default. Before
8041 almost all types were #define'd to ASN1_STRING which was causing
8042 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8046 *) Change names of new functions to the new get1/get0 naming
8047 convention: After 'get1', the caller owns a reference count
8048 and has to call ..._free; 'get0' returns a pointer to some
8049 data structure without incrementing reference counters.
8050 (Some of the existing 'get' functions increment a reference
8051 counter, some don't.)
8052 Similarly, 'set1' and 'add1' functions increase reference
8053 counters or duplicate objects.
8056 *) Allow for the possibility of temp RSA key generation failure:
8057 the code used to assume it always worked and crashed on failure.
8060 *) Fix potential buffer overrun problem in BIO_printf().
8061 [Ulf Möller, using public domain code by Patrick Powell; problem
8062 pointed out by David Sacerdote <das33@cornell.edu>]
8064 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8065 RAND_egd() and RAND_status(). In the command line application,
8066 the EGD socket can be specified like a seed file using RANDFILE
8070 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8071 Some CAs (e.g. Verisign) distribute certificates in this form.
8074 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8075 list to exclude them. This means that no special compilation option
8076 is needed to use anonymous DH: it just needs to be included in the
8080 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8081 EVP_MD_type. The old functionality is available in a new macro called
8082 EVP_MD_md(). Change code that uses it and update docs.
8085 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8086 where the 'void *' argument is replaced by a function pointer argument.
8087 Previously 'void *' was abused to point to functions, which works on
8088 many platforms, but is not correct. As these functions are usually
8089 called by macros defined in OpenSSL header files, most source code
8090 should work without changes.
8093 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8094 sections with information on -D... compiler switches used for
8095 compiling the library so that applications can see them. To enable
8096 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8097 must be defined. E.g.,
8098 #define OPENSSL_ALGORITHM_DEFINES
8099 #include <openssl/opensslconf.h>
8100 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8101 [Richard Levitte, Ulf and Bodo Möller]
8103 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8107 *) Change the 'other' type in certificate aux info to a STACK_OF
8108 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8109 the required ASN1 format: arbitrary types determined by an OID.
8112 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8113 argument to 'req'. This is not because the function is newer or
8114 better than others it just uses the work 'NEW' in the certificate
8115 request header lines. Some software needs this.
8118 *) Reorganise password command line arguments: now passwords can be
8119 obtained from various sources. Delete the PEM_cb function and make
8120 it the default behaviour: i.e. if the callback is NULL and the
8121 usrdata argument is not NULL interpret it as a null terminated pass
8122 phrase. If usrdata and the callback are NULL then the pass phrase
8123 is prompted for as usual.
8126 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8127 the support is automatically enabled. The resulting binaries will
8128 autodetect the card and use it if present.
8129 [Ben Laurie and Compaq Inc.]
8131 *) Work around for Netscape hang bug. This sends certificate request
8132 and server done in one record. Since this is perfectly legal in the
8133 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8134 the bugs/SSLv3 entry for more info.
8137 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8140 *) Add -rand argument to smime and pkcs12 applications and read/write
8144 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8147 *) Add command line password options to the remaining applications.
8150 *) Bug fix for BN_div_recp() for numerators with an even number of
8154 *) More tests in bntest.c, and changed test_bn output.
8157 *) ./config recognizes MacOS X now.
8160 *) Bug fix for BN_div() when the first words of num and divsor are
8161 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8164 *) Add support for various broken PKCS#8 formats, and command line
8165 options to produce them.
8168 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8169 get temporary BIGNUMs from a BN_CTX.
8172 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8176 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8177 include a #define from the old name to the new. The original intent
8178 was that statically linked binaries could for example just call
8179 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8180 link with digests. This never worked becayse SSLeay_add_all_digests()
8181 and SSLeay_add_all_ciphers() were in the same source file so calling
8182 one would link with the other. They are now in separate source files.
8185 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8188 *) Use a less unusual form of the Miller-Rabin primality test (it used
8189 a binary algorithm for exponentiation integrated into the Miller-Rabin
8190 loop, our standard modexp algorithms are faster).
8193 *) Support for the EBCDIC character set completed.
8194 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8196 *) Source code cleanups: use const where appropriate, eliminate casts,
8197 use void * instead of char * in lhash.
8200 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8201 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8202 this the server could overwrite ephemeral keys that the client
8206 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8207 using 50 iterations of the Rabin-Miller test.
8209 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8210 iterations of the Rabin-Miller test as required by the appendix
8211 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8212 As BN_is_prime_fasttest includes trial division, DSA parameter
8213 generation becomes much faster.
8215 This implies a change for the callback functions in DSA_is_prime
8216 and DSA_generate_parameters: The callback function is called once
8217 for each positive witness in the Rabin-Miller test, not just
8218 occasionally in the inner loop; and the parameters to the
8219 callback function now provide an iteration count for the outer
8220 loop rather than for the current invocation of the inner loop.
8221 DSA_generate_parameters additionally can call the callback
8222 function with an 'iteration count' of -1, meaning that a
8223 candidate has passed the trial division test (when q is generated
8224 from an application-provided seed, trial division is skipped).
8227 *) New function BN_is_prime_fasttest that optionally does trial
8228 division before starting the Rabin-Miller test and has
8229 an additional BN_CTX * argument (whereas BN_is_prime always
8230 has to allocate at least one BN_CTX).
8231 'callback(1, -1, cb_arg)' is called when a number has passed the
8232 trial division stage.
8235 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8239 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8242 *) New function BN_pseudo_rand().
8245 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8246 bignum version of BN_from_montgomery() with the working code from
8247 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8251 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8252 made it impossible to use the same SSL_SESSION data structure in
8253 SSL2 clients in multiple threads.
8256 *) The return value of RAND_load_file() no longer counts bytes obtained
8257 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8258 to seed the PRNG (previously an explicit byte count was required).
8259 [Ulf Möller, Bodo Möller]
8261 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8262 used (char *) instead of (void *) and had casts all over the place.
8265 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8268 *) Retain source code compatibility for BN_prime_checks macro:
8269 BN_is_prime(..., BN_prime_checks, ...) now uses
8270 BN_prime_checks_for_size to determine the appropriate number of
8271 Rabin-Miller iterations.
8274 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8275 DH_CHECK_P_NOT_SAFE_PRIME.
8276 (Check if this is true? OpenPGP calls them "strong".)
8279 *) Merge the functionality of "dh" and "gendh" programs into a new program
8280 "dhparam". The old programs are retained for now but will handle DH keys
8281 (instead of parameters) in future.
8284 *) Make the ciphers, s_server and s_client programs check the return values
8285 when a new cipher list is set.
8288 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8289 ciphers. Before when the 56bit ciphers were enabled the sorting was
8292 The syntax for the cipher sorting has been extended to support sorting by
8293 cipher-strength (using the strength_bits hard coded in the tables).
8294 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8296 Fix a bug in the cipher-command parser: when supplying a cipher command
8297 string with an "undefined" symbol (neither command nor alphanumeric
8298 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8299 an error is flagged.
8301 Due to the strength-sorting extension, the code of the
8302 ssl_create_cipher_list() function was completely rearranged. I hope that
8303 the readability was also increased :-)
8304 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8306 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8307 for the first serial number and places 2 in the serial number file. This
8308 avoids problems when the root CA is created with serial number zero and
8309 the first user certificate has the same issuer name and serial number
8313 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8314 the new code. Add documentation for this stuff.
8317 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8318 X509_*() to X509at_*() on the grounds that they don't handle X509
8319 structures and behave in an analagous way to the X509v3 functions:
8320 they shouldn't be called directly but wrapper functions should be used
8323 So we also now have some wrapper functions that call the X509at functions
8324 when passed certificate requests. (TO DO: similar things can be done with
8325 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8326 things. Some of these need some d2i or i2d and print functionality
8327 because they handle more complex structures.)
8330 *) Add missing #ifndefs that caused missing symbols when building libssl
8331 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8332 NO_RSA in ssl/s2*.c.
8333 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8335 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8336 has a return value which indicates the quality of the random data
8337 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8338 error queue. New function RAND_pseudo_bytes() generates output that is
8339 guaranteed to be unique but not unpredictable. RAND_add is like
8340 RAND_seed, but takes an extra argument for an entropy estimate
8341 (RAND_seed always assumes full entropy).
8344 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8345 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8346 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8347 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8348 false-positive rate of at most 2^-80 for random input.
8351 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8354 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8355 in the 0.9.5 release), this returns the chain
8356 from an X509_CTX structure with a dup of the stack and all
8357 the X509 reference counts upped: so the stack will exist
8358 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8361 Also make SSL_SESSION_print() print out the verify return
8365 *) Add manpage for the pkcs12 command. Also change the default
8366 behaviour so MAC iteration counts are used unless the new
8367 -nomaciter option is used. This improves file security and
8368 only older versions of MSIE (4.0 for example) need it.
8371 *) Honor the no-xxx Configure options when creating .DEF files.
8374 *) Add PKCS#10 attributes to field table: challengePassword,
8375 unstructuredName and unstructuredAddress. These are taken from
8376 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8377 international characters are used.
8379 More changes to X509_ATTRIBUTE code: allow the setting of types
8380 based on strings. Remove the 'loc' parameter when adding
8381 attributes because these will be a SET OF encoding which is sorted
8385 *) Initial changes to the 'req' utility to allow request generation
8386 automation. This will allow an application to just generate a template
8387 file containing all the field values and have req construct the
8390 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8391 used all over the place including certificate requests and PKCS#7
8392 structures. They are currently handled manually where necessary with
8393 some primitive wrappers for PKCS#7. The new functions behave in a
8394 manner analogous to the X509 extension functions: they allow
8395 attributes to be looked up by NID and added.
8397 Later something similar to the X509V3 code would be desirable to
8398 automatically handle the encoding, decoding and printing of the
8399 more complex types. The string types like challengePassword can
8400 be handled by the string table functions.
8402 Also modified the multi byte string table handling. Now there is
8403 a 'global mask' which masks out certain types. The table itself
8404 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8405 is useful when for example there is only one permissible type
8406 (as in countryName) and using the mask might result in no valid
8410 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8411 SSL_get_peer_finished to allow applications to obtain the latest
8412 Finished messages sent to the peer or expected from the peer,
8413 respectively. (SSL_get_peer_finished is usually the Finished message
8414 actually received from the peer, otherwise the protocol will be aborted.)
8416 As the Finished message are message digests of the complete handshake
8417 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8418 be used for external authentication procedures when the authentication
8419 provided by SSL/TLS is not desired or is not enough.
8422 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8423 the host supports BWX extension and if Compaq C is present on the
8424 $PATH. Just exploiting of the BWX extension results in 20-30%
8425 performance kick for some algorithms, e.g. DES and RC4 to mention
8426 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8430 *) Add support for MS "fast SGC". This is arguably a violation of the
8431 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8432 weak crypto and after checking the certificate is SGC a second one
8433 with strong crypto. MS SGC stops the first handshake after receiving
8434 the server certificate message and sends a second client hello. Since
8435 a server will typically do all the time consuming operations before
8436 expecting any further messages from the client (server key exchange
8437 is the most expensive) there is little difference between the two.
8439 To get OpenSSL to support MS SGC we have to permit a second client
8440 hello message after we have sent server done. In addition we have to
8441 reset the MAC if we do get this second client hello.
8444 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8445 if a DER encoded private key is RSA or DSA traditional format. Changed
8446 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8447 format DER encoded private key. Newer code should use PKCS#8 format which
8448 has the key type encoded in the ASN1 structure. Added DER private key
8449 support to pkcs8 application.
8452 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8453 ciphersuites has been selected (as required by the SSL 3/TLS 1
8454 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8455 is set, we interpret this as a request to violate the specification
8456 (the worst that can happen is a handshake failure, and 'correct'
8457 behaviour would result in a handshake failure anyway).
8460 *) In SSL_CTX_add_session, take into account that there might be multiple
8461 SSL_SESSION structures with the same session ID (e.g. when two threads
8462 concurrently obtain them from an external cache).
8463 The internal cache can handle only one SSL_SESSION with a given ID,
8464 so if there's a conflict, we now throw out the old one to achieve
8468 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8469 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8470 some routines that use cipher OIDs: some ciphers do not have OIDs
8471 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8475 *) Simplify the trust setting structure and code. Now we just have
8476 two sequences of OIDs for trusted and rejected settings. These will
8477 typically have values the same as the extended key usage extension
8478 and any application specific purposes.
8480 The trust checking code now has a default behaviour: it will just
8481 check for an object with the same NID as the passed id. Functions can
8482 be provided to override either the default behaviour or the behaviour
8483 for a given id. SSL client, server and email already have functions
8484 in place for compatibility: they check the NID and also return "trusted"
8485 if the certificate is self signed.
8488 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8489 traditional format into an EVP_PKEY structure.
8492 *) Add a password callback function PEM_cb() which either prompts for
8493 a password if usr_data is NULL or otherwise assumes it is a null
8494 terminated password. Allow passwords to be passed on command line
8495 environment or config files in a few more utilities.
8498 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8499 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8500 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8501 Update documentation.
8504 *) Support for ASN1 "NULL" type. This could be handled before by using
8505 ASN1_TYPE but there wasn't any function that would try to read a NULL
8506 and produce an error if it couldn't. For compatibility we also have
8507 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8508 don't allocate anything because they don't need to.
8511 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8513 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8515 *) Rebuild of the memory allocation routines used by OpenSSL code and
8516 possibly others as well. The purpose is to make an interface that
8517 provide hooks so anyone can build a separate set of allocation and
8518 deallocation routines to be used by OpenSSL, for example memory
8519 pool implementations, or something else, which was previously hard
8520 since Malloc(), Realloc() and Free() were defined as macros having
8521 the values malloc, realloc and free, respectively (except for Win32
8522 compilations). The same is provided for memory debugging code.
8523 OpenSSL already comes with functionality to find memory leaks, but
8524 this gives people a chance to debug other memory problems.
8526 With these changes, a new set of functions and macros have appeared:
8528 CRYPTO_set_mem_debug_functions() [F]
8529 CRYPTO_get_mem_debug_functions() [F]
8530 CRYPTO_dbg_set_options() [F]
8531 CRYPTO_dbg_get_options() [F]
8532 CRYPTO_malloc_debug_init() [M]
8534 The memory debug functions are NULL by default, unless the library
8535 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8536 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8537 gives the standard debugging functions that come with OpenSSL) or
8538 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8539 provided by the library user) must be used. When the standard
8540 debugging functions are used, CRYPTO_dbg_set_options can be used to
8541 request additional information:
8542 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8543 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8545 Also, things like CRYPTO_set_mem_functions will always give the
8546 expected result (the new set of functions is used for allocation
8547 and deallocation) at all times, regardless of platform and compiler
8550 To finish it up, some functions that were never use in any other
8551 way than through macros have a new API and new semantic:
8554 CRYPTO_dbg_realloc()
8557 All macros of value have retained their old syntax.
8558 [Richard Levitte and Bodo Moeller]
8560 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8561 ordering of SMIMECapabilities wasn't in "strength order" and there
8562 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8566 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8567 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8568 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8570 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8571 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8572 functionality to handle multipart/signed properly) and a utility
8573 called 'smime' to call all this stuff. This is based on code I
8574 originally wrote for Celo who have kindly allowed it to be
8575 included in OpenSSL.
8578 *) Add variants des_set_key_checked and des_set_key_unchecked of
8579 des_set_key (aka des_key_sched). Global variable des_check_key
8580 decides which of these is called by des_set_key; this way
8581 des_check_key behaves as it always did, but applications and
8582 the library itself, which was buggy for des_check_key == 1,
8583 have a cleaner way to pick the version they need.
8586 *) New function PKCS12_newpass() which changes the password of a
8590 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8591 dynamic mix. In both cases the ids can be used as an index into the
8592 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8593 functions so they accept a list of the field values and the
8594 application doesn't need to directly manipulate the X509_TRUST
8598 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8602 *) Modify the way the V3 extension code looks up extensions. This now
8603 works in a similar way to the object code: we have some "standard"
8604 extensions in a static table which is searched with OBJ_bsearch()
8605 and the application can add dynamic ones if needed. The file
8606 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8607 updated whenever a new extension is added to the core code and kept
8608 in ext_nid order. There is a simple program 'tabtest.c' which checks
8609 this. New extensions are not added too often so this file can readily
8610 be maintained manually.
8612 There are two big advantages in doing things this way. The extensions
8613 can be looked up immediately and no longer need to be "added" using
8614 X509V3_add_standard_extensions(): this function now does nothing.
8615 [Side note: I get *lots* of email saying the extension code doesn't
8616 work because people forget to call this function]
8617 Also no dynamic allocation is done unless new extensions are added:
8618 so if we don't add custom extensions there is no need to call
8619 X509V3_EXT_cleanup().
8622 *) Modify enc utility's salting as follows: make salting the default. Add a
8623 magic header, so unsalted files fail gracefully instead of just decrypting
8624 to garbage. This is because not salting is a big security hole, so people
8625 should be discouraged from doing it.
8628 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8629 digest to be passed on the command line but it only used this
8630 parameter when signing a certificate. Modified so all relevant
8631 operations are affected by the digest parameter including the
8632 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8633 DSA key was used because it didn't fix the digest.
8636 *) Initial certificate chain verify code. Currently tests the untrusted
8637 certificates for consistency with the verify purpose (which is set
8638 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8640 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8641 this is because it will reject chains with invalid extensions whereas
8642 every previous version of OpenSSL and SSLeay made no checks at all.
8644 Trust code: checks the root CA for the relevant trust settings. Trust
8645 settings have an initial value consistent with the verify purpose: e.g.
8646 if the verify purpose is for SSL client use it expects the CA to be
8647 trusted for SSL client use. However the default value can be changed to
8648 permit custom trust settings: one example of this would be to only trust
8649 certificates from a specific "secure" set of CAs.
8651 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8652 which should be used for version portability: especially since the
8653 verify structure is likely to change more often now.
8655 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8656 to set them. If not set then assume SSL clients will verify SSL servers
8659 Two new options to the verify program: -untrusted allows a set of
8660 untrusted certificates to be passed in and -purpose which sets the
8661 intended purpose of the certificate. If a purpose is set then the
8662 new chain verify code is used to check extension consistency.
8665 *) Support for the authority information access extension.
8668 *) Modify RSA and DSA PEM read routines to transparently handle
8669 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8670 public keys in a format compatible with certificate
8671 SubjectPublicKeyInfo structures. Unfortunately there were already
8672 functions called *_PublicKey_* which used various odd formats so
8673 these are retained for compatibility: however the DSA variants were
8674 never in a public release so they have been deleted. Changed dsa/rsa
8675 utilities to handle the new format: note no releases ever handled public
8676 keys so we should be OK.
8678 The primary motivation for this change is to avoid the same fiasco
8679 that dogs private keys: there are several incompatible private key
8680 formats some of which are standard and some OpenSSL specific and
8681 require various evil hacks to allow partial transparent handling and
8682 even then it doesn't work with DER formats. Given the option anything
8683 other than PKCS#8 should be dumped: but the other formats have to
8684 stay in the name of compatibility.
8686 With public keys and the benefit of hindsight one standard format
8687 is used which works with EVP_PKEY, RSA or DSA structures: though
8688 it clearly returns an error if you try to read the wrong kind of key.
8690 Added a -pubkey option to the 'x509' utility to output the public key.
8691 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8692 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8693 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8694 that do the same as the EVP_PKEY_assign_*() except they up the
8695 reference count of the added key (they don't "swallow" the
8699 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8700 CRLs would fail if the file contained no certificates or no CRLs:
8701 added a new function to read in both types and return the number
8702 read: this means that if none are read it will be an error. The
8703 DER versions of the certificate and CRL reader would always fail
8704 because it isn't possible to mix certificates and CRLs in DER format
8705 without choking one or the other routine. Changed this to just read
8706 a certificate: this is the best we can do. Also modified the code
8707 in apps/verify.c to take notice of return codes: it was previously
8708 attempting to read in certificates from NULL pointers and ignoring
8709 any errors: this is one reason why the cert and CRL reader seemed
8710 to work. It doesn't check return codes from the default certificate
8711 routines: these may well fail if the certificates aren't installed.
8714 *) Code to support otherName option in GeneralName.
8717 *) First update to verify code. Change the verify utility
8718 so it warns if it is passed a self signed certificate:
8719 for consistency with the normal behaviour. X509_verify
8720 has been modified to it will now verify a self signed
8721 certificate if *exactly* the same certificate appears
8722 in the store: it was previously impossible to trust a
8723 single self signed certificate. This means that:
8724 openssl verify ss.pem
8725 now gives a warning about a self signed certificate but
8726 openssl verify -CAfile ss.pem ss.pem
8730 *) For servers, store verify_result in SSL_SESSION data structure
8731 (and add it to external session representation).
8732 This is needed when client certificate verifications fails,
8733 but an application-provided verification callback (set by
8734 SSL_CTX_set_cert_verify_callback) allows accepting the session
8735 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8736 but returns 1): When the session is reused, we have to set
8737 ssl->verify_result to the appropriate error code to avoid
8739 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8741 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8742 case in PKCS7_dataInit() where the signed PKCS7 structure
8743 didn't contain any existing data because it was being created.
8744 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8746 *) Add a salt to the key derivation routines in enc.c. This
8747 forms the first 8 bytes of the encrypted file. Also add a
8748 -S option to allow a salt to be input on the command line.
8751 *) New function X509_cmp(). Oddly enough there wasn't a function
8752 to compare two certificates. We do this by working out the SHA1
8753 hash and comparing that. X509_cmp() will be needed by the trust
8757 *) SSL_get1_session() is like SSL_get_session(), but increments
8758 the reference count in the SSL_SESSION returned.
8759 [Geoff Thorpe <geoff@eu.c2.net>]
8761 *) Fix for 'req': it was adding a null to request attributes.
8762 Also change the X509_LOOKUP and X509_INFO code to handle
8763 certificate auxiliary information.
8766 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8770 *) Add the possibility to add extra information to the memory leak
8771 detecting output, to form tracebacks, showing from where each
8772 allocation was originated: CRYPTO_push_info("constant string") adds
8773 the string plus current file name and line number to a per-thread
8774 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8775 is like calling CYRPTO_pop_info() until the stack is empty.
8776 Also updated memory leak detection code to be multi-thread-safe.
8779 *) Add options -text and -noout to pkcs7 utility and delete the
8780 encryption options which never did anything. Update docs.
8783 *) Add options to some of the utilities to allow the pass phrase
8784 to be included on either the command line (not recommended on
8785 OSes like Unix) or read from the environment. Update the
8786 manpages and fix a few bugs.
8789 *) Add a few manpages for some of the openssl commands.
8792 *) Fix the -revoke option in ca. It was freeing up memory twice,
8793 leaking and not finding already revoked certificates.
8796 *) Extensive changes to support certificate auxiliary information.
8797 This involves the use of X509_CERT_AUX structure and X509_AUX
8798 functions. An X509_AUX function such as PEM_read_X509_AUX()
8799 can still read in a certificate file in the usual way but it
8800 will also read in any additional "auxiliary information". By
8801 doing things this way a fair degree of compatibility can be
8802 retained: existing certificates can have this information added
8803 using the new 'x509' options.
8805 Current auxiliary information includes an "alias" and some trust
8806 settings. The trust settings will ultimately be used in enhanced
8807 certificate chain verification routines: currently a certificate
8808 can only be trusted if it is self signed and then it is trusted
8812 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8813 The problem was that one of the replacement routines had not been working
8814 since SSLeay releases. For now the offending routine has been replaced
8815 with non-optimised assembler. Even so, this now gives around 95%
8816 performance improvement for 1024 bit RSA signs.
8819 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8820 handling. Most clients have the effective key size in bits equal to
8821 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8822 A few however don't do this and instead use the size of the decrypted key
8823 to determine the RC2 key length and the AlgorithmIdentifier to determine
8824 the effective key length. In this case the effective key length can still
8825 be 40 bits but the key length can be 168 bits for example. This is fixed
8826 by manually forcing an RC2 key into the EVP_PKEY structure because the
8827 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8828 the key length and effective key length are equal.
8831 *) Add a bunch of functions that should simplify the creation of
8832 X509_NAME structures. Now you should be able to do:
8833 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8834 and have it automatically work out the correct field type and fill in
8835 the structures. The more adventurous can try:
8836 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8837 and it will (hopefully) work out the correct multibyte encoding.
8840 *) Change the 'req' utility to use the new field handling and multibyte
8841 copy routines. Before the DN field creation was handled in an ad hoc
8842 way in req, ca, and x509 which was rather broken and didn't support
8843 BMPStrings or UTF8Strings. Since some software doesn't implement
8844 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8845 using the dirstring_type option. See the new comment in the default
8846 openssl.cnf for more info.
8849 *) Make crypto/rand/md_rand.c more robust:
8850 - Assure unique random numbers after fork().
8851 - Make sure that concurrent threads access the global counter and
8852 md serializably so that we never lose entropy in them
8853 or use exactly the same state in multiple threads.
8854 Access to the large state is not always serializable because
8855 the additional locking could be a performance killer, and
8856 md should be large enough anyway.
8859 *) New file apps/app_rand.c with commonly needed functionality
8860 for handling the random seed file.
8862 Use the random seed file in some applications that previously did not:
8864 dsaparam -genkey (which also ignored its '-rand' option),
8867 x509 (when signing).
8868 Except on systems with /dev/urandom, it is crucial to have a random
8869 seed file at least for key creation, DSA signing, and for DH exchanges;
8870 for RSA signatures we could do without one.
8872 gendh and gendsa (unlike genrsa) used to read only the first byte
8873 of each file listed in the '-rand' option. The function as previously
8874 found in genrsa is now in app_rand.c and is used by all programs
8875 that support '-rand'.
8878 *) In RAND_write_file, use mode 0600 for creating files;
8879 don't just chmod when it may be too late.
8882 *) Report an error from X509_STORE_load_locations
8883 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8886 *) New function ASN1_mbstring_copy() this copies a string in either
8887 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8888 into an ASN1_STRING type. A mask of permissible types is passed
8889 and it chooses the "minimal" type to use or an error if not type
8893 *) Add function equivalents to the various macros in asn1.h. The old
8894 macros are retained with an M_ prefix. Code inside the library can
8895 use the M_ macros. External code (including the openssl utility)
8896 should *NOT* in order to be "shared library friendly".
8899 *) Add various functions that can check a certificate's extensions
8900 to see if it usable for various purposes such as SSL client,
8901 server or S/MIME and CAs of these types. This is currently
8902 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8903 verification. Also added a -purpose flag to x509 utility to
8904 print out all the purposes.
8907 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8911 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8912 for, obtain and decode and extension and obtain its critical flag.
8913 This allows all the necessary extension code to be handled in a
8914 single function call.
8917 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8918 platforms. See crypto/rc4/rc4_enc.c for further details.
8921 *) New -noout option to asn1parse. This causes no output to be produced
8922 its main use is when combined with -strparse and -out to extract data
8923 from a file (which may not be in ASN.1 format).
8926 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8927 when producing the local key id.
8928 [Richard Levitte <levitte@stacken.kth.se>]
8930 *) New option -dhparam in s_server. This allows a DH parameter file to be
8931 stated explicitly. If it is not stated then it tries the first server
8932 certificate file. The previous behaviour hard coded the filename
8936 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8937 a public key to be input or output. For example:
8938 openssl rsa -in key.pem -pubout -out pubkey.pem
8939 Also added necessary DSA public key functions to handle this.
8942 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8943 in the message. This was handled by allowing
8944 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8945 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8947 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8948 to the end of the strings whereas this didn't. This would cause problems
8949 if strings read with d2i_ASN1_bytes() were later modified.
8950 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8952 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8953 data and it contains EOF it will end up returning an error. This is
8954 caused by input 46 bytes long. The cause is due to the way base64
8955 BIOs find the start of base64 encoded data. They do this by trying a
8956 trial decode on each line until they find one that works. When they
8957 do a flag is set and it starts again knowing it can pass all the
8958 data directly through the decoder. Unfortunately it doesn't reset
8959 the context it uses. This means that if EOF is reached an attempt
8960 is made to pass two EOFs through the context and this causes the
8961 resulting error. This can also cause other problems as well. As is
8962 usual with these problems it takes *ages* to find and the fix is
8963 trivial: move one line.
8964 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8966 *) Ugly workaround to get s_client and s_server working under Windows. The
8967 old code wouldn't work because it needed to select() on sockets and the
8968 tty (for keypresses and to see if data could be written). Win32 only
8969 supports select() on sockets so we select() with a 1s timeout on the
8970 sockets and then see if any characters are waiting to be read, if none
8971 are present then we retry, we also assume we can always write data to
8972 the tty. This isn't nice because the code then blocks until we've
8973 received a complete line of data and it is effectively polling the
8974 keyboard at 1s intervals: however it's quite a bit better than not
8975 working at all :-) A dedicated Windows application might handle this
8976 with an event loop for example.
8979 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8980 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8981 will be called when RSA_sign() and RSA_verify() are used. This is useful
8982 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8983 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8984 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8985 This necessitated the support of an extra signature type NID_md5_sha1
8986 for SSL signatures and modifications to the SSL library to use it instead
8987 of calling RSA_public_decrypt() and RSA_private_encrypt().
8990 *) Add new -verify -CAfile and -CApath options to the crl program, these
8991 will lookup a CRL issuers certificate and verify the signature in a
8992 similar way to the verify program. Tidy up the crl program so it
8993 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8994 less strict. It will now permit CRL extensions even if it is not
8995 a V2 CRL: this will allow it to tolerate some broken CRLs.
8998 *) Initialize all non-automatic variables each time one of the openssl
8999 sub-programs is started (this is necessary as they may be started
9000 multiple times from the "OpenSSL>" prompt).
9001 [Lennart Bang, Bodo Moeller]
9003 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9004 removing all other RSA functionality (this is what NO_RSA does). This
9005 is so (for example) those in the US can disable those operations covered
9006 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9010 *) Non-copying interface to BIO pairs.
9011 (still largely untested)
9014 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9015 ASCII string. This was handled independently in various places before.
9018 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9019 UTF8 strings a character at a time.
9022 *) Use client_version from client hello to select the protocol
9023 (s23_srvr.c) and for RSA client key exchange verification
9024 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9027 *) Add various utility functions to handle SPKACs, these were previously
9028 handled by poking round in the structure internals. Added new function
9029 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9030 print, verify and generate SPKACs. Based on an original idea from
9031 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9034 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9037 *) Allow the config file extension section to be overwritten on the
9038 command line. Based on an original idea from Massimiliano Pala
9039 <madwolf@comune.modena.it>. The new option is called -extensions
9040 and can be applied to ca, req and x509. Also -reqexts to override
9041 the request extensions in req and -crlexts to override the crl extensions
9045 *) Add new feature to the SPKAC handling in ca. Now you can include
9046 the same field multiple times by preceding it by "XXXX." for example:
9049 this is the same syntax as used in the req config file.
9052 *) Allow certificate extensions to be added to certificate requests. These
9053 are specified in a 'req_extensions' option of the req section of the
9054 config file. They can be printed out with the -text option to req but
9055 are otherwise ignored at present.
9058 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9059 data read consists of only the final block it would not decrypted because
9060 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9061 A misplaced 'break' also meant the decrypted final block might not be
9062 copied until the next read.
9065 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9066 a few extra parameters to the DH structure: these will be useful if
9067 for example we want the value of 'q' or implement X9.42 DH.
9070 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9071 provides hooks that allow the default DSA functions or functions on a
9072 "per key" basis to be replaced. This allows hardware acceleration and
9073 hardware key storage to be handled without major modification to the
9074 library. Also added low level modexp hooks and CRYPTO_EX structure and
9075 associated functions.
9078 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9079 as "read only": it can't be written to and the buffer it points to will
9080 not be freed. Reading from a read only BIO is much more efficient than
9081 a normal memory BIO. This was added because there are several times when
9082 an area of memory needs to be read from a BIO. The previous method was
9083 to create a memory BIO and write the data to it, this results in two
9084 copies of the data and an O(n^2) reading algorithm. There is a new
9085 function BIO_new_mem_buf() which creates a read only memory BIO from
9086 an area of memory. Also modified the PKCS#7 routines to use read only
9090 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9091 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9092 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9093 but a retry condition occured while trying to read the rest.
9096 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9097 NID_pkcs7_encrypted by default: this was wrong since this should almost
9098 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9099 the encrypted data type: this is a more sensible place to put it and it
9100 allows the PKCS#12 code to be tidied up that duplicated this
9104 *) Changed obj_dat.pl script so it takes its input and output files on
9105 the command line. This should avoid shell escape redirection problems
9109 *) Initial support for certificate extension requests, these are included
9110 in things like Xenroll certificate requests. Included functions to allow
9111 extensions to be obtained and added.
9114 *) -crlf option to s_client and s_server for sending newlines as
9115 CRLF (as required by many protocols).
9118 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9120 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9121 [Ralf S. Engelschall]
9123 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9124 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9126 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9130 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9131 DH parameters/keys (q is lost during that conversion, but the resulting
9132 DH parameters contain its length).
9134 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9135 much faster than DH_generate_parameters (which creates parameters
9136 where p = 2*q + 1), and also the smaller q makes DH computations
9137 much more efficient (160-bit exponentiation instead of 1024-bit
9138 exponentiation); so this provides a convenient way to support DHE
9139 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9140 utter importance to use
9141 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9143 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9144 when such DH parameters are used, because otherwise small subgroup
9145 attacks may become possible!
9148 *) Avoid memory leak in i2d_DHparams.
9151 *) Allow the -k option to be used more than once in the enc program:
9152 this allows the same encrypted message to be read by multiple recipients.
9155 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9156 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9157 it will always use the numerical form of the OID, even if it has a short
9161 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9162 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9163 otherwise bn_mod_exp was called. In the case of hardware keys for example
9164 no private key components need be present and it might store extra data
9165 in the RSA structure, which cannot be accessed from bn_mod_exp.
9166 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9167 private key operations.
9170 *) Added support for SPARC Linux.
9173 *) pem_password_cb function type incompatibly changed from
9174 typedef int pem_password_cb(char *buf, int size, int rwflag);
9176 ....(char *buf, int size, int rwflag, void *userdata);
9177 so that applications can pass data to their callbacks:
9178 The PEM[_ASN1]_{read,write}... functions and macros now take an
9179 additional void * argument, which is just handed through whenever
9180 the password callback is called.
9181 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9183 New function SSL_CTX_set_default_passwd_cb_userdata.
9185 Compatibility note: As many C implementations push function arguments
9186 onto the stack in reverse order, the new library version is likely to
9187 interoperate with programs that have been compiled with the old
9188 pem_password_cb definition (PEM_whatever takes some data that
9189 happens to be on the stack as its last argument, and the callback
9190 just ignores this garbage); but there is no guarantee whatsoever that
9193 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9194 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9195 problems not only on Windows, but also on some Unix platforms.
9196 To avoid problematic command lines, these definitions are now in an
9197 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9198 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9201 *) MIPS III/IV assembler module is reimplemented.
9204 *) More DES library cleanups: remove references to srand/rand and
9205 delete an unused file.
9208 *) Add support for the the free Netwide assembler (NASM) under Win32,
9209 since not many people have MASM (ml) and it can be hard to obtain.
9210 This is currently experimental but it seems to work OK and pass all
9211 the tests. Check out INSTALL.W32 for info.
9214 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9215 without temporary keys kept an extra copy of the server key,
9216 and connections with temporary keys did not free everything in case
9220 *) New function RSA_check_key and new openssl rsa option -check
9221 for verifying the consistency of RSA keys.
9222 [Ulf Moeller, Bodo Moeller]
9224 *) Various changes to make Win32 compile work:
9225 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9226 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9227 comparison" warnings.
9228 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9231 *) Add a debugging option to PKCS#5 v2 key generation function: when
9232 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9233 derived keys are printed to stderr.
9236 *) Copy the flags in ASN1_STRING_dup().
9237 [Roman E. Pavlov <pre@mo.msk.ru>]
9239 *) The x509 application mishandled signing requests containing DSA
9240 keys when the signing key was also DSA and the parameters didn't match.
9242 It was supposed to omit the parameters when they matched the signing key:
9243 the verifying software was then supposed to automatically use the CA's
9244 parameters if they were absent from the end user certificate.
9246 Omitting parameters is no longer recommended. The test was also
9247 the wrong way round! This was probably due to unusual behaviour in
9248 EVP_cmp_parameters() which returns 1 if the parameters match.
9249 This meant that parameters were omitted when they *didn't* match and
9250 the certificate was useless. Certificates signed with 'ca' didn't have
9252 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9254 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9255 The interface is as follows:
9256 Applications can use
9257 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9258 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9259 "off" is now the default.
9260 The library internally uses
9261 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9262 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9263 to disable memory-checking temporarily.
9265 Some inconsistent states that previously were possible (and were
9266 even the default) are now avoided.
9268 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9269 with each memory chunk allocated; this is occasionally more helpful
9270 than just having a counter.
9272 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9274 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9278 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9279 which largely parallels "options", but is for changing API behaviour,
9280 whereas "options" are about protocol behaviour.
9281 Initial "mode" flags are:
9283 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9284 a single record has been written.
9285 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9286 retries use the same buffer location.
9287 (But all of the contents must be
9291 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9294 *) Fix problems with no-hmac etc.
9295 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9297 *) New functions RSA_get_default_method(), RSA_set_method() and
9298 RSA_get_method(). These allows replacement of RSA_METHODs without having
9299 to mess around with the internals of an RSA structure.
9302 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9303 Also really enable memory leak checks in openssl.c and in some
9305 [Chad C. Mulligan, Bodo Moeller]
9307 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9308 up the length of negative integers. This has now been simplified to just
9309 store the length when it is first determined and use it later, rather
9310 than trying to keep track of where data is copied and updating it to
9312 [Steve Henson, reported by Brien Wheeler
9313 <bwheeler@authentica-security.com>]
9315 *) Add a new function PKCS7_signatureVerify. This allows the verification
9316 of a PKCS#7 signature but with the signing certificate passed to the
9317 function itself. This contrasts with PKCS7_dataVerify which assumes the
9318 certificate is present in the PKCS#7 structure. This isn't always the
9319 case: certificates can be omitted from a PKCS#7 structure and be
9320 distributed by "out of band" means (such as a certificate database).
9323 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9324 function prototypes in pem.h, also change util/mkdef.pl to add the
9325 necessary function names.
9328 *) mk1mf.pl (used by Windows builds) did not properly read the
9329 options set by Configure in the top level Makefile, and Configure
9330 was not even able to write more than one option correctly.
9331 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9334 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9335 file to be loaded from a BIO or FILE pointer. The BIO version will
9336 for example allow memory BIOs to contain config info.
9339 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9340 Whoever hopes to achieve shared-library compatibility across versions
9341 must use this, not the compile-time macro.
9342 (Exercise 0.9.4: Which is the minimum library version required by
9344 Note: All this applies only to multi-threaded programs, others don't
9348 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9349 through a BIO pair triggered the default case, i.e.
9350 SSLerr(...,SSL_R_UNKNOWN_STATE).
9353 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9354 can use the SSL library even if none of the specific BIOs is
9358 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9359 for the encoded length.
9360 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9362 *) Add initial documentation of the X509V3 functions.
9365 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9366 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9367 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9368 secure PKCS#8 private key format with a high iteration count.
9371 *) Fix determination of Perl interpreter: A perl or perl5
9372 _directory_ in $PATH was also accepted as the interpreter.
9373 [Ralf S. Engelschall]
9375 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9376 wrong with it but it was very old and did things like calling
9377 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9381 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9382 to use the new extension code.
9385 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9386 with macros. This should make it easier to change their form, add extra
9387 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9391 *) Add to configuration table a new entry that can specify an alternative
9392 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9393 according to Mark Crispin <MRC@Panda.COM>.
9397 *) DES CBC did not update the IV. Weird.
9400 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9401 Changing the behaviour of the former might break existing programs --
9402 where IV updating is needed, des_ncbc_encrypt can be used.
9405 *) When bntest is run from "make test" it drives bc to check its
9406 calculations, as well as internally checking them. If an internal check
9407 fails, it needs to cause bc to give a non-zero result or make test carries
9408 on without noticing the failure. Fixed.
9411 *) DES library cleanups.
9414 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9415 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9416 ciphers. NOTE: although the key derivation function has been verified
9417 against some published test vectors it has not been extensively tested
9418 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9422 *) Instead of "mkdir -p", which is not fully portable, use new
9423 Perl script "util/mkdir-p.pl".
9426 *) Rewrite the way password based encryption (PBE) is handled. It used to
9427 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9428 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9429 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9430 the 'parameter' field of the AlgorithmIdentifier is passed to the
9431 underlying key generation function so it must do its own ASN1 parsing.
9432 This has also changed the EVP_PBE_CipherInit() function which now has a
9433 'parameter' argument instead of literal salt and iteration count values
9434 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9437 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9438 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9439 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9440 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9441 value was just used as a "magic string" and not used directly its
9442 value doesn't matter.
9445 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9449 *) "linux-sparc64" configuration (ultrapenguin).
9450 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9451 "linux-sparc" configuration.
9452 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9454 *) config now generates no-xxx options for missing ciphers.
9457 *) Support the EBCDIC character set (work in progress).
9458 File ebcdic.c not yet included because it has a different license.
9459 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9461 *) Support BS2000/OSD-POSIX.
9462 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9464 *) Make callbacks for key generation use void * instead of char *.
9467 *) Make S/MIME samples compile (not yet tested).
9470 *) Additional typesafe stacks.
9473 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9477 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9479 *) New configuration variant "sco5-gcc".
9481 *) Updated some demos.
9482 [Sean O Riordain, Wade Scholine]
9484 *) Add missing BIO_free at exit of pkcs12 application.
9487 *) Fix memory leak in conf.c.
9490 *) Updates for Win32 to assembler version of MD5.
9493 *) Set #! path to perl in apps/der_chop to where we found it
9494 instead of using a fixed path.
9497 *) SHA library changes for irix64-mips4-cc.
9500 *) Improvements for VMS support.
9504 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9506 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9507 This also avoids the problems with SC4.2 and unpatched SC5.
9508 [Andy Polyakov <appro@fy.chalmers.se>]
9510 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9511 These are required because of the typesafe stack would otherwise break
9512 existing code. If old code used a structure member which used to be STACK
9513 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9514 sk_num or sk_value it would produce an error because the num, data members
9515 are not present in STACK_OF. Now it just produces a warning. sk_set
9516 replaces the old method of assigning a value to sk_value
9517 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9518 that does this will no longer work (and should use sk_set instead) but
9519 this could be regarded as a "questionable" behaviour anyway.
9522 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9523 correctly handle encrypted S/MIME data.
9526 *) Change type of various DES function arguments from des_cblock
9527 (which means, in function argument declarations, pointer to char)
9528 to des_cblock * (meaning pointer to array with 8 char elements),
9529 which allows the compiler to do more typechecking; it was like
9530 that back in SSLeay, but with lots of ugly casts.
9532 Introduce new type const_des_cblock.
9535 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9536 problems: find RecipientInfo structure that matches recipient certificate
9537 and initialise the ASN1 structures properly based on passed cipher.
9540 *) Belatedly make the BN tests actually check the results.
9543 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9544 to and from BNs: it was completely broken. New compilation option
9545 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9546 key elements as negative integers.
9549 *) Reorganize and speed up MD5.
9550 [Andy Polyakov <appro@fy.chalmers.se>]
9553 [Richard Levitte <richard@levitte.org>]
9555 *) New option -out to asn1parse to allow the parsed structure to be
9556 output to a file. This is most useful when combined with the -strparse
9557 option to examine the output of things like OCTET STRINGS.
9560 *) Make SSL library a little more fool-proof by not requiring any longer
9561 that SSL_set_{accept,connect}_state be called before
9562 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9563 in many applications because usually everything *appeared* to work as
9564 intended anyway -- now it really works as intended).
9567 *) Move openssl.cnf out of lib/.
9570 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9571 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9572 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9573 [Ralf S. Engelschall]
9575 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9576 handle PKCS#7 enveloped data properly.
9577 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9579 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9580 copying pointers. The cert_st handling is changed by this in
9581 various ways (and thus what used to be known as ctx->default_cert
9582 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9583 any longer when s->cert does not give us what we need).
9584 ssl_cert_instantiate becomes obsolete by this change.
9585 As soon as we've got the new code right (possibly it already is?),
9586 we have solved a couple of bugs of the earlier code where s->cert
9587 was used as if it could not have been shared with other SSL structures.
9589 Note that using the SSL API in certain dirty ways now will result
9590 in different behaviour than observed with earlier library versions:
9591 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9592 does not influence s as it used to.
9594 In order to clean up things more thoroughly, inside SSL_SESSION
9595 we don't use CERT any longer, but a new structure SESS_CERT
9596 that holds per-session data (if available); currently, this is
9597 the peer's certificate chain and, for clients, the server's certificate
9598 and temporary key. CERT holds only those values that can have
9599 meaningful defaults in an SSL_CTX.
9602 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9603 from the internal representation. Various PKCS#7 fixes: remove some
9604 evil casts and set the enc_dig_alg field properly based on the signing
9608 *) Allow PKCS#12 password to be set from the command line or the
9609 environment. Let 'ca' get its config file name from the environment
9610 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9614 *) Allow certificate policies extension to use an IA5STRING for the
9615 organization field. This is contrary to the PKIX definition but
9616 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9620 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9621 without disallowing inline assembler and the like for non-pedantic builds.
9624 *) Support Borland C++ builder.
9625 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9630 *) SHA-1 cleanups and performance enhancements.
9631 [Andy Polyakov <appro@fy.chalmers.se>]
9633 *) Sparc v8plus assembler for the bignum library.
9634 [Andy Polyakov <appro@fy.chalmers.se>]
9636 *) Accept any -xxx and +xxx compiler options in Configure.
9639 *) Update HPUX configuration.
9642 *) Add missing sk_<type>_unshift() function to safestack.h
9643 [Ralf S. Engelschall]
9645 *) New function SSL_CTX_use_certificate_chain_file that sets the
9646 "extra_cert"s in addition to the certificate. (This makes sense
9647 only for "PEM" format files, as chains as a whole are not
9651 *) Support verify_depth from the SSL API.
9652 x509_vfy.c had what can be considered an off-by-one-error:
9653 Its depth (which was not part of the external interface)
9654 was actually counting the number of certificates in a chain;
9655 now it really counts the depth.
9658 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9659 instead of X509err, which often resulted in confusing error
9660 messages since the error codes are not globally unique
9661 (e.g. an alleged error in ssl3_accept when a certificate
9662 didn't match the private key).
9664 *) New function SSL_CTX_set_session_id_context that allows to set a default
9665 value (so that you don't need SSL_set_session_id_context for each
9666 connection using the SSL_CTX).
9669 *) OAEP decoding bug fix.
9672 *) Support INSTALL_PREFIX for package builders, as proposed by
9676 *) New Configure options "threads" and "no-threads". For systems
9677 where the proper compiler options are known (currently Solaris
9678 and Linux), "threads" is the default.
9681 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9684 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9685 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9686 such as /usr/local/bin.
9689 *) "make linux-shared" to build shared libraries.
9690 [Niels Poppe <niels@netbox.org>]
9692 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9695 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9696 extension adding in x509 utility.
9699 *) Remove NOPROTO sections and error code comments.
9702 *) Partial rewrite of the DEF file generator to now parse the ANSI
9706 *) New Configure options --prefix=DIR and --openssldir=DIR.
9709 *) Complete rewrite of the error code script(s). It is all now handled
9710 by one script at the top level which handles error code gathering,
9711 header rewriting and C source file generation. It should be much better
9712 than the old method: it now uses a modified version of Ulf's parser to
9713 read the ANSI prototypes in all header files (thus the old K&R definitions
9714 aren't needed for error creation any more) and do a better job of
9715 translating function codes into names. The old 'ASN1 error code imbedded
9716 in a comment' is no longer necessary and it doesn't use .err files which
9717 have now been deleted. Also the error code call doesn't have to appear all
9718 on one line (which resulted in some large lines...).
9721 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9724 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9725 0 (which usually indicates a closed connection), but continue reading.
9728 *) Fix some race conditions.
9731 *) Add support for CRL distribution points extension. Add Certificate
9732 Policies and CRL distribution points documentation.
9735 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9738 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9739 8 of keying material. Merlin has also confirmed interop with this fix
9740 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9741 [Merlin Hughes <merlin@baltimore.ie>]
9743 *) Fix lots of warnings.
9744 [Richard Levitte <levitte@stacken.kth.se>]
9746 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9747 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9748 [Richard Levitte <levitte@stacken.kth.se>]
9750 *) Fix problems with sizeof(long) == 8.
9751 [Andy Polyakov <appro@fy.chalmers.se>]
9753 *) Change functions to ANSI C.
9756 *) Fix typos in error codes.
9757 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9759 *) Remove defunct assembler files from Configure.
9762 *) SPARC v8 assembler BIGNUM implementation.
9763 [Andy Polyakov <appro@fy.chalmers.se>]
9765 *) Support for Certificate Policies extension: both print and set.
9766 Various additions to support the r2i method this uses.
9769 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9770 return a const string when you are expecting an allocated buffer.
9773 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9774 types DirectoryString and DisplayText.
9777 *) Add code to allow r2i extensions to access the configuration database,
9778 add an LHASH database driver and add several ctx helper functions.
9781 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9782 fail when they extended the size of a BIGNUM.
9785 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9786 support typesafe stack.
9789 *) Fix typo in SSL_[gs]et_options().
9790 [Nils Frostberg <nils@medcom.se>]
9792 *) Delete various functions and files that belonged to the (now obsolete)
9793 old X509V3 handling code.
9796 *) New Configure option "rsaref".
9799 *) Don't auto-generate pem.h.
9802 *) Introduce type-safe ASN.1 SETs.
9805 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9806 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9808 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9809 that links with OpenSSL (well at least cause lots of warnings), but fear
9810 not: the conversion is trivial, and it eliminates loads of evil casts. A
9811 few STACKed things have been converted already. Feel free to convert more.
9812 In the fullness of time, I'll do away with the STACK type altogether.
9815 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9816 specified in <certfile> by updating the entry in the index.txt file.
9817 This way one no longer has to edit the index.txt file manually for
9818 revoking a certificate. The -revoke option does the gory details now.
9819 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9821 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9822 `-text' option at all and this way the `-noout -text' combination was
9823 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9824 [Ralf S. Engelschall]
9826 *) Make sure a corresponding plain text error message exists for the
9827 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9828 verify callback function determined that a certificate was revoked.
9829 [Ralf S. Engelschall]
9831 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9832 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9833 all available cipers including rc5, which was forgotten until now.
9834 In order to let the testing shell script know which algorithms
9835 are available, a new (up to now undocumented) command
9836 "openssl list-cipher-commands" is used.
9839 *) Bugfix: s_client occasionally would sleep in select() when
9840 it should have checked SSL_pending() first.
9843 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9844 the raw DSA values prior to ASN.1 encoding.
9847 *) Tweaks to Configure
9848 [Niels Poppe <niels@netbox.org>]
9850 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9854 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9857 *) New config option to avoid instructions that are illegal on the 80386.
9858 The default code is faster, but requires at least a 486.
9861 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9862 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9863 same as SSL2_VERSION anyway.
9866 *) New "-showcerts" option for s_client.
9869 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9870 application. Various cleanups and fixes.
9873 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9874 modify error routines to work internally. Add error codes and PBE init
9875 to library startup routines.
9878 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9879 packing functions to asn1 and evp. Changed function names and error
9880 codes along the way.
9883 *) PKCS12 integration: and so it begins... First of several patches to
9884 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9885 objects to objects.h
9888 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9889 and display support for Thawte strong extranet extension.
9892 *) Add LinuxPPC support.
9893 [Jeff Dubrule <igor@pobox.org>]
9895 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9896 bn_div_words in alpha.s.
9897 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9899 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9900 OAEP isn't supported when OpenSSL is built with RSAref.
9901 [Ulf Moeller <ulf@fitug.de>]
9903 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9904 so they no longer are missing under -DNOPROTO.
9905 [Soren S. Jorvang <soren@t.dk>]
9908 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9910 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9911 doesn't work when the session is reused. Coming soon!
9914 *) Fix a security hole, that allows sessions to be reused in the wrong
9915 context thus bypassing client cert protection! All software that uses
9916 client certs and session caches in multiple contexts NEEDS PATCHING to
9917 allow session reuse! A fuller solution is in the works.
9918 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9920 *) Some more source tree cleanups (removed obsolete files
9921 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9922 permission on "config" script to be executable) and a fix for the INSTALL
9924 [Ulf Moeller <ulf@fitug.de>]
9926 *) Remove some legacy and erroneous uses of malloc, free instead of
9928 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9930 *) Make rsa_oaep_test return non-zero on error.
9931 [Ulf Moeller <ulf@fitug.de>]
9933 *) Add support for native Solaris shared libraries. Configure
9934 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9935 if someone would make that last step automatic.
9936 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9938 *) ctx_size was not built with the right compiler during "make links". Fixed.
9941 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9942 except NULL ciphers". This means the default cipher list will no longer
9943 enable NULL ciphers. They need to be specifically enabled e.g. with
9944 the string "DEFAULT:eNULL".
9947 *) Fix to RSA private encryption routines: if p < q then it would
9948 occasionally produce an invalid result. This will only happen with
9949 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9952 *) Be less restrictive and allow also `perl util/perlpath.pl
9953 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9954 because this way one can also use an interpreter named `perl5' (which is
9955 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9956 installed as `perl').
9957 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9959 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9960 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9962 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9963 advapi32.lib to Win32 build and change the pem test comparision
9964 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9965 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9966 and crypto/des/ede_cbcm_enc.c.
9969 *) DES quad checksum was broken on big-endian architectures. Fixed.
9972 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9973 Win32 test batch file so it (might) work again. The Win32 test batch file
9974 is horrible: I feel ill....
9977 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9978 in e_os.h. Audit of header files to check ANSI and non ANSI
9979 sections: 10 functions were absent from non ANSI section and not exported
9980 from Windows DLLs. Fixed up libeay.num for new functions.
9983 *) Make `openssl version' output lines consistent.
9984 [Ralf S. Engelschall]
9986 *) Fix Win32 symbol export lists for BIO functions: Added
9987 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9988 to ms/libeay{16,32}.def.
9989 [Ralf S. Engelschall]
9991 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9992 fine under Unix and passes some trivial tests I've now added. But the
9993 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9994 added to make sure no one expects that this stuff really works in the
9995 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9996 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9998 [Ralf S. Engelschall]
10000 *) Fix the generation of two part addresses in perl.
10001 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10003 *) Add config entry for Linux on MIPS.
10004 [John Tobey <jtobey@channel1.com>]
10006 *) Make links whenever Configure is run, unless we are on Windoze.
10009 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10010 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10014 *) Add a useful kludge to allow package maintainers to specify compiler and
10015 other platforms details on the command line without having to patch the
10016 Configure script everytime: One now can use ``perl Configure
10017 <id>:<details>'', i.e. platform ids are allowed to have details appended
10018 to them (seperated by colons). This is treated as there would be a static
10019 pre-configured entry in Configure's %table under key <id> with value
10020 <details> and ``perl Configure <id>'' is called. So, when you want to
10021 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10022 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10023 now, which overrides the FreeBSD-elf entry on-the-fly.
10024 [Ralf S. Engelschall]
10026 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10029 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10030 on the `perl Configure ...' command line. This way one can compile
10031 OpenSSL libraries with Position Independent Code (PIC) which is needed
10032 for linking it into DSOs.
10033 [Ralf S. Engelschall]
10035 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10039 *) Cleaned up the LICENSE document: The official contact for any license
10040 questions now is the OpenSSL core team under openssl-core@openssl.org.
10041 And add a paragraph about the dual-license situation to make sure people
10042 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10043 to the OpenSSL toolkit.
10044 [Ralf S. Engelschall]
10046 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10047 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10048 Additonally cleaned up the `make links' target: Remove unnecessary
10049 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10050 to speed processing and no longer clutter the display with confusing
10051 stuff. Instead only the actually done links are displayed.
10052 [Ralf S. Engelschall]
10054 *) Permit null encryption ciphersuites, used for authentication only. It used
10055 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10056 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10060 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10061 signed attributes when verifying signatures (this would break them),
10062 the detached data encoding was wrong and public keys obtained using
10063 X509_get_pubkey() weren't freed.
10066 *) Add text documentation for the BUFFER functions. Also added a work around
10067 to a Win95 console bug. This was triggered by the password read stuff: the
10068 last character typed gets carried over to the next fread(). If you were
10069 generating a new cert request using 'req' for example then the last
10070 character of the passphrase would be CR which would then enter the first
10074 *) Added the new `Includes OpenSSL Cryptography Software' button as
10075 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10076 button and can be used by applications based on OpenSSL to show the
10077 relationship to the OpenSSL project.
10078 [Ralf S. Engelschall]
10080 *) Remove confusing variables in function signatures in files
10081 ssl/ssl_lib.c and ssl/ssl.h.
10082 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10084 *) Don't install bss_file.c under PREFIX/include/
10085 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10087 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10088 functions that return function pointers and has support for NT specific
10089 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10090 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10091 unsigned to signed types: this was killing the Win32 compile.
10094 *) Add new certificate file to stack functions,
10095 SSL_add_dir_cert_subjects_to_stack() and
10096 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10097 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10098 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10099 This means that Apache-SSL and similar packages don't have to mess around
10100 to add as many CAs as they want to the preferred list.
10103 *) Experiment with doxygen documentation. Currently only partially applied to
10105 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10106 openssl.doxy as the configuration file.
10109 *) Get rid of remaining C++-style comments which strict C compilers hate.
10110 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10112 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10113 compiled in by default: it has problems with large keys.
10116 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10117 DH private keys and/or callback functions which directly correspond to
10118 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10119 is needed for applications which have to configure certificates on a
10120 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10122 For the RSA certificate situation is makes no difference, but
10123 for the DSA certificate situation this fixes the "no shared cipher"
10124 problem where the OpenSSL cipher selection procedure failed because the
10125 temporary keys were not overtaken from the context and the API provided
10126 no way to reconfigure them.
10127 The new functions now let applications reconfigure the stuff and they
10128 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10129 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10130 non-public-API function ssl_cert_instantiate() is used as a helper
10131 function and also to reduce code redundancy inside ssl_rsa.c.
10132 [Ralf S. Engelschall]
10134 *) Move s_server -dcert and -dkey options out of the undocumented feature
10135 area because they are useful for the DSA situation and should be
10136 recognized by the users.
10137 [Ralf S. Engelschall]
10139 *) Fix the cipher decision scheme for export ciphers: the export bits are
10140 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10141 SSL_EXP_MASK. So, the original variable has to be used instead of the
10142 already masked variable.
10143 [Richard Levitte <levitte@stacken.kth.se>]
10145 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10146 [Richard Levitte <levitte@stacken.kth.se>]
10148 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10149 from `int' to `unsigned int' because it's a length and initialized by
10150 EVP_DigestFinal() which expects an `unsigned int *'.
10151 [Richard Levitte <levitte@stacken.kth.se>]
10153 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10154 script. Instead use the usual Shell->Perl transition trick.
10155 [Ralf S. Engelschall]
10157 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10158 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10159 -noout -modulus' as it's already the case for `openssl rsa -noout
10160 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10161 currently the public key is printed (a decision which was already done by
10162 `openssl dsa -modulus' in the past) which serves a similar purpose.
10163 Additionally the NO_RSA no longer completely removes the whole -modulus
10164 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10166 [Ralf S. Engelschall]
10168 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10169 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10170 [Arne Ansper <arne@ats.cyber.ee>]
10172 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10173 to be added. Now both 'req' and 'ca' can use new objects defined in the
10177 *) Add cool BIO that does syslog (or event log on NT).
10178 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10180 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10181 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10182 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10183 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10186 *) Add preliminary config info for new extension code.
10189 *) Make RSA_NO_PADDING really use no padding.
10190 [Ulf Moeller <ulf@fitug.de>]
10192 *) Generate errors when private/public key check is done.
10195 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10196 for some CRL extensions and new objects added.
10199 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10200 key usage extension and fuller support for authority key id.
10203 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10204 padding method for RSA, which is recommended for new applications in PKCS
10205 #1 v2.0 (RFC 2437, October 1998).
10206 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10207 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10208 against Bleichbacher's attack on RSA.
10209 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10212 *) Updates to the new SSL compression code
10213 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10215 *) Fix so that the version number in the master secret, when passed
10216 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10217 (because the server will not accept higher), that the version number
10218 is 0x03,0x01, not 0x03,0x00
10219 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10221 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10222 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10223 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10226 *) Support for RAW extensions where an arbitrary extension can be
10227 created by including its DER encoding. See apps/openssl.cnf for
10231 *) Make sure latest Perl versions don't interpret some generated C array
10232 code as Perl array code in the crypto/err/err_genc.pl script.
10233 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10235 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10236 not many people have the assembler. Various Win32 compilation fixes and
10237 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10238 build instructions.
10241 *) Modify configure script 'Configure' to automatically create crypto/date.h
10242 file under Win32 and also build pem.h from pem.org. New script
10243 util/mkfiles.pl to create the MINFO file on environments that can't do a
10244 'make files': perl util/mkfiles.pl >MINFO should work.
10247 *) Major rework of DES function declarations, in the pursuit of correctness
10248 and purity. As a result, many evil casts evaporated, and some weirdness,
10249 too. You may find this causes warnings in your code. Zapping your evil
10250 casts will probably fix them. Mostly.
10253 *) Fix for a typo in asn1.h. Bug fix to object creation script
10254 obj_dat.pl. It considered a zero in an object definition to mean
10255 "end of object": none of the objects in objects.h have any zeros
10256 so it wasn't spotted.
10257 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10259 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10260 Masking (CBCM). In the absence of test vectors, the best I have been able
10261 to do is check that the decrypt undoes the encrypt, so far. Send me test
10262 vectors if you have them.
10265 *) Correct calculation of key length for export ciphers (too much space was
10266 allocated for null ciphers). This has not been tested!
10269 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10270 message is now correct (it understands "crypto" and "ssl" on its
10271 command line). There is also now an "update" option. This will update
10272 the util/ssleay.num and util/libeay.num files with any new functions.
10274 perl util/mkdef.pl crypto ssl update
10275 it will update them.
10278 *) Overhauled the Perl interface (perl/*):
10279 - ported BN stuff to OpenSSL's different BN library
10280 - made the perl/ source tree CVS-aware
10281 - renamed the package from SSLeay to OpenSSL (the files still contain
10282 their history because I've copied them in the repository)
10283 - removed obsolete files (the test scripts will be replaced
10284 by better Test::Harness variants in the future)
10285 [Ralf S. Engelschall]
10287 *) First cut for a very conservative source tree cleanup:
10288 1. merge various obsolete readme texts into doc/ssleay.txt
10289 where we collect the old documents and readme texts.
10290 2. remove the first part of files where I'm already sure that we no
10291 longer need them because of three reasons: either they are just temporary
10292 files which were left by Eric or they are preserved original files where
10293 I've verified that the diff is also available in the CVS via "cvs diff
10294 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10295 the crypto/md/ stuff).
10296 [Ralf S. Engelschall]
10298 *) More extension code. Incomplete support for subject and issuer alt
10299 name, issuer and authority key id. Change the i2v function parameters
10300 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10301 what that's for :-) Fix to ASN1 macro which messed up
10302 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10305 *) Preliminary support for ENUMERATED type. This is largely copied from the
10309 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10310 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10312 *) Make sure `make rehash' target really finds the `openssl' program.
10313 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10315 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10316 like to hear about it if this slows down other processors.
10319 *) Add CygWin32 platform information to Configure script.
10320 [Alan Batie <batie@aahz.jf.intel.com>]
10322 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10323 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10325 *) New program nseq to manipulate netscape certificate sequences
10328 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10332 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10333 but the BN code had some problems that would cause failures when
10334 doing certificate verification and some other functions.
10335 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10337 *) Add ASN1 and PEM code to support netscape certificate sequences.
10340 *) Add ASN1 and PEM code to support netscape certificate sequences.
10343 *) Add several PKIX and private extended key usage OIDs.
10346 *) Modify the 'ca' program to handle the new extension code. Modify
10347 openssl.cnf for new extension format, add comments.
10350 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10351 and add a sample to openssl.cnf so req -x509 now adds appropriate
10355 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10356 error code, add initial support to X509_print() and x509 application.
10359 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10360 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10361 stuff is currently isolated and isn't even compiled yet.
10364 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10365 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10366 Removed the versions check from X509 routines when loading extensions:
10367 this allows certain broken certificates that don't set the version
10368 properly to be processed.
10371 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10372 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10373 can still be regenerated with "make depend".
10376 *) Spelling mistake in C version of CAST-128.
10377 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10379 *) Changes to the error generation code. The perl script err-code.pl
10380 now reads in the old error codes and retains the old numbers, only
10381 adding new ones if necessary. It also only changes the .err files if new
10382 codes are added. The makefiles have been modified to only insert errors
10383 when needed (to avoid needlessly modifying header files). This is done
10384 by only inserting errors if the .err file is newer than the auto generated
10385 C file. To rebuild all the error codes from scratch (the old behaviour)
10386 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10387 or delete all the .err files.
10390 *) CAST-128 was incorrectly implemented for short keys. The C version has
10391 been fixed, but is untested. The assembler versions are also fixed, but
10392 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10393 to regenerate it if needed.
10394 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10395 Hagino <itojun@kame.net>]
10397 *) File was opened incorrectly in randfile.c.
10398 [Ulf Möller <ulf@fitug.de>]
10400 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10401 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10402 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10403 al: it's just almost always a UTCTime. Note this patch adds new error
10404 codes so do a "make errors" if there are problems.
10407 *) Correct Linux 1 recognition in config.
10408 [Ulf Möller <ulf@fitug.de>]
10410 *) Remove pointless MD5 hash when using DSA keys in ca.
10411 [Anonymous <nobody@replay.com>]
10413 *) Generate an error if given an empty string as a cert directory. Also
10414 generate an error if handed NULL (previously returned 0 to indicate an
10415 error, but didn't set one).
10416 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10418 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10421 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10422 parameters. This was causing a warning which killed off the Win32 compile.
10425 *) Remove C++ style comments from crypto/bn/bn_local.h.
10426 [Neil Costigan <neil.costigan@celocom.com>]
10428 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10429 based on a text string, looking up short and long names and finally
10430 "dot" format. The "dot" format stuff didn't work. Added new function
10431 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10432 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10433 OID is not part of the table.
10436 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10437 X509_LOOKUP_by_alias().
10440 *) Sort openssl functions by name.
10443 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10444 encryption from sample DSA keys (in case anyone is interested the password
10448 *) Make _all_ *_free functions accept a NULL pointer.
10449 [Frans Heymans <fheymans@isaserver.be>]
10451 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10453 [Anonymous <nobody@replay.com>]
10455 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10456 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10458 *) Don't blow it for numeric -newkey arguments to apps/req.
10459 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10461 *) Temp key "for export" tests were wrong in s3_srvr.c.
10462 [Anonymous <nobody@replay.com>]
10464 *) Add prototype for temp key callback functions
10465 SSL_CTX_set_tmp_{rsa,dh}_callback().
10468 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10469 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10472 *) X509_name_add_entry() freed the wrong thing after an error.
10473 [Arne Ansper <arne@ats.cyber.ee>]
10475 *) rsa_eay.c would attempt to free a NULL context.
10476 [Arne Ansper <arne@ats.cyber.ee>]
10478 *) BIO_s_socket() had a broken should_retry() on Windoze.
10479 [Arne Ansper <arne@ats.cyber.ee>]
10481 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10482 [Arne Ansper <arne@ats.cyber.ee>]
10484 *) Make sure the already existing X509_STORE->depth variable is initialized
10485 in X509_STORE_new(), but document the fact that this variable is still
10486 unused in the certificate verification process.
10487 [Ralf S. Engelschall]
10489 *) Fix the various library and apps files to free up pkeys obtained from
10490 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10493 *) Fix reference counting in X509_PUBKEY_get(). This makes
10494 demos/maurice/example2.c work, amongst others, probably.
10495 [Steve Henson and Ben Laurie]
10497 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10498 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10499 are no longer created. This way we have a single and consistent command
10500 line interface `openssl <command>', similar to `cvs <command>'.
10501 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10503 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10504 BIT STRING wrapper always have zero unused bits.
10507 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10510 *) Make the top-level INSTALL documentation easier to understand.
10513 *) Makefiles updated to exit if an error occurs in a sub-directory
10514 make (including if user presses ^C) [Paul Sutton]
10516 *) Make Montgomery context stuff explicit in RSA data structure.
10519 *) Fix build order of pem and err to allow for generated pem.h.
10522 *) Fix renumbering bug in X509_NAME_delete_entry().
10525 *) Enhanced the err-ins.pl script so it makes the error library number
10526 global and can add a library name. This is needed for external ASN1 and
10527 other error libraries.
10530 *) Fixed sk_insert which never worked properly.
10533 *) Fix ASN1 macros so they can handle indefinite length construted
10534 EXPLICIT tags. Some non standard certificates use these: they can now
10538 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10539 into a single doc/ssleay.txt bundle. This way the information is still
10540 preserved but no longer messes up this directory. Now it's new room for
10541 the new set of documenation files.
10542 [Ralf S. Engelschall]
10544 *) SETs were incorrectly DER encoded. This was a major pain, because they
10545 shared code with SEQUENCEs, which aren't coded the same. This means that
10546 almost everything to do with SETs or SEQUENCEs has either changed name or
10547 number of arguments.
10548 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10550 *) Fix test data to work with the above.
10553 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10554 was already fixed by Eric for 0.9.1 it seems.
10555 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10557 *) Autodetect FreeBSD3.
10560 *) Fix various bugs in Configure. This affects the following platforms:
10564 unixware-2.0-pentium
10568 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10569 before they are needed.
10572 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10576 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10578 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10579 changed SSLeay to OpenSSL in version strings.
10580 [Ralf S. Engelschall]
10582 *) Some fixups to the top-level documents.
10585 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10586 because the symlink to include/ was missing.
10587 [Ralf S. Engelschall]
10589 *) Incorporated the popular no-RSA/DSA-only patches
10590 which allow to compile a RSA-free SSLeay.
10591 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10593 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10594 when "ssleay" is still not found.
10595 [Ralf S. Engelschall]
10597 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10598 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10600 *) Updated the README file.
10601 [Ralf S. Engelschall]
10603 *) Added various .cvsignore files in the CVS repository subdirs
10604 to make a "cvs update" really silent.
10605 [Ralf S. Engelschall]
10607 *) Recompiled the error-definition header files and added
10608 missing symbols to the Win32 linker tables.
10609 [Ralf S. Engelschall]
10611 *) Cleaned up the top-level documents;
10612 o new files: CHANGES and LICENSE
10613 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10614 o merged COPYRIGHT into LICENSE
10615 o removed obsolete TODO file
10616 o renamed MICROSOFT to INSTALL.W32
10617 [Ralf S. Engelschall]
10619 *) Removed dummy files from the 0.9.1b source tree:
10620 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10621 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10622 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10623 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10624 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10625 [Ralf S. Engelschall]
10627 *) Added various platform portability fixes.
10630 *) The Genesis of the OpenSSL rpject:
10631 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10632 Young and Tim J. Hudson created while they were working for C2Net until
10634 [The OpenSSL Project]
10637 Changes between 0.9.0b and 0.9.1b [not released]
10639 *) Updated a few CA certificates under certs/
10642 *) Changed some BIGNUM api stuff.
10645 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10646 DGUX x86, Linux Alpha, etc.
10649 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10650 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10654 *) Add -strparse option to asn1pars program which parses nested
10656 [Dr Stephen Henson <shenson@bigfoot.com>]
10658 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10661 *) DSA fix for "ca" program.
10664 *) Added "-genkey" option to "dsaparam" program.
10667 *) Added RIPE MD160 (rmd160) message digest.
10670 *) Added -a (all) option to "ssleay version" command.
10673 *) Added PLATFORM define which is the id given to Configure.
10676 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10679 *) Extended the ASN.1 parser routines.
10682 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10685 *) Added a BN_CTX to the BN library.
10688 *) Fixed the weak key values in DES library
10691 *) Changed API in EVP library for cipher aliases.
10694 *) Added support for RC2/64bit cipher.
10697 *) Converted the lhash library to the crypto/mem.c functions.
10700 *) Added more recognized ASN.1 object ids.
10703 *) Added more RSA padding checks for SSL/TLS.
10706 *) Added BIO proxy/filter functionality.
10709 *) Added extra_certs to SSL_CTX which can be used
10710 send extra CA certificates to the client in the CA cert chain sending
10711 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10714 *) Now Fortezza is denied in the authentication phase because
10715 this is key exchange mechanism is not supported by SSLeay at all.
10718 *) Additional PKCS1 checks.
10721 *) Support the string "TLSv1" for all TLS v1 ciphers.
10724 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10725 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10728 *) Fixed a few memory leaks.
10731 *) Fixed various code and comment typos.
10734 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10735 bytes sent in the client random.
10736 [Edward Bishop <ebishop@spyglass.com>]