5 Changes between 1.0.1t and 1.0.1u [xx XXX xxxx]
7 *) In order to mitigate the SWEET32 attack (CVE-2016-2183),
8 the DES ciphers were moved from HIGH to MEDIUM.
11 Changes between 1.0.1s and 1.0.1t [3 May 2016]
13 *) Prevent padding oracle in AES-NI CBC MAC check
15 A MITM attacker can use a padding oracle attack to decrypt traffic
16 when the connection uses an AES CBC cipher and the server support
19 This issue was introduced as part of the fix for Lucky 13 padding
20 attack (CVE-2013-0169). The padding check was rewritten to be in
21 constant time by making sure that always the same bytes are read and
22 compared against either the MAC or padding bytes. But it no longer
23 checked that there was enough data to have both the MAC and padding
26 This issue was reported by Juraj Somorovsky using TLS-Attacker.
30 *) Fix EVP_EncodeUpdate overflow
32 An overflow can occur in the EVP_EncodeUpdate() function which is used for
33 Base64 encoding of binary data. If an attacker is able to supply very large
34 amounts of input data then a length check can overflow resulting in a heap
37 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
38 the PEM_write_bio* family of functions. These are mainly used within the
39 OpenSSL command line applications, so any application which processes data
40 from an untrusted source and outputs it as a PEM file should be considered
41 vulnerable to this issue. User applications that call these APIs directly
42 with large amounts of untrusted data may also be vulnerable.
44 This issue was reported by Guido Vranken.
48 *) Fix EVP_EncryptUpdate overflow
50 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
51 is able to supply very large amounts of input data after a previous call to
52 EVP_EncryptUpdate() with a partial block then a length check can overflow
53 resulting in a heap corruption. Following an analysis of all OpenSSL
54 internal usage of the EVP_EncryptUpdate() function all usage is one of two
55 forms. The first form is where the EVP_EncryptUpdate() call is known to be
56 the first called function after an EVP_EncryptInit(), and therefore that
57 specific call must be safe. The second form is where the length passed to
58 EVP_EncryptUpdate() can be seen from the code to be some small value and
59 therefore there is no possibility of an overflow. Since all instances are
60 one of these two forms, it is believed that there can be no overflows in
61 internal code due to this problem. It should be noted that
62 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
63 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
64 of these calls have also been analysed too and it is believed there are no
65 instances in internal usage where an overflow could occur.
67 This issue was reported by Guido Vranken.
71 *) Prevent ASN.1 BIO excessive memory allocation
73 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
74 a short invalid encoding can casuse allocation of large amounts of memory
75 potentially consuming excessive resources or exhausting memory.
77 Any application parsing untrusted data through d2i BIO functions is
78 affected. The memory based functions such as d2i_X509() are *not* affected.
79 Since the memory based functions are used by the TLS library, TLS
80 applications are not affected.
82 This issue was reported by Brian Carpenter.
88 ASN1 Strings that are over 1024 bytes can cause an overread in applications
89 using the X509_NAME_oneline() function on EBCDIC systems. This could result
90 in arbitrary stack data being returned in the buffer.
92 This issue was reported by Guido Vranken.
96 *) Modify behavior of ALPN to invoke callback after SNI/servername
97 callback, such that updates to the SSL_CTX affect ALPN.
100 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
104 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
105 methods are enabled and ssl2 is disabled the methods return NULL.
108 Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
110 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
111 Builds that are not configured with "enable-weak-ssl-ciphers" will not
112 provide any "EXPORT" or "LOW" strength ciphers.
115 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
116 is by default disabled at build-time. Builds that are not configured with
117 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
118 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
119 will need to explicitly call either of:
121 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
123 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
125 as appropriate. Even if either of those is used, or the application
126 explicitly uses the version-specific SSLv2_method() or its client and
127 server variants, SSLv2 ciphers vulnerable to exhaustive search key
128 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
129 ciphers, and SSLv2 56-bit DES are no longer available.
133 *) Fix a double-free in DSA code
135 A double free bug was discovered when OpenSSL parses malformed DSA private
136 keys and could lead to a DoS attack or memory corruption for applications
137 that receive DSA private keys from untrusted sources. This scenario is
140 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
145 *) Disable SRP fake user seed to address a server memory leak.
147 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
149 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
150 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
151 was changed to ignore the "fake user" SRP seed, even if the seed
154 Users should use SRP_VBASE_get1_by_user instead. Note that in
155 SRP_VBASE_get1_by_user, caller must free the returned value. Note
156 also that even though configuring the SRP seed attempts to hide
157 invalid usernames by continuing the handshake with fake
158 credentials, this behaviour is not constant time and no strong
159 guarantees are made that the handshake is indistinguishable from
160 that of a valid user.
164 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
166 In the BN_hex2bn function the number of hex digits is calculated using an
167 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
168 large values of |i| this can result in |bn_expand| not allocating any
169 memory because |i * 4| is negative. This can leave the internal BIGNUM data
170 field as NULL leading to a subsequent NULL ptr deref. For very large values
171 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
172 In this case memory is allocated to the internal BIGNUM data field, but it
173 is insufficiently sized leading to heap corruption. A similar issue exists
174 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
175 is ever called by user applications with very large untrusted hex/dec data.
176 This is anticipated to be a rare occurrence.
178 All OpenSSL internal usage of these functions use data that is not expected
179 to be untrusted, e.g. config file data or application command line
180 arguments. If user developed applications generate config file data based
181 on untrusted data then it is possible that this could also lead to security
182 consequences. This is also anticipated to be rare.
184 This issue was reported to OpenSSL by Guido Vranken.
188 *) Fix memory issues in BIO_*printf functions
190 The internal |fmtstr| function used in processing a "%s" format string in
191 the BIO_*printf functions could overflow while calculating the length of a
192 string and cause an OOB read when printing very long strings.
194 Additionally the internal |doapr_outch| function can attempt to write to an
195 OOB memory location (at an offset from the NULL pointer) in the event of a
196 memory allocation failure. In 1.0.2 and below this could be caused where
197 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
198 could be in processing a very long "%s" format string. Memory leaks can
201 The first issue may mask the second issue dependent on compiler behaviour.
202 These problems could enable attacks where large amounts of untrusted data
203 is passed to the BIO_*printf functions. If applications use these functions
204 in this way then they could be vulnerable. OpenSSL itself uses these
205 functions when printing out human-readable dumps of ASN.1 data. Therefore
206 applications that print this data could be vulnerable if the data is from
207 untrusted sources. OpenSSL command line applications could also be
208 vulnerable where they print out ASN.1 data, or if untrusted data is passed
209 as command line arguments.
211 Libssl is not considered directly vulnerable. Additionally certificates etc
212 received via remote connections via libssl are also unlikely to be able to
213 trigger these issues because of message size limits enforced within libssl.
215 This issue was reported to OpenSSL Guido Vranken.
219 *) Side channel attack on modular exponentiation
221 A side-channel attack was found which makes use of cache-bank conflicts on
222 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
223 of RSA keys. The ability to exploit this issue is limited as it relies on
224 an attacker who has control of code in a thread running on the same
225 hyper-threaded core as the victim thread which is performing decryptions.
227 This issue was reported to OpenSSL by Yuval Yarom, The University of
228 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
229 Nadia Heninger, University of Pennsylvania with more information at
230 http://cachebleed.info.
234 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
235 if no keysize is specified with default_bits. This fixes an
236 omission in an earlier change that changed all RSA/DSA key generation
237 apps to use 2048 bits by default.
240 Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
242 *) Protection for DH small subgroup attacks
244 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
245 switched on by default and cannot be disabled. This could have some
249 *) SSLv2 doesn't block disabled ciphers
251 A malicious client can negotiate SSLv2 ciphers that have been disabled on
252 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
253 been disabled, provided that the SSLv2 protocol was not also disabled via
256 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
257 and Sebastian Schinzel.
261 *) Reject DH handshakes with parameters shorter than 1024 bits.
264 Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
266 *) Certificate verify crash with missing PSS parameter
268 The signature verification routines will crash with a NULL pointer
269 dereference if presented with an ASN.1 signature using the RSA PSS
270 algorithm and absent mask generation function parameter. Since these
271 routines are used to verify certificate signature algorithms this can be
272 used to crash any certificate verification operation and exploited in a
273 DoS attack. Any application which performs certificate verification is
274 vulnerable including OpenSSL clients and servers which enable client
277 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
281 *) X509_ATTRIBUTE memory leak
283 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
284 memory. This structure is used by the PKCS#7 and CMS routines so any
285 application which reads PKCS#7 or CMS data from untrusted sources is
286 affected. SSL/TLS is not affected.
288 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
293 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
294 This changes the decoding behaviour for some invalid messages,
295 though the change is mostly in the more lenient direction, and
296 legacy behaviour is preserved as much as possible.
299 *) In DSA_generate_parameters_ex, if the provided seed is too short,
300 use a random seed, as already documented.
301 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
303 Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
305 *) Alternate chains certificate forgery
307 During certificate verfification, OpenSSL will attempt to find an
308 alternative certificate chain if the first attempt to build such a chain
309 fails. An error in the implementation of this logic can mean that an
310 attacker could cause certain checks on untrusted certificates to be
311 bypassed, such as the CA flag, enabling them to use a valid leaf
312 certificate to act as a CA and "issue" an invalid certificate.
314 This issue was reported to OpenSSL by Adam Langley/David Benjamin
319 *) Race condition handling PSK identify hint
321 If PSK identity hints are received by a multi-threaded client then
322 the values are wrongly updated in the parent SSL_CTX structure. This can
323 result in a race condition potentially leading to a double free of the
328 Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
329 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
330 incompatibility in the handling of HMAC. The previous ABI has now been
333 Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
335 *) Malformed ECParameters causes infinite loop
337 When processing an ECParameters structure OpenSSL enters an infinite loop
338 if the curve specified is over a specially malformed binary polynomial
341 This can be used to perform denial of service against any
342 system which processes public keys, certificate requests or
343 certificates. This includes TLS clients and TLS servers with
344 client authentication enabled.
346 This issue was reported to OpenSSL by Joseph Barr-Pixton.
350 *) Exploitable out-of-bounds read in X509_cmp_time
352 X509_cmp_time does not properly check the length of the ASN1_TIME
353 string and can read a few bytes out of bounds. In addition,
354 X509_cmp_time accepts an arbitrary number of fractional seconds in the
357 An attacker can use this to craft malformed certificates and CRLs of
358 various sizes and potentially cause a segmentation fault, resulting in
359 a DoS on applications that verify certificates or CRLs. TLS clients
360 that verify CRLs are affected. TLS clients and servers with client
361 authentication enabled may be affected if they use custom verification
364 This issue was reported to OpenSSL by Robert Swiecki (Google), and
365 independently by Hanno Böck.
369 *) PKCS7 crash with missing EnvelopedContent
371 The PKCS#7 parsing code does not handle missing inner EncryptedContent
372 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
373 with missing content and trigger a NULL pointer dereference on parsing.
375 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
376 structures from untrusted sources are affected. OpenSSL clients and
377 servers are not affected.
379 This issue was reported to OpenSSL by Michal Zalewski (Google).
383 *) CMS verify infinite loop with unknown hash function
385 When verifying a signedData message the CMS code can enter an infinite loop
386 if presented with an unknown hash function OID. This can be used to perform
387 denial of service against any system which verifies signedData messages using
389 This issue was reported to OpenSSL by Johannes Bauer.
393 *) Race condition handling NewSessionTicket
395 If a NewSessionTicket is received by a multi-threaded client when attempting to
396 reuse a previous ticket then a race condition can occur potentially leading to
397 a double free of the ticket data.
401 *) Reject DH handshakes with parameters shorter than 768 bits.
402 [Kurt Roeckx and Emilia Kasper]
404 *) dhparam: generate 2048-bit parameters by default.
405 [Kurt Roeckx and Emilia Kasper]
407 Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
409 *) Segmentation fault in ASN1_TYPE_cmp fix
411 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
412 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
413 certificate signature algorithm consistency this can be used to crash any
414 certificate verification operation and exploited in a DoS attack. Any
415 application which performs certificate verification is vulnerable including
416 OpenSSL clients and servers which enable client authentication.
420 *) ASN.1 structure reuse memory corruption fix
422 Reusing a structure in ASN.1 parsing may allow an attacker to cause
423 memory corruption via an invalid write. Such reuse is and has been
424 strongly discouraged and is believed to be rare.
426 Applications that parse structures containing CHOICE or ANY DEFINED BY
427 components may be affected. Certificate parsing (d2i_X509 and related
428 functions) are however not affected. OpenSSL clients and servers are
433 *) PKCS7 NULL pointer dereferences fix
435 The PKCS#7 parsing code does not handle missing outer ContentInfo
436 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
437 missing content and trigger a NULL pointer dereference on parsing.
439 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
440 otherwise parse PKCS#7 structures from untrusted sources are
441 affected. OpenSSL clients and servers are not affected.
443 This issue was reported to OpenSSL by Michal Zalewski (Google).
447 *) DoS via reachable assert in SSLv2 servers fix
449 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
450 servers that both support SSLv2 and enable export cipher suites by sending
451 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
453 This issue was discovered by Sean Burford (Google) and Emilia Käsper
454 (OpenSSL development team).
458 *) Use After Free following d2i_ECPrivatekey error fix
460 A malformed EC private key file consumed via the d2i_ECPrivateKey function
461 could cause a use after free condition. This, in turn, could cause a double
462 free in several private key parsing functions (such as d2i_PrivateKey
463 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
464 for applications that receive EC private keys from untrusted
465 sources. This scenario is considered rare.
467 This issue was discovered by the BoringSSL project and fixed in their
472 *) X509_to_X509_REQ NULL pointer deref fix
474 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
475 the certificate key is invalid. This function is rarely used in practice.
477 This issue was discovered by Brian Carpenter.
481 *) Removed the export ciphers from the DEFAULT ciphers
484 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
486 *) Build fixes for the Windows and OpenVMS platforms
487 [Matt Caswell and Richard Levitte]
489 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
491 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
492 message can cause a segmentation fault in OpenSSL due to a NULL pointer
493 dereference. This could lead to a Denial Of Service attack. Thanks to
494 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
498 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
499 dtls1_buffer_record function under certain conditions. In particular this
500 could occur if an attacker sent repeated DTLS records with the same
501 sequence number but for the next epoch. The memory leak could be exploited
502 by an attacker in a Denial of Service attack through memory exhaustion.
503 Thanks to Chris Mueller for reporting this issue.
507 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
508 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
509 method would be set to NULL which could later result in a NULL pointer
510 dereference. Thanks to Frank Schmirler for reporting this issue.
514 *) Abort handshake if server key exchange message is omitted for ephemeral
517 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
518 reporting this issue.
522 *) Remove non-export ephemeral RSA code on client and server. This code
523 violated the TLS standard by allowing the use of temporary RSA keys in
524 non-export ciphersuites and could be used by a server to effectively
525 downgrade the RSA key length used to a value smaller than the server
526 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
527 INRIA or reporting this issue.
531 *) Fixed issue where DH client certificates are accepted without verification.
532 An OpenSSL server will accept a DH certificate for client authentication
533 without the certificate verify message. This effectively allows a client to
534 authenticate without the use of a private key. This only affects servers
535 which trust a client certificate authority which issues certificates
536 containing DH keys: these are extremely rare and hardly ever encountered.
537 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
542 *) Ensure that the session ID context of an SSL is updated when its
543 SSL_CTX is updated via SSL_set_SSL_CTX.
545 The session ID context is typically set from the parent SSL_CTX,
546 and can vary with the CTX.
549 *) Fix various certificate fingerprint issues.
551 By using non-DER or invalid encodings outside the signed portion of a
552 certificate the fingerprint can be changed without breaking the signature.
553 Although no details of the signed portion of the certificate can be changed
554 this can cause problems with some applications: e.g. those using the
555 certificate fingerprint for blacklists.
557 1. Reject signatures with non zero unused bits.
559 If the BIT STRING containing the signature has non zero unused bits reject
560 the signature. All current signature algorithms require zero unused bits.
562 2. Check certificate algorithm consistency.
564 Check the AlgorithmIdentifier inside TBS matches the one in the
565 certificate signature. NB: this will result in signature failure
566 errors for some broken certificates.
568 Thanks to Konrad Kraszewski from Google for reporting this issue.
570 3. Check DSA/ECDSA signatures use DER.
572 Reencode DSA/ECDSA signatures and compare with the original received
573 signature. Return an error if there is a mismatch.
575 This will reject various cases including garbage after signature
576 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
577 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
578 (negative or with leading zeroes).
580 Further analysis was conducted and fixes were developed by Stephen Henson
581 of the OpenSSL core team.
586 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
587 results on some platforms, including x86_64. This bug occurs at random
588 with a very low probability, and is not known to be exploitable in any
589 way, though its exact impact is difficult to determine. Thanks to Pieter
590 Wuille (Blockstream) who reported this issue and also suggested an initial
591 fix. Further analysis was conducted by the OpenSSL development team and
592 Adam Langley of Google. The final fix was developed by Andy Polyakov of
593 the OpenSSL core team.
597 *) Do not resume sessions on the server if the negotiated protocol
598 version does not match the session's version. Resuming with a different
599 version, while not strictly forbidden by the RFC, is of questionable
600 sanity and breaks all known clients.
601 [David Benjamin, Emilia Käsper]
603 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
604 early CCS messages during renegotiation. (Note that because
605 renegotiation is encrypted, this early CCS was not exploitable.)
608 *) Tighten client-side session ticket handling during renegotiation:
609 ensure that the client only accepts a session ticket if the server sends
610 the extension anew in the ServerHello. Previously, a TLS client would
611 reuse the old extension state and thus accept a session ticket if one was
612 announced in the initial ServerHello.
614 Similarly, ensure that the client requires a session ticket if one
615 was advertised in the ServerHello. Previously, a TLS client would
616 ignore a missing NewSessionTicket message.
619 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
623 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
624 sends a carefully crafted handshake message, to cause OpenSSL to fail
625 to free up to 64k of memory causing a memory leak. This could be
626 exploited in a Denial Of Service attack. This issue affects OpenSSL
627 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
628 whether SRTP is used or configured. Implementations of OpenSSL that
629 have been compiled with OPENSSL_NO_SRTP defined are not affected.
631 The fix was developed by the OpenSSL team.
635 *) Session Ticket Memory Leak.
637 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
638 integrity of that ticket is first verified. In the event of a session
639 ticket integrity check failing, OpenSSL will fail to free memory
640 causing a memory leak. By sending a large number of invalid session
641 tickets an attacker could exploit this issue in a Denial Of Service
646 *) Build option no-ssl3 is incomplete.
648 When OpenSSL is configured with "no-ssl3" as a build option, servers
649 could accept and complete a SSL 3.0 handshake, and clients could be
650 configured to send them.
652 [Akamai and the OpenSSL team]
654 *) Add support for TLS_FALLBACK_SCSV.
655 Client applications doing fallback retries should call
656 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
658 [Adam Langley, Bodo Moeller]
660 *) Add additional DigestInfo checks.
662 Reencode DigestInto in DER and check against the original when
663 verifying RSA signature: this will reject any improperly encoded
664 DigestInfo structures.
666 Note: this is a precautionary measure and no attacks are currently known.
670 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
672 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
673 SRP code can be overrun an internal buffer. Add sanity check that
674 g, A, B < N to SRP code.
676 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
677 Group for discovering this issue.
681 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
682 TLS 1.0 instead of higher protocol versions when the ClientHello message
683 is badly fragmented. This allows a man-in-the-middle attacker to force a
684 downgrade to TLS 1.0 even if both the server and the client support a
685 higher protocol version, by modifying the client's TLS records.
687 Thanks to David Benjamin and Adam Langley (Google) for discovering and
688 researching this issue.
692 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
693 to a denial of service attack. A malicious server can crash the client
694 with a null pointer dereference (read) by specifying an anonymous (EC)DH
695 ciphersuite and sending carefully crafted handshake messages.
697 Thanks to Felix Gröbert (Google) for discovering and researching this
702 *) By sending carefully crafted DTLS packets an attacker could cause openssl
703 to leak memory. This can be exploited through a Denial of Service attack.
704 Thanks to Adam Langley for discovering and researching this issue.
708 *) An attacker can force openssl to consume large amounts of memory whilst
709 processing DTLS handshake messages. This can be exploited through a
710 Denial of Service attack.
711 Thanks to Adam Langley for discovering and researching this issue.
715 *) An attacker can force an error condition which causes openssl to crash
716 whilst processing DTLS packets due to memory being freed twice. This
717 can be exploited through a Denial of Service attack.
718 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
723 *) If a multithreaded client connects to a malicious server using a resumed
724 session and the server sends an ec point format extension it could write
725 up to 255 bytes to freed memory.
727 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
732 *) A malicious server can crash an OpenSSL client with a null pointer
733 dereference (read) by specifying an SRP ciphersuite even though it was not
734 properly negotiated with the client. This can be exploited through a
735 Denial of Service attack.
737 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
738 discovering and researching this issue.
742 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
743 X509_name_oneline, X509_name_print_ex et al. to leak some information
744 from the stack. Applications may be affected if they echo pretty printing
745 output to the attacker.
747 Thanks to Ivan Fratric (Google) for discovering this issue.
749 [Emilia Käsper, and Steve Henson]
751 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
752 for corner cases. (Certain input points at infinity could lead to
753 bogus results, with non-infinity inputs mapped to infinity too.)
756 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
758 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
759 handshake can force the use of weak keying material in OpenSSL
760 SSL/TLS clients and servers.
762 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
763 researching this issue. (CVE-2014-0224)
764 [KIKUCHI Masashi, Steve Henson]
766 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
767 OpenSSL DTLS client the code can be made to recurse eventually crashing
770 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
772 [Imre Rad, Steve Henson]
774 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
775 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
776 client or server. This is potentially exploitable to run arbitrary
777 code on a vulnerable client or server.
779 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
780 [Jüri Aedla, Steve Henson]
782 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
783 are subject to a denial of service attack.
785 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
786 this issue. (CVE-2014-3470)
787 [Felix Gröbert, Ivan Fratric, Steve Henson]
789 *) Harmonize version and its documentation. -f flag is used to display
791 [mancha <mancha1@zoho.com>]
793 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
795 [mancha <mancha1@zoho.com>]
797 *) Fix some double frees. These are not thought to be exploitable.
798 [mancha <mancha1@zoho.com>]
800 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
802 *) A missing bounds check in the handling of the TLS heartbeat extension
803 can be used to reveal up to 64k of memory to a connected client or
806 Thanks for Neel Mehta of Google Security for discovering this bug and to
807 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
808 preparing the fix (CVE-2014-0160)
809 [Adam Langley, Bodo Moeller]
811 *) Fix for the attack described in the paper "Recovering OpenSSL
812 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
813 by Yuval Yarom and Naomi Benger. Details can be obtained from:
814 http://eprint.iacr.org/2014/140
816 Thanks to Yuval Yarom and Naomi Benger for discovering this
817 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
818 [Yuval Yarom and Naomi Benger]
820 *) TLS pad extension: draft-agl-tls-padding-03
822 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
823 TLS client Hello record length value would otherwise be > 255 and
824 less that 512 pad with a dummy extension containing zeroes so it
825 is at least 512 bytes long.
827 [Adam Langley, Steve Henson]
829 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
831 *) Fix for TLS record tampering bug. A carefully crafted invalid
832 handshake could crash OpenSSL with a NULL pointer exception.
833 Thanks to Anton Johansson for reporting this issues.
836 *) Keep original DTLS digest and encryption contexts in retransmission
837 structures so we can use the previous session parameters if they need
838 to be resent. (CVE-2013-6450)
841 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
842 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
843 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
844 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
845 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
846 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
847 [Rob Stradling, Adam Langley]
849 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
851 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
852 supporting platforms or when small records were transferred.
853 [Andy Polyakov, Steve Henson]
855 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
857 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
859 This addresses the flaw in CBC record processing discovered by
860 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
861 at: http://www.isg.rhul.ac.uk/tls/
863 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
864 Security Group at Royal Holloway, University of London
865 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
866 Emilia Käsper for the initial patch.
868 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
870 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
871 ciphersuites which can be exploited in a denial of service attack.
872 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
873 and detecting this bug and to Wolfgang Ettlinger
874 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
878 *) Return an error when checking OCSP signatures when key is NULL.
879 This fixes a DoS attack. (CVE-2013-0166)
882 *) Make openssl verify return errors.
883 [Chris Palmer <palmer@google.com> and Ben Laurie]
885 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
886 the right response is stapled. Also change SSL_get_certificate()
887 so it returns the certificate actually sent.
888 See http://rt.openssl.org/Ticket/Display.html?id=2836.
889 [Rob Stradling <rob.stradling@comodo.com>]
891 *) Fix possible deadlock when decoding public keys.
894 *) Don't use TLS 1.0 record version number in initial client hello
898 Changes between 1.0.1b and 1.0.1c [10 May 2012]
900 *) Sanity check record length before skipping explicit IV in TLS
901 1.2, 1.1 and DTLS to fix DoS attack.
903 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
904 fuzzing as a service testing platform.
908 *) Initialise tkeylen properly when encrypting CMS messages.
909 Thanks to Solar Designer of Openwall for reporting this issue.
912 *) In FIPS mode don't try to use composite ciphers as they are not
916 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
918 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
919 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
920 mean any application compiled against OpenSSL 1.0.0 headers setting
921 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
922 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
923 0x10000000L Any application which was previously compiled against
924 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
925 will need to be recompiled as a result. Letting be results in
926 inability to disable specifically TLS 1.1 and in client context,
927 in unlike event, limit maximum offered version to TLS 1.0 [see below].
930 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
931 disable just protocol X, but all protocols above X *if* there are
932 protocols *below* X still enabled. In more practical terms it means
933 that if application wants to disable TLS1.0 in favor of TLS1.1 and
934 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
935 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
939 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
941 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
942 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
943 in CRYPTO_realloc_clean.
945 Thanks to Tavis Ormandy, Google Security Team, for discovering this
946 issue and to Adam Langley <agl@chromium.org> for fixing it.
948 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
950 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
953 *) Workarounds for some broken servers that "hang" if a client hello
954 record length exceeds 255 bytes.
956 1. Do not use record version number > TLS 1.0 in initial client
957 hello: some (but not all) hanging servers will now work.
958 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
959 the number of ciphers sent in the client hello. This should be
960 set to an even number, such as 50, for example by passing:
961 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
962 Most broken servers should now work.
963 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
964 TLS 1.2 client support entirely.
967 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
970 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
972 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
973 STRING form instead of a DigestInfo.
976 *) The format used for MDC2 RSA signatures is inconsistent between EVP
977 and the RSA_sign/RSA_verify functions. This was made more apparent when
978 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
979 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
980 the correct format in RSA_verify so both forms transparently work.
983 *) Some servers which support TLS 1.0 can choke if we initially indicate
984 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
985 encrypted premaster secret. As a workaround use the maximum pemitted
986 client version in client hello, this should keep such servers happy
987 and still work with previous versions of OpenSSL.
990 *) Add support for TLS/DTLS heartbeats.
991 [Robin Seggelmann <seggelmann@fh-muenster.de>]
993 *) Add support for SCTP.
994 [Robin Seggelmann <seggelmann@fh-muenster.de>]
996 *) Improved PRNG seeding for VOS.
997 [Paul Green <Paul.Green@stratus.com>]
999 *) Extensive assembler packs updates, most notably:
1001 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1002 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1003 - x86_64: bit-sliced AES implementation;
1004 - ARM: NEON support, contemporary platforms optimizations;
1005 - s390x: z196 support;
1006 - *: GHASH and GF(2^m) multiplication implementations;
1010 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1011 (removal of unnecessary code)
1012 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1014 *) Add TLS key material exporter from RFC 5705.
1017 *) Add DTLS-SRTP negotiation from RFC 5764.
1020 *) Add Next Protocol Negotiation,
1021 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1022 disabled with a no-npn flag to config or Configure. Code donated
1024 [Adam Langley <agl@google.com> and Ben Laurie]
1026 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1027 NIST-P256, NIST-P521, with constant-time single point multiplication on
1028 typical inputs. Compiler support for the nonstandard type __uint128_t is
1029 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1030 Code made available under Apache License version 2.0.
1032 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1033 line to include this in your build of OpenSSL, and run "make depend" (or
1034 "make update"). This enables the following EC_METHODs:
1036 EC_GFp_nistp224_method()
1037 EC_GFp_nistp256_method()
1038 EC_GFp_nistp521_method()
1040 EC_GROUP_new_by_curve_name() will automatically use these (while
1041 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1043 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1045 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1046 all platforms. Move ssize_t definition from e_os.h to the public
1047 header file e_os2.h as it now appears in public header file cms.h
1050 *) New -sigopt option to the ca, req and x509 utilities. Additional
1051 signature parameters can be passed using this option and in
1055 *) Add RSA PSS signing function. This will generate and set the
1056 appropriate AlgorithmIdentifiers for PSS based on those in the
1057 corresponding EVP_MD_CTX structure. No application support yet.
1060 *) Support for companion algorithm specific ASN1 signing routines.
1061 New function ASN1_item_sign_ctx() signs a pre-initialised
1062 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1063 the appropriate parameters.
1066 *) Add new algorithm specific ASN1 verification initialisation function
1067 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1068 handling will be the same no matter what EVP_PKEY_METHOD is used.
1069 Add a PSS handler to support verification of PSS signatures: checked
1070 against a number of sample certificates.
1073 *) Add signature printing for PSS. Add PSS OIDs.
1074 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1076 *) Add algorithm specific signature printing. An individual ASN1 method
1077 can now print out signatures instead of the standard hex dump.
1079 More complex signatures (e.g. PSS) can print out more meaningful
1080 information. Include DSA version that prints out the signature
1084 *) Password based recipient info support for CMS library: implementing
1088 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1089 neatly separates the code into cipher and PBE sections and is required
1090 for some algorithms that split PBES2 into separate pieces (such as
1091 password based CMS).
1094 *) Session-handling fixes:
1095 - Fix handling of connections that are resuming with a session ID,
1096 but also support Session Tickets.
1097 - Fix a bug that suppressed issuing of a new ticket if the client
1098 presented a ticket with an expired session.
1099 - Try to set the ticket lifetime hint to something reasonable.
1100 - Make tickets shorter by excluding irrelevant information.
1101 - On the client side, don't ignore renewed tickets.
1102 [Adam Langley, Bodo Moeller (Google)]
1104 *) Fix PSK session representation.
1107 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1109 This work was sponsored by Intel.
1112 *) Add GCM support to TLS library. Some custom code is needed to split
1113 the IV between the fixed (from PRF) and explicit (from TLS record)
1114 portions. This adds all GCM ciphersuites supported by RFC5288 and
1115 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
1116 add a special AESGCM string for GCM only.
1119 *) Expand range of ctrls for AES GCM. Permit setting invocation
1120 field on decrypt and retrieval of invocation field only on encrypt.
1123 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1124 As required by RFC5289 these ciphersuites cannot be used if for
1125 versions of TLS earlier than 1.2.
1128 *) For FIPS capable OpenSSL interpret a NULL default public key method
1129 as unset and return the appopriate default but do *not* set the default.
1130 This means we can return the appopriate method in applications that
1131 swicth between FIPS and non-FIPS modes.
1134 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1135 ENGINE is used then we cannot handle that in the FIPS module so we
1136 keep original code iff non-FIPS operations are allowed.
1139 *) Add -attime option to openssl utilities.
1140 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1142 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1145 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1146 FIPS EC methods unconditionally for now.
1149 *) New build option no-ec2m to disable characteristic 2 code.
1152 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1153 all cases can be covered as some introduce binary incompatibilities.
1156 *) Redirect RSA operations to FIPS module including keygen,
1157 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1160 *) Add similar low level API blocking to ciphers.
1163 *) Low level digest APIs are not approved in FIPS mode: any attempt
1164 to use these will cause a fatal error. Applications that *really* want
1165 to use them can use the private_* version instead.
1168 *) Redirect cipher operations to FIPS module for FIPS builds.
1171 *) Redirect digest operations to FIPS module for FIPS builds.
1174 *) Update build system to add "fips" flag which will link in fipscanister.o
1175 for static and shared library builds embedding a signature if needed.
1178 *) Output TLS supported curves in preference order instead of numerical
1179 order. This is currently hardcoded for the highest order curves first.
1180 This should be configurable so applications can judge speed vs strength.
1183 *) Add TLS v1.2 server support for client authentication.
1186 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1190 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1191 FIPS modules versions.
1194 *) Add TLS v1.2 client side support for client authentication. Keep cache
1195 of handshake records longer as we don't know the hash algorithm to use
1196 until after the certificate request message is received.
1199 *) Initial TLS v1.2 client support. Add a default signature algorithms
1200 extension including all the algorithms we support. Parse new signature
1201 format in client key exchange. Relax some ECC signing restrictions for
1202 TLS v1.2 as indicated in RFC5246.
1205 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1206 to new signature format when needed using client digest preference.
1207 All server ciphersuites should now work correctly in TLS v1.2. No client
1208 support yet and no support for client certificates.
1211 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1212 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1213 ciphersuites. At present only RSA key exchange ciphersuites work with
1214 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1215 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1216 and version checking.
1219 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1220 with this defined it will not be affected by any changes to ssl internal
1221 structures. Add several utility functions to allow openssl application
1222 to work with OPENSSL_NO_SSL_INTERN defined.
1226 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1228 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1231 *) Permit abbreviated handshakes when renegotiating using the function
1232 SSL_renegotiate_abbreviated().
1233 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1235 *) Add call to ENGINE_register_all_complete() to
1236 ENGINE_load_builtin_engines(), so some implementations get used
1237 automatically instead of needing explicit application support.
1240 *) Add support for TLS key exporter as described in RFC5705.
1241 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1243 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1244 a few changes are required:
1246 Add SSL_OP_NO_TLSv1_1 flag.
1247 Add TLSv1_1 methods.
1248 Update version checking logic to handle version 1.1.
1249 Add explicit IV handling (ported from DTLS code).
1250 Add command line options to s_client/s_server.
1253 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1255 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1256 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1257 content decryption and always return the same error. Note: this attack
1258 needs on average 2^20 messages so it only affects automated senders. The
1259 old behaviour can be reenabled in the CMS code by setting the
1260 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1261 an MMA defence is not necessary.
1262 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1263 this issue. (CVE-2012-0884)
1266 *) Fix CVE-2011-4619: make sure we really are receiving a
1267 client hello before rejecting multiple SGC restarts. Thanks to
1268 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1271 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1273 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1274 Thanks to Antonio Martin, Enterprise Secure Access Research and
1275 Development, Cisco Systems, Inc. for discovering this bug and
1276 preparing a fix. (CVE-2012-0050)
1279 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1281 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1282 of the Vaudenay padding oracle attack on CBC mode encryption
1283 which enables an efficient plaintext recovery attack against
1284 the OpenSSL implementation of DTLS. Their attack exploits timing
1285 differences arising during decryption processing. A research
1286 paper describing this attack can be found at:
1287 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1288 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1289 Security Group at Royal Holloway, University of London
1290 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1291 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1292 for preparing the fix. (CVE-2011-4108)
1293 [Robin Seggelmann, Michael Tuexen]
1295 *) Clear bytes used for block padding of SSL 3.0 records.
1297 [Adam Langley (Google)]
1299 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1300 Kadianakis <desnacked@gmail.com> for discovering this issue and
1301 Adam Langley for preparing the fix. (CVE-2011-4619)
1302 [Adam Langley (Google)]
1304 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1305 [Andrey Kulikov <amdeich@gmail.com>]
1307 *) Prevent malformed RFC3779 data triggering an assertion failure.
1308 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1309 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1310 [Rob Austein <sra@hactrn.net>]
1312 *) Improved PRNG seeding for VOS.
1313 [Paul Green <Paul.Green@stratus.com>]
1315 *) Fix ssl_ciph.c set-up race.
1316 [Adam Langley (Google)]
1318 *) Fix spurious failures in ecdsatest.c.
1319 [Emilia Käsper (Google)]
1321 *) Fix the BIO_f_buffer() implementation (which was mixing different
1322 interpretations of the '..._len' fields).
1323 [Adam Langley (Google)]
1325 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1326 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1327 threads won't reuse the same blinding coefficients.
1329 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1330 lock to call BN_BLINDING_invert_ex, and avoids one use of
1331 BN_BLINDING_update for each BN_BLINDING structure (previously,
1332 the last update always remained unused).
1333 [Emilia Käsper (Google)]
1335 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1336 [Bob Buckholz (Google)]
1338 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1340 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1341 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1342 [Kaspar Brand <ossl@velox.ch>]
1344 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1345 for multi-threaded use of ECDH. (CVE-2011-3210)
1346 [Adam Langley (Google)]
1348 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1351 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1352 signature public key algorithm by using OID xref utilities instead.
1353 Before this you could only use some ECC ciphersuites with SHA1 only.
1356 *) Add protection against ECDSA timing attacks as mentioned in the paper
1357 by Billy Bob Brumley and Nicola Tuveri, see:
1359 http://eprint.iacr.org/2011/232.pdf
1361 [Billy Bob Brumley and Nicola Tuveri]
1363 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1365 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1366 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1368 *) Fix bug in string printing code: if *any* escaping is enabled we must
1369 escape the escape character (backslash) or the resulting string is
1373 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1375 *) Disable code workaround for ancient and obsolete Netscape browsers
1376 and servers: an attacker can use it in a ciphersuite downgrade attack.
1377 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1380 *) Fixed J-PAKE implementation error, originally discovered by
1381 Sebastien Martini, further info and confirmation from Stefan
1382 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1385 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1387 *) Fix extension code to avoid race conditions which can result in a buffer
1388 overrun vulnerability: resumed sessions must not be modified as they can
1389 be shared by multiple threads. CVE-2010-3864
1392 *) Fix WIN32 build system to correctly link an ENGINE directory into
1396 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1398 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1400 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1402 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1404 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1405 context. The operation can be customised via the ctrl mechanism in
1406 case ENGINEs want to include additional functionality.
1409 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1412 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1413 output hashes compatible with older versions of OpenSSL.
1414 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1416 *) Fix compression algorithm handling: if resuming a session use the
1417 compression algorithm of the resumed session instead of determining
1418 it from client hello again. Don't allow server to change algorithm.
1421 *) Add load_crls() function to apps tidying load_certs() too. Add option
1422 to verify utility to allow additional CRLs to be included.
1425 *) Update OCSP request code to permit adding custom headers to the request:
1426 some responders need this.
1429 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1431 [Julia Lawall <julia@diku.dk>]
1433 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1434 needlessly dereferenced structures, used obsolete functions and
1435 didn't handle all updated verify codes correctly.
1438 *) Disable MD2 in the default configuration.
1441 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1442 indicate the initial BIO being pushed or popped. This makes it possible
1443 to determine whether the BIO is the one explicitly called or as a result
1444 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1445 it handles reference counts correctly and doesn't zero out the I/O bio
1446 when it is not being explicitly popped. WARNING: applications which
1447 included workarounds for the old buggy behaviour will need to be modified
1448 or they could free up already freed BIOs.
1451 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1452 renaming to all platforms (within the 0.9.8 branch, this was
1453 done conditionally on Netware platforms to avoid a name clash).
1454 [Guenter <lists@gknw.net>]
1456 *) Add ECDHE and PSK support to DTLS.
1457 [Michael Tuexen <tuexen@fh-muenster.de>]
1459 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1463 *) Add "missing" function EVP_MD_flags() (without this the only way to
1464 retrieve a digest flags is by accessing the structure directly. Update
1465 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1466 or cipher is registered as in the "from" argument. Print out all
1467 registered digests in the dgst usage message instead of manually
1468 attempting to work them out.
1471 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1472 this allows the use of compression and extensions. Change default cipher
1473 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1474 by default unless an application cipher string requests it.
1477 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1478 key ids to find matching certificates and keys but some PKCS#12 files
1479 don't follow the (somewhat unwritten) rules and this strategy fails.
1480 Now just gather all certificates together and the first private key
1481 then look for the first certificate that matches the key.
1484 *) Support use of registered digest and cipher names for dgst and cipher
1485 commands instead of having to add each one as a special case. So now
1492 openssl dgst -sha256 foo
1494 and this works for ENGINE based algorithms too.
1498 *) Update Gost ENGINE to support parameter files.
1499 [Victor B. Wagner <vitus@cryptocom.ru>]
1501 *) Support GeneralizedTime in ca utility.
1502 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1504 *) Enhance the hash format used for certificate directory links. The new
1505 form uses the canonical encoding (meaning equivalent names will work
1506 even if they aren't identical) and uses SHA1 instead of MD5. This form
1507 is incompatible with the older format and as a result c_rehash should
1508 be used to rebuild symbolic links.
1511 *) Make PKCS#8 the default write format for private keys, replacing the
1512 traditional format. This form is standardised, more secure and doesn't
1513 include an implicit MD5 dependency.
1516 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1517 committed to OpenSSL should pass this lot as a minimum.
1520 *) Add session ticket override functionality for use by EAP-FAST.
1521 [Jouni Malinen <j@w1.fi>]
1523 *) Modify HMAC functions to return a value. Since these can be implemented
1524 in an ENGINE errors can occur.
1527 *) Type-checked OBJ_bsearch_ex.
1530 *) Type-checked OBJ_bsearch. Also some constification necessitated
1531 by type-checking. Still to come: TXT_DB, bsearch(?),
1532 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1536 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1537 seconds to a tm structure directly, instead of going through OS
1538 specific date routines. This avoids any issues with OS routines such
1539 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1540 and X509_time_adj_ex() to cover the extended range. The existing
1541 X509_time_adj() is still usable and will no longer have any date issues.
1544 *) Delta CRL support. New use deltas option which will attempt to locate
1545 and search any appropriate delta CRLs available.
1547 This work was sponsored by Google.
1550 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1551 code and add additional score elements. Validate alternate CRL paths
1552 as part of the CRL checking and indicate a new error "CRL path validation
1553 error" in this case. Applications wanting additional details can use
1554 the verify callback and check the new "parent" field. If this is not
1555 NULL CRL path validation is taking place. Existing applications wont
1556 see this because it requires extended CRL support which is off by
1559 This work was sponsored by Google.
1562 *) Support for freshest CRL extension.
1564 This work was sponsored by Google.
1567 *) Initial indirect CRL support. Currently only supported in the CRLs
1568 passed directly and not via lookup. Process certificate issuer
1569 CRL entry extension and lookup CRL entries by bother issuer name
1570 and serial number. Check and process CRL issuer entry in IDP extension.
1572 This work was sponsored by Google.
1575 *) Add support for distinct certificate and CRL paths. The CRL issuer
1576 certificate is validated separately in this case. Only enabled if
1577 an extended CRL support flag is set: this flag will enable additional
1578 CRL functionality in future.
1580 This work was sponsored by Google.
1583 *) Add support for policy mappings extension.
1585 This work was sponsored by Google.
1588 *) Fixes to pathlength constraint, self issued certificate handling,
1589 policy processing to align with RFC3280 and PKITS tests.
1591 This work was sponsored by Google.
1594 *) Support for name constraints certificate extension. DN, email, DNS
1595 and URI types are currently supported.
1597 This work was sponsored by Google.
1600 *) To cater for systems that provide a pointer-based thread ID rather
1601 than numeric, deprecate the current numeric thread ID mechanism and
1602 replace it with a structure and associated callback type. This
1603 mechanism allows a numeric "hash" to be extracted from a thread ID in
1604 either case, and on platforms where pointers are larger than 'long',
1605 mixing is done to help ensure the numeric 'hash' is usable even if it
1606 can't be guaranteed unique. The default mechanism is to use "&errno"
1607 as a pointer-based thread ID to distinguish between threads.
1609 Applications that want to provide their own thread IDs should now use
1610 CRYPTO_THREADID_set_callback() to register a callback that will call
1611 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1613 Note that ERR_remove_state() is now deprecated, because it is tied
1614 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1615 to free the current thread's error state should be replaced by
1616 ERR_remove_thread_state(NULL).
1618 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1619 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1620 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1621 application was previously providing a numeric thread callback that
1622 was inappropriate for distinguishing threads, then uniqueness might
1623 have been obtained with &errno that happened immediately in the
1624 intermediate development versions of OpenSSL; this is no longer the
1625 case, the numeric thread callback will now override the automatic use
1627 [Geoff Thorpe, with help from Bodo Moeller]
1629 *) Initial support for different CRL issuing certificates. This covers a
1630 simple case where the self issued certificates in the chain exist and
1631 the real CRL issuer is higher in the existing chain.
1633 This work was sponsored by Google.
1636 *) Removed effectively defunct crypto/store from the build.
1639 *) Revamp of STACK to provide stronger type-checking. Still to come:
1640 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1641 ASN1_STRING, CONF_VALUE.
1644 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1645 RAM on SSL connections. This option can save about 34k per idle SSL.
1648 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1649 STACK, TXT_DB, bsearch, qsort.
1652 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1653 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1654 support for data, signedData, compressedData, digestedData and
1655 encryptedData, envelopedData types included. Scripts to check against
1656 RFC4134 examples draft and interop and consistency checks of many
1657 content types and variants.
1660 *) Add options to enc utility to support use of zlib compression BIO.
1663 *) Extend mk1mf to support importing of options and assembly language
1664 files from Configure script, currently only included in VC-WIN32.
1665 The assembly language rules can now optionally generate the source
1666 files from the associated perl scripts.
1669 *) Implement remaining functionality needed to support GOST ciphersuites.
1670 Interop testing has been performed using CryptoPro implementations.
1671 [Victor B. Wagner <vitus@cryptocom.ru>]
1673 *) s390x assembler pack.
1676 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1680 *) Implement Opaque PRF Input TLS extension as specified in
1681 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1682 official specification yet and no extension type assignment by
1683 IANA exists, this extension (for now) will have to be explicitly
1684 enabled when building OpenSSL by providing the extension number
1685 to use. For example, specify an option
1687 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1689 to the "config" or "Configure" script to enable the extension,
1690 assuming extension number 0x9527 (which is a completely arbitrary
1691 and unofficial assignment based on the MD5 hash of the Internet
1692 Draft). Note that by doing so, you potentially lose
1693 interoperability with other TLS implementations since these might
1694 be using the same extension number for other purposes.
1696 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1697 opaque PRF input value to use in the handshake. This will create
1698 an interal copy of the length-'len' string at 'src', and will
1699 return non-zero for success.
1701 To get more control and flexibility, provide a callback function
1704 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1705 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1709 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1712 Callback function 'cb' will be called in handshakes, and is
1713 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1714 Argument 'arg' is for application purposes (the value as given to
1715 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1716 be provided to the callback function). The callback function
1717 has to return non-zero to report success: usually 1 to use opaque
1718 PRF input just if possible, or 2 to enforce use of the opaque PRF
1719 input. In the latter case, the library will abort the handshake
1720 if opaque PRF input is not successfully negotiated.
1722 Arguments 'peerinput' and 'len' given to the callback function
1723 will always be NULL and 0 in the case of a client. A server will
1724 see the client's opaque PRF input through these variables if
1725 available (NULL and 0 otherwise). Note that if the server
1726 provides an opaque PRF input, the length must be the same as the
1727 length of the client's opaque PRF input.
1729 Note that the callback function will only be called when creating
1730 a new session (session resumption can resume whatever was
1731 previously negotiated), and will not be called in SSL 2.0
1732 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1733 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1734 for applications that need to enforce opaque PRF input.
1738 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1741 [Victor B. Wagner <vitus@cryptocom.ru>]
1743 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1744 RFC4507bis. The encrypted ticket format is an encrypted encoded
1745 SSL_SESSION structure, that way new session features are automatically
1748 If a client application caches session in an SSL_SESSION structure
1749 support is transparent because tickets are now stored in the encoded
1752 The SSL_CTX structure automatically generates keys for ticket
1753 protection in servers so again support should be possible
1754 with no application modification.
1756 If a client or server wishes to disable RFC4507 support then the option
1757 SSL_OP_NO_TICKET can be set.
1759 Add a TLS extension debugging callback to allow the contents of any client
1760 or server extensions to be examined.
1762 This work was sponsored by Google.
1765 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1766 OpenSSL should now compile cleanly on gcc 4.2
1767 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1769 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1770 support including streaming MAC support: this is required for GOST
1771 ciphersuite support.
1772 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1774 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1775 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1776 to output in BER and PEM format.
1779 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1780 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1781 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1782 ENGINE support for HMAC keys which are unextractable. New -mac and
1783 -macopt options to dgst utility.
1786 *) New option -sigopt to dgst utility. Update dgst to use
1787 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1788 alternative signing paramaters such as X9.31 or PSS in the dgst
1792 *) Change ssl_cipher_apply_rule(), the internal function that does
1793 the work each time a ciphersuite string requests enabling
1794 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1795 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1796 the order of disabled ciphersuites such that those ciphersuites
1797 that most recently went from enabled to disabled not only stay
1798 in order with respect to each other, but also have higher priority
1799 than other disabled ciphersuites the next time ciphersuites are
1802 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1803 the same ciphersuites as with "HIGH" alone, but in a specific
1804 order where the PSK ciphersuites come first (since they are the
1805 most recently disabled ciphersuites when "HIGH" is parsed).
1807 Also, change ssl_create_cipher_list() (using this new
1808 funcionality) such that between otherwise identical
1809 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1813 *) Change ssl_create_cipher_list() so that it automatically
1814 arranges the ciphersuites in reasonable order before starting
1815 to process the rule string. Thus, the definition for "DEFAULT"
1816 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1817 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1818 This makes it much easier to arrive at a reasonable default order
1819 in applications for which anonymous ciphers are OK (meaning
1820 that you can't actually use DEFAULT).
1821 [Bodo Moeller; suggested by Victor Duchovni]
1823 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1824 processing) into multiple integers instead of setting
1825 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1826 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1827 (These masks as well as the individual bit definitions are hidden
1828 away into the non-exported interface ssl/ssl_locl.h, so this
1829 change to the definition of the SSL_CIPHER structure shouldn't
1830 affect applications.) This give us more bits for each of these
1831 categories, so there is no longer a need to coagulate AES128 and
1832 AES256 into a single algorithm bit, and to coagulate Camellia128
1833 and Camellia256 into a single algorithm bit, which has led to all
1836 Thus, among other things, the kludge introduced in 0.9.7m and
1837 0.9.8e for masking out AES256 independently of AES128 or masking
1838 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1840 With the change, we also introduce new ciphersuite aliases that
1841 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1845 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1846 Use the leftmost N bytes of the signature input if the input is
1847 larger than the prime q (with N being the size in bytes of q).
1850 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1851 it yet and it is largely untested.
1854 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1857 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1858 some compilers (gcc 4.2 and later) reject their use. Safestack is
1859 reimplemented. Update ASN1 to avoid use of legacy functions.
1862 *) Win32/64 targets are linked with Winsock2.
1865 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1866 to external functions. This can be used to increase CRL handling
1867 efficiency especially when CRLs are very large by (for example) storing
1868 the CRL revoked certificates in a database.
1871 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1872 new CRLs added to a directory can be used. New command line option
1873 -verify_return_error to s_client and s_server. This causes real errors
1874 to be returned by the verify callback instead of carrying on no matter
1875 what. This reflects the way a "real world" verify callback would behave.
1878 *) GOST engine, supporting several GOST algorithms and public key formats.
1879 Kindly donated by Cryptocom.
1882 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1883 partitioned by DP are handled but no indirect CRL or reason partitioning
1884 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1885 selected via a scoring technique which handles IDP and AKID in CRLs.
1888 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1889 will ultimately be used for all verify operations: this will remove the
1890 X509_STORE dependency on certificate verification and allow alternative
1891 lookup methods. X509_STORE based implementations of these two callbacks.
1894 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1895 Modify get_crl() to find a valid (unexpired) CRL if possible.
1898 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1899 this would be called X509_CRL_cmp() but that name is already used by
1900 a function that just compares CRL issuer names. Cache several CRL
1901 extensions in X509_CRL structure and cache CRLDP in X509.
1904 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1905 this maps equivalent X509_NAME structures into a consistent structure.
1906 Name comparison can then be performed rapidly using memcmp().
1909 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1913 *) Allow digests to supply their own micalg string for S/MIME type using
1914 the ctrl EVP_MD_CTRL_MICALG.
1917 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1918 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1919 ctrl. It can then customise the structure before and/or after signing
1923 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1924 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1925 to free up any added signature OIDs.
1928 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1929 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1930 digest and cipher tables. New options added to openssl utility:
1931 list-message-digest-algorithms and list-cipher-algorithms.
1934 *) Change the array representation of binary polynomials: the list
1935 of degrees of non-zero coefficients is now terminated with -1.
1936 Previously it was terminated with 0, which was also part of the
1937 value; thus, the array representation was not applicable to
1938 polynomials where t^0 has coefficient zero. This change makes
1939 the array representation useful in a more general context.
1942 *) Various modifications and fixes to SSL/TLS cipher string
1943 handling. For ECC, the code now distinguishes between fixed ECDH
1944 with RSA certificates on the one hand and with ECDSA certificates
1945 on the other hand, since these are separate ciphersuites. The
1946 unused code for Fortezza ciphersuites has been removed.
1948 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1949 (not "ECDHE"). For consistency with the code for DH
1950 certificates, use of ECDH certificates is now considered ECDH
1951 authentication, not RSA or ECDSA authentication (the latter is
1952 merely the CA's signing algorithm and not actively used in the
1955 The temporary ciphersuite alias "ECCdraft" is no longer
1956 available, and ECC ciphersuites are no longer excluded from "ALL"
1957 and "DEFAULT". The following aliases now exist for RFC 4492
1958 ciphersuites, most of these by analogy with the DH case:
1960 kECDHr - ECDH cert, signed with RSA
1961 kECDHe - ECDH cert, signed with ECDSA
1962 kECDH - ECDH cert (signed with either RSA or ECDSA)
1963 kEECDH - ephemeral ECDH
1964 ECDH - ECDH cert or ephemeral ECDH
1970 AECDH - anonymous ECDH
1971 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1975 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1976 Use correct micalg parameters depending on digest(s) in signed message.
1979 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1980 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1983 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1984 an engine to register a method. Add ENGINE lookups for methods and
1985 functional reference processing.
1988 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1989 EVP_{Sign,Verify}* which allow an application to customise the signature
1993 *) New -resign option to smime utility. This adds one or more signers
1994 to an existing PKCS#7 signedData structure. Also -md option to use an
1995 alternative message digest algorithm for signing.
1998 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1999 create PKCS7 structures containing multiple signers. Update smime
2000 application to support multiple signers.
2003 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2007 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2008 Reorganize PBE internals to lookup from a static table using NIDs,
2009 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2010 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2011 PRF which will be automatically used with PBES2.
2014 *) Replace the algorithm specific calls to generate keys in "req" with the
2018 *) Update PKCS#7 enveloped data routines to use new API. This is now
2019 supported by any public key method supporting the encrypt operation. A
2020 ctrl is added to allow the public key algorithm to examine or modify
2021 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2025 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2026 a default digest type to use. In most cases this will be SHA1 but some
2027 algorithms (such as GOST) need to specify an alternative digest. The
2028 return value indicates how strong the prefernce is 1 means optional and
2029 2 is mandatory (that is it is the only supported type). Modify
2030 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2031 use the default md. Update openssl utilities to use the default digest
2032 type for signing if it is not explicitly indicated.
2035 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2036 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2037 signing method from the key type. This effectively removes the link
2038 between digests and public key types.
2041 *) Add an OID cross reference table and utility functions. Its purpose is to
2042 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2043 rsaEncryption. This will allow some of the algorithm specific hackery
2044 needed to use the correct OID to be removed.
2047 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2048 structures for PKCS7_sign(). They are now set up by the relevant public
2052 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2055 *) Add support for key derivation (agreement) in the API, DH method and
2059 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2060 public and private key formats. As a side effect these add additional
2061 command line functionality not previously available: DSA signatures can be
2062 generated and verified using pkeyutl and DH key support and generation in
2067 [Oliver Tappe <zooey@hirschkaefer.de>]
2069 *) New make target "install_html_docs" installs HTML renditions of the
2071 [Oliver Tappe <zooey@hirschkaefer.de>]
2073 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
2074 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2075 support key and parameter generation and add initial key generation
2076 functionality for RSA.
2079 *) Add functions for main EVP_PKEY_method operations. The undocumented
2080 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2081 EVP_PKEY_{encrypt,decrypt}_old.
2084 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2085 key API, doesn't do much yet.
2088 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2089 public key algorithms. New option to openssl utility:
2090 "list-public-key-algorithms" to print out info.
2093 *) Implement the Supported Elliptic Curves Extension for
2094 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2097 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2098 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2101 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2102 utilities such as rsa, dsa, dsaparam etc except they process any key
2106 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2107 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2108 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2112 *) Initial support for pluggable public key ASN1.
2113 De-spaghettify the public key ASN1 handling. Move public and private
2114 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2115 algorithm specific handling to a single module within the relevant
2116 algorithm directory. Add functions to allow (near) opaque processing
2117 of public and private key structures.
2120 *) Implement the Supported Point Formats Extension for
2121 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2124 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2125 for the psk identity [hint] and the psk callback functions to the
2126 SSL_SESSION, SSL and SSL_CTX structure.
2129 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2133 SSL_CTX_use_psk_identity_hint
2134 SSL_get_psk_identity_hint
2135 SSL_get_psk_identity
2136 SSL_use_psk_identity_hint
2138 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2140 *) Add RFC 3161 compliant time stamp request creation, response generation
2141 and response verification functionality.
2142 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2144 *) Add initial support for TLS extensions, specifically for the server_name
2145 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2146 have new members for a host name. The SSL data structure has an
2147 additional member SSL_CTX *initial_ctx so that new sessions can be
2148 stored in that context to allow for session resumption, even after the
2149 SSL has been switched to a new SSL_CTX in reaction to a client's
2150 server_name extension.
2152 New functions (subject to change):
2154 SSL_get_servername()
2155 SSL_get_servername_type()
2158 New CTRL codes and macros (subject to change):
2160 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2161 - SSL_CTX_set_tlsext_servername_callback()
2162 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2163 - SSL_CTX_set_tlsext_servername_arg()
2164 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2166 openssl s_client has a new '-servername ...' option.
2168 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2169 '-key2 ...', '-servername_fatal' (subject to change). This allows
2170 testing the HostName extension for a specific single host name ('-cert'
2171 and '-key' remain fallbacks for handshakes without HostName
2172 negotiation). If the unrecogninzed_name alert has to be sent, this by
2173 default is a warning; it becomes fatal with the '-servername_fatal'
2176 [Peter Sylvester, Remy Allais, Christophe Renou]
2178 *) Whirlpool hash implementation is added.
2181 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2182 bn(64,32). Because of instruction set limitations it doesn't have
2183 any negative impact on performance. This was done mostly in order
2184 to make it possible to share assembler modules, such as bn_mul_mont
2185 implementations, between 32- and 64-bit builds without hassle.
2188 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2189 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2193 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2194 dedicated Montgomery multiplication procedure, is introduced.
2195 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2196 "64-bit" performance on certain 32-bit targets.
2199 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2200 in SSL structures. New SSL ctrl to set maximum send fragment size.
2201 Save memory by seeting the I/O buffer sizes dynamically instead of
2202 using the maximum available value.
2205 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2206 in addition to the text details.
2209 *) Very, very preliminary EXPERIMENTAL support for printing of general
2210 ASN1 structures. This currently produces rather ugly output and doesn't
2211 handle several customised structures at all.
2214 *) Integrated support for PVK file format and some related formats such
2215 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2216 these in the 'rsa' and 'dsa' utilities.
2219 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2222 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2223 place for the (very old) "NETSCAPE" format certificates which are now
2224 handled using new ASN1 code equivalents.
2227 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2228 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2229 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2232 *) Modify CRL distribution points extension code to print out previously
2233 unsupported fields. Enhance extension setting code to allow setting of
2237 *) Add print and set support for Issuing Distribution Point CRL extension.
2240 *) Change 'Configure' script to enable Camellia by default.
2243 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2245 *) When rejecting SSL/TLS records due to an incorrect version number, never
2246 update s->server with a new major version number. As of
2247 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2248 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2249 the previous behavior could result in a read attempt at NULL when
2250 receiving specific incorrect SSL/TLS records once record payload
2251 protection is active. (CVE-2010-0740)
2252 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2254 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2255 could be crashed if the relevant tables were not present (e.g. chrooted).
2256 [Tomas Hoger <thoger@redhat.com>]
2258 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2260 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2261 [Martin Olsson, Neel Mehta]
2263 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2264 accommodate for stack sorting, always a write lock!).
2267 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2268 excessive delays in the RAND_poll(): over a minute. As a workaround
2269 include a time check in the inner Heap32Next loop too.
2272 *) The code that handled flushing of data in SSL/TLS originally used the
2273 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2274 the problem outlined in PR#1949. The fix suggested there however can
2275 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2276 of Apache). So instead simplify the code to flush unconditionally.
2277 This should be fine since flushing with no data to flush is a no op.
2280 *) Handle TLS versions 2.0 and later properly and correctly use the
2281 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2282 off ancient servers have a habit of sticking around for a while...
2285 *) Modify compression code so it frees up structures without using the
2286 ex_data callbacks. This works around a problem where some applications
2287 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2288 restarting) then use compression (e.g. SSL with compression) later.
2289 This results in significant per-connection memory leaks and
2290 has caused some security issues including CVE-2008-1678 and
2294 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2295 change when encrypting or decrypting.
2298 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2299 connect and renegotiate with servers which do not support RI.
2300 Until RI is more widely deployed this option is enabled by default.
2303 *) Add "missing" ssl ctrls to clear options and mode.
2306 *) If client attempts to renegotiate and doesn't support RI respond with
2307 a no_renegotiation alert as required by RFC5746. Some renegotiating
2308 TLS clients will continue a connection gracefully when they receive
2309 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2310 waiting for a server hello which it will never receive. Now we treat a
2311 received no_renegotiation alert as a fatal error. This is because
2312 applications requesting a renegotiation might well expect it to succeed
2313 and would have no code in place to handle the server denying it so the
2314 only safe thing to do is to terminate the connection.
2317 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2318 peer supports secure renegotiation and 0 otherwise. Print out peer
2319 renegotiation support in s_client/s_server.
2322 *) Replace the highly broken and deprecated SPKAC certification method with
2323 the updated NID creation version. This should correctly handle UTF8.
2326 *) Implement RFC5746. Re-enable renegotiation but require the extension
2327 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2328 turns out to be a bad idea. It has been replaced by
2329 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2330 SSL_CTX_set_options(). This is really not recommended unless you
2331 know what you are doing.
2332 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2334 *) Fixes to stateless session resumption handling. Use initial_ctx when
2335 issuing and attempting to decrypt tickets in case it has changed during
2336 servername handling. Use a non-zero length session ID when attempting
2337 stateless session resumption: this makes it possible to determine if
2338 a resumption has occurred immediately after receiving server hello
2339 (several places in OpenSSL subtly assume this) instead of later in
2343 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2344 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2345 fixes for a few places where the return code is not checked
2347 [Julia Lawall <julia@diku.dk>]
2349 *) Add --strict-warnings option to Configure script to include devteam
2350 warnings in other configurations.
2353 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2354 makes it possible to install openssl libraries in locations which
2355 have names other than "lib", for example "/usr/lib64" which some
2357 [Steve Henson, based on patch from Jeremy Utley]
2359 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2360 X690 8.9.12 and can produce some misleading textual output of OIDs.
2361 [Steve Henson, reported by Dan Kaminsky]
2363 *) Delete MD2 from algorithm tables. This follows the recommendation in
2364 several standards that it is not used in new applications due to
2365 several cryptographic weaknesses. For binary compatibility reasons
2366 the MD2 API is still compiled in by default.
2369 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2373 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2374 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2376 [Guenter <lists@gknw.net>]
2378 *) Fix the server certificate chain building code to use X509_verify_cert(),
2379 it used to have an ad-hoc builder which was unable to cope with anything
2380 other than a simple chain.
2381 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2383 *) Don't check self signed certificate signatures in X509_verify_cert()
2384 by default (a flag can override this): it just wastes time without
2385 adding any security. As a useful side effect self signed root CAs
2386 with non-FIPS digests are now usable in FIPS mode.
2389 *) In dtls1_process_out_of_seq_message() the check if the current message
2390 is already buffered was missing. For every new message was memory
2391 allocated, allowing an attacker to perform an denial of service attack
2392 with sending out of seq handshake messages until there is no memory
2393 left. Additionally every future messege was buffered, even if the
2394 sequence number made no sense and would be part of another handshake.
2395 So only messages with sequence numbers less than 10 in advance will be
2396 buffered. (CVE-2009-1378)
2397 [Robin Seggelmann, discovered by Daniel Mentz]
2399 *) Records are buffered if they arrive with a future epoch to be
2400 processed after finishing the corresponding handshake. There is
2401 currently no limitation to this buffer allowing an attacker to perform
2402 a DOS attack with sending records with future epochs until there is no
2403 memory left. This patch adds the pqueue_size() function to detemine
2404 the size of a buffer and limits the record buffer to 100 entries.
2406 [Robin Seggelmann, discovered by Daniel Mentz]
2408 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2409 parent structure is freed. (CVE-2009-1379)
2412 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2413 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2416 [Ilya O. <vrghost@gmail.com>]
2418 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2420 *) Disable renegotiation completely - this fixes a severe security
2421 problem (CVE-2009-3555) at the cost of breaking all
2422 renegotiation. Renegotiation can be re-enabled by setting
2423 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2424 run-time. This is really not recommended unless you know what
2428 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2430 *) Don't set val to NULL when freeing up structures, it is freed up by
2431 underlying code. If sizeof(void *) > sizeof(long) this can result in
2432 zeroing past the valid field. (CVE-2009-0789)
2433 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2435 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2436 checked correctly. This would allow some invalid signed attributes to
2437 appear to verify correctly. (CVE-2009-0591)
2438 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2440 *) Reject UniversalString and BMPString types with invalid lengths. This
2441 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2442 a legal length. (CVE-2009-0590)
2445 *) Set S/MIME signing as the default purpose rather than setting it
2446 unconditionally. This allows applications to override it at the store
2450 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2451 to handle some structures.
2454 *) Improve efficiency of mem_gets: don't search whole buffer each time
2456 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2458 *) New -hex option for openssl rand.
2461 *) Print out UTF8String and NumericString when parsing ASN1.
2464 *) Support NumericString type for name components.
2467 *) Allow CC in the environment to override the automatically chosen
2468 compiler. Note that nothing is done to ensure flags work with the
2472 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2474 *) Properly check EVP_VerifyFinal() and similar return values
2476 [Ben Laurie, Bodo Moeller, Google Security Team]
2478 *) Enable TLS extensions by default.
2481 *) Allow the CHIL engine to be loaded, whether the application is
2482 multithreaded or not. (This does not release the developer from the
2483 obligation to set up the dynamic locking callbacks.)
2484 [Sander Temme <sander@temme.net>]
2486 *) Use correct exit code if there is an error in dgst command.
2487 [Steve Henson; problem pointed out by Roland Dirlewanger]
2489 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2490 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2493 *) Add experimental JPAKE support, including demo authentication in
2494 s_client and s_server.
2497 *) Set the comparison function in v3_addr_canonize().
2498 [Rob Austein <sra@hactrn.net>]
2500 *) Add support for XMPP STARTTLS in s_client.
2501 [Philip Paeps <philip@freebsd.org>]
2503 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2504 to ensure that even with this option, only ciphersuites in the
2505 server's preference list will be accepted. (Note that the option
2506 applies only when resuming a session, so the earlier behavior was
2507 just about the algorithm choice for symmetric cryptography.)
2510 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2512 *) Fix NULL pointer dereference if a DTLS server received
2513 ChangeCipherSpec as first record (CVE-2009-1386).
2516 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2517 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2520 *) The fix in 0.9.8c that supposedly got rid of unsafe
2521 double-checked locking was incomplete for RSA blinding,
2522 addressing just one layer of what turns out to have been
2523 doubly unsafe triple-checked locking.
2525 So now fix this for real by retiring the MONT_HELPER macro
2526 in crypto/rsa/rsa_eay.c.
2528 [Bodo Moeller; problem pointed out by Marius Schilder]
2530 *) Various precautionary measures:
2532 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2534 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2535 (NB: This would require knowledge of the secret session ticket key
2536 to exploit, in which case you'd be SOL either way.)
2538 - Change bn_nist.c so that it will properly handle input BIGNUMs
2539 outside the expected range.
2541 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2544 [Neel Mehta, Bodo Moeller]
2546 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2547 the load fails. Useful for distros.
2548 [Ben Laurie and the FreeBSD team]
2550 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2553 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2556 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2558 This work was sponsored by Logica.
2561 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2562 keystores. Support for SSL/TLS client authentication too.
2563 Not compiled unless enable-capieng specified to Configure.
2565 This work was sponsored by Logica.
2568 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2569 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2570 attribute creation routines such as certifcate requests and PKCS#12
2574 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2576 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2577 handshake which could lead to a cilent crash as found using the
2578 Codenomicon TLS test suite (CVE-2008-1672)
2579 [Steve Henson, Mark Cox]
2581 *) Fix double free in TLS server name extensions which could lead to
2582 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2585 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2587 Clear the error queue to ensure that error entries left from
2588 older function calls do not interfere with the correct operation.
2589 [Lutz Jaenicke, Erik de Castro Lopo]
2591 *) Remove root CA certificates of commercial CAs:
2593 The OpenSSL project does not recommend any specific CA and does not
2594 have any policy with respect to including or excluding any CA.
2595 Therefore it does not make any sense to ship an arbitrary selection
2596 of root CA certificates with the OpenSSL software.
2599 *) RSA OAEP patches to fix two separate invalid memory reads.
2600 The first one involves inputs when 'lzero' is greater than
2601 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2602 before the beginning of from). The second one involves inputs where
2603 the 'db' section contains nothing but zeroes (there is a one-byte
2604 invalid read after the end of 'db').
2605 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2607 *) Partial backport from 0.9.9-dev:
2609 Introduce bn_mul_mont (dedicated Montgomery multiplication
2610 procedure) as a candidate for BIGNUM assembler implementation.
2611 While 0.9.9-dev uses assembler for various architectures, only
2612 x86_64 is available by default here in the 0.9.8 branch, and
2613 32-bit x86 is available through a compile-time setting.
2615 To try the 32-bit x86 assembler implementation, use Configure
2616 option "enable-montasm" (which exists only for this backport).
2618 As "enable-montasm" for 32-bit x86 disclaims code stability
2619 anyway, in this constellation we activate additional code
2620 backported from 0.9.9-dev for further performance improvements,
2621 namely BN_from_montgomery_word. (To enable this otherwise,
2622 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2624 [Andy Polyakov (backport partially by Bodo Moeller)]
2626 *) Add TLS session ticket callback. This allows an application to set
2627 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2628 values. This is useful for key rollover for example where several key
2629 sets may exist with different names.
2632 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2633 This was broken until now in 0.9.8 releases, such that the only way
2634 a registered ENGINE could be used (assuming it initialises
2635 successfully on the host) was to explicitly set it as the default
2636 for the relevant algorithms. This is in contradiction with 0.9.7
2637 behaviour and the documentation. With this fix, when an ENGINE is
2638 registered into a given algorithm's table of implementations, the
2639 'uptodate' flag is reset so that auto-discovery will be used next
2640 time a new context for that algorithm attempts to select an
2642 [Ian Lister (tweaked by Geoff Thorpe)]
2644 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2645 implemention in the following ways:
2647 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2650 Lack of BER streaming support means one pass streaming processing is
2651 only supported if data is detached: setting the streaming flag is
2652 ignored for embedded content.
2654 CMS support is disabled by default and must be explicitly enabled
2655 with the enable-cms configuration option.
2658 *) Update the GMP engine glue to do direct copies between BIGNUM and
2659 mpz_t when openssl and GMP use the same limb size. Otherwise the
2660 existing "conversion via a text string export" trick is still used.
2661 [Paul Sheer <paulsheer@gmail.com>]
2663 *) Zlib compression BIO. This is a filter BIO which compressed and
2664 uncompresses any data passed through it.
2667 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2668 RFC3394 compatible AES key wrapping.
2671 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2672 sets string data without copying. X509_ALGOR_set0() and
2673 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2674 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2675 from an X509_ATTRIBUTE structure optionally checking it occurs only
2676 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2680 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2681 to get the expected BN_FLG_CONSTTIME behavior.
2682 [Bodo Moeller (Google)]
2686 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2687 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2688 - added some more tests to do_tests.pl
2689 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2690 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2691 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2692 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2693 - various changes to netware.pl to enable gcc-cross builds on Win32
2695 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2696 - various changes to fix missing prototype warnings
2697 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2698 - added AES, WHIRLPOOL and CPUID assembler code to build files
2699 - added missing AES assembler make rules to mk1mf.pl
2700 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2701 [Guenter Knauf <eflash@gmx.net>]
2703 *) Implement certificate status request TLS extension defined in RFC3546.
2704 A client can set the appropriate parameters and receive the encoded
2705 OCSP response via a callback. A server can query the supplied parameters
2706 and set the encoded OCSP response in the callback. Add simplified examples
2707 to s_client and s_server.
2710 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2712 *) Fix various bugs:
2713 + Binary incompatibility of ssl_ctx_st structure
2714 + DTLS interoperation with non-compliant servers
2715 + Don't call get_session_cb() without proposed session
2716 + Fix ia64 assembler code
2717 [Andy Polyakov, Steve Henson]
2719 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2721 *) DTLS Handshake overhaul. There were longstanding issues with
2722 OpenSSL DTLS implementation, which were making it impossible for
2723 RFC 4347 compliant client to communicate with OpenSSL server.
2724 Unfortunately just fixing these incompatibilities would "cut off"
2725 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2726 server keeps tolerating non RFC compliant syntax. The opposite is
2727 not true, 0.9.8f client can not communicate with earlier server.
2728 This update even addresses CVE-2007-4995.
2731 *) Changes to avoid need for function casts in OpenSSL: some compilers
2732 (gcc 4.2 and later) reject their use.
2733 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2736 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2737 RFC4507bis. The encrypted ticket format is an encrypted encoded
2738 SSL_SESSION structure, that way new session features are automatically
2741 If a client application caches session in an SSL_SESSION structure
2742 support is transparent because tickets are now stored in the encoded
2745 The SSL_CTX structure automatically generates keys for ticket
2746 protection in servers so again support should be possible
2747 with no application modification.
2749 If a client or server wishes to disable RFC4507 support then the option
2750 SSL_OP_NO_TICKET can be set.
2752 Add a TLS extension debugging callback to allow the contents of any client
2753 or server extensions to be examined.
2755 This work was sponsored by Google.
2758 *) Add initial support for TLS extensions, specifically for the server_name
2759 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2760 have new members for a host name. The SSL data structure has an
2761 additional member SSL_CTX *initial_ctx so that new sessions can be
2762 stored in that context to allow for session resumption, even after the
2763 SSL has been switched to a new SSL_CTX in reaction to a client's
2764 server_name extension.
2766 New functions (subject to change):
2768 SSL_get_servername()
2769 SSL_get_servername_type()
2772 New CTRL codes and macros (subject to change):
2774 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2775 - SSL_CTX_set_tlsext_servername_callback()
2776 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2777 - SSL_CTX_set_tlsext_servername_arg()
2778 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2780 openssl s_client has a new '-servername ...' option.
2782 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2783 '-key2 ...', '-servername_fatal' (subject to change). This allows
2784 testing the HostName extension for a specific single host name ('-cert'
2785 and '-key' remain fallbacks for handshakes without HostName
2786 negotiation). If the unrecogninzed_name alert has to be sent, this by
2787 default is a warning; it becomes fatal with the '-servername_fatal'
2790 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2792 *) Add AES and SSE2 assembly language support to VC++ build.
2795 *) Mitigate attack on final subtraction in Montgomery reduction.
2798 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2799 (which previously caused an internal error).
2802 *) Squeeze another 10% out of IGE mode when in != out.
2805 *) AES IGE mode speedup.
2806 [Dean Gaudet (Google)]
2808 *) Add the Korean symmetric 128-bit cipher SEED (see
2809 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2810 add SEED ciphersuites from RFC 4162:
2812 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2813 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2814 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2815 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2817 To minimize changes between patchlevels in the OpenSSL 0.9.8
2818 series, SEED remains excluded from compilation unless OpenSSL
2819 is configured with 'enable-seed'.
2820 [KISA, Bodo Moeller]
2822 *) Mitigate branch prediction attacks, which can be practical if a
2823 single processor is shared, allowing a spy process to extract
2824 information. For detailed background information, see
2825 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2826 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2827 and Necessary Software Countermeasures"). The core of the change
2828 are new versions BN_div_no_branch() and
2829 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2830 respectively, which are slower, but avoid the security-relevant
2831 conditional branches. These are automatically called by BN_div()
2832 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2833 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2834 remove a conditional branch.
2836 BN_FLG_CONSTTIME is the new name for the previous
2837 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2838 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2839 in the exponent causes BN_mod_exp_mont() to use the alternative
2840 implementation in BN_mod_exp_mont_consttime().) The old name
2841 remains as a deprecated alias.
2843 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2844 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2845 constant-time implementations for more than just exponentiation.
2846 Here too the old name is kept as a deprecated alias.
2848 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2849 the BN_BLINDING structure gets an independent copy of the
2850 modulus. This means that the previous "BIGNUM *m" argument to
2851 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2852 essentially becomes "const BIGNUM *m", although we can't actually
2853 change this in the header file before 0.9.9. It allows
2854 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2855 enable BN_FLG_CONSTTIME.
2857 [Matthew D Wood (Intel Corp)]
2859 *) In the SSL/TLS server implementation, be strict about session ID
2860 context matching (which matters if an application uses a single
2861 external cache for different purposes). Previously,
2862 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2863 set. This did ensure strict client verification, but meant that,
2864 with applications using a single external cache for quite
2865 different requirements, clients could circumvent ciphersuite
2866 restrictions for a given session ID context by starting a session
2867 in a different context.
2870 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2871 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2872 authentication-only ciphersuites.
2875 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2876 not complete and could lead to a possible single byte overflow
2877 (CVE-2007-5135) [Ben Laurie]
2879 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2881 *) Since AES128 and AES256 (and similarly Camellia128 and
2882 Camellia256) share a single mask bit in the logic of
2883 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2884 kludge to work properly if AES128 is available and AES256 isn't
2885 (or if Camellia128 is available and Camellia256 isn't).
2888 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2889 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2890 When a point or a seed is encoded in a BIT STRING, we need to
2891 prevent the removal of trailing zero bits to get the proper DER
2892 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2893 of a NamedBitList, for which trailing 0 bits need to be removed.)
2896 *) Have SSL/TLS server implementation tolerate "mismatched" record
2897 protocol version while receiving ClientHello even if the
2898 ClientHello is fragmented. (The server can't insist on the
2899 particular protocol version it has chosen before the ServerHello
2900 message has informed the client about his choice.)
2903 *) Add RFC 3779 support.
2904 [Rob Austein for ARIN, Ben Laurie]
2906 *) Load error codes if they are not already present instead of using a
2907 static variable. This allows them to be cleanly unloaded and reloaded.
2908 Improve header file function name parsing.
2911 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2912 or CAPABILITY handshake as required by RFCs.
2915 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2917 *) Introduce limits to prevent malicious keys being able to
2918 cause a denial of service. (CVE-2006-2940)
2919 [Steve Henson, Bodo Moeller]
2921 *) Fix ASN.1 parsing of certain invalid structures that can result
2922 in a denial of service. (CVE-2006-2937) [Steve Henson]
2924 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2925 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2927 *) Fix SSL client code which could crash if connecting to a
2928 malicious SSLv2 server. (CVE-2006-4343)
2929 [Tavis Ormandy and Will Drewry, Google Security Team]
2931 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2932 match only those. Before that, "AES256-SHA" would be interpreted
2933 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2934 the same strength classification in 0.9.7h) as we currently only
2935 have a single AES bit in the ciphersuite description bitmap.
2936 That change, however, also applied to ciphersuite strings such as
2937 "RC4-MD5" that intentionally matched multiple ciphersuites --
2938 namely, SSL 2.0 ciphersuites in addition to the more common ones
2939 from SSL 3.0/TLS 1.0.
2941 So we change the selection algorithm again: Naming an explicit
2942 ciphersuite selects this one ciphersuite, and any other similar
2943 ciphersuite (same bitmap) from *other* protocol versions.
2944 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2945 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2947 Since SSL 2.0 does not have any ciphersuites for which the
2948 128/256 bit distinction would be relevant, this works for now.
2949 The proper fix will be to use different bits for AES128 and
2950 AES256, which would have avoided the problems from the beginning;
2951 however, bits are scarce, so we can only do this in a new release
2952 (not just a patchlevel) when we can change the SSL_CIPHER
2953 definition to split the single 'unsigned long mask' bitmap into
2954 multiple values to extend the available space.
2958 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2960 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2961 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2963 *) Add AES IGE and biIGE modes.
2966 *) Change the Unix randomness entropy gathering to use poll() when
2967 possible instead of select(), since the latter has some
2968 undesirable limitations.
2969 [Darryl Miles via Richard Levitte and Bodo Moeller]
2971 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2972 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2973 cannot be implicitly activated as part of, e.g., the "AES" alias.
2974 However, please upgrade to OpenSSL 0.9.9[-dev] for
2975 non-experimental use of the ECC ciphersuites to get TLS extension
2976 support, which is required for curve and point format negotiation
2977 to avoid potential handshake problems.
2980 *) Disable rogue ciphersuites:
2982 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2983 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2984 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2986 The latter two were purportedly from
2987 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2990 Also deactivate the remaining ciphersuites from
2991 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2992 unofficial, and the ID has long expired.
2995 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2996 dual-core machines) and other potential thread-safety issues.
2999 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3000 versions), which is now available for royalty-free use
3001 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3002 Also, add Camellia TLS ciphersuites from RFC 4132.
3004 To minimize changes between patchlevels in the OpenSSL 0.9.8
3005 series, Camellia remains excluded from compilation unless OpenSSL
3006 is configured with 'enable-camellia'.
3009 *) Disable the padding bug check when compression is in use. The padding
3010 bug check assumes the first packet is of even length, this is not
3011 necessarily true if compresssion is enabled and can result in false
3012 positives causing handshake failure. The actual bug test is ancient
3013 code so it is hoped that implementations will either have fixed it by
3014 now or any which still have the bug do not support compression.
3017 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3019 *) When applying a cipher rule check to see if string match is an explicit
3020 cipher suite and only match that one cipher suite if it is.
3023 *) Link in manifests for VC++ if needed.
3024 [Austin Ziegler <halostatue@gmail.com>]
3026 *) Update support for ECC-based TLS ciphersuites according to
3027 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3028 TLS extensions, which are supported starting with the 0.9.9
3029 branch, not in the OpenSSL 0.9.8 branch).
3032 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3033 opaque EVP_CIPHER_CTX handling.
3036 *) Fixes and enhancements to zlib compression code. We now only use
3037 "zlib1.dll" and use the default __cdecl calling convention on Win32
3038 to conform with the standards mentioned here:
3039 http://www.zlib.net/DLL_FAQ.txt
3040 Static zlib linking now works on Windows and the new --with-zlib-include
3041 --with-zlib-lib options to Configure can be used to supply the location
3042 of the headers and library. Gracefully handle case where zlib library
3046 *) Several fixes and enhancements to the OID generation code. The old code
3047 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3048 handle numbers larger than ULONG_MAX, truncated printing and had a
3049 non standard OBJ_obj2txt() behaviour.
3052 *) Add support for building of engines under engine/ as shared libraries
3053 under VC++ build system.
3056 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3057 Hopefully, we will not see any false combination of paths any more.
3060 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3062 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3063 (part of SSL_OP_ALL). This option used to disable the
3064 countermeasure against man-in-the-middle protocol-version
3065 rollback in the SSL 2.0 server implementation, which is a bad
3066 idea. (CVE-2005-2969)
3068 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3069 for Information Security, National Institute of Advanced Industrial
3070 Science and Technology [AIST], Japan)]
3072 *) Add two function to clear and return the verify parameter flags.
3075 *) Keep cipherlists sorted in the source instead of sorting them at
3076 runtime, thus removing the need for a lock.
3079 *) Avoid some small subgroup attacks in Diffie-Hellman.
3080 [Nick Mathewson and Ben Laurie]
3082 *) Add functions for well-known primes.
3085 *) Extended Windows CE support.
3086 [Satoshi Nakamura and Andy Polyakov]
3088 *) Initialize SSL_METHOD structures at compile time instead of during
3089 runtime, thus removing the need for a lock.
3092 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3093 attempting to decrypt each encrypted key in turn. Add support to
3097 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3099 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3102 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3105 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3106 key into the same file any more.
3109 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3112 *) Add -utf8 command line and config file option to 'ca'.
3113 [Stefan <stf@udoma.org]
3115 *) Removed the macro des_crypt(), as it seems to conflict with some
3116 libraries. Use DES_crypt().
3119 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3120 involves renaming the source and generated shared-libs for
3121 both. The engines will accept the corrected or legacy ids
3122 ('ncipher' and '4758_cca' respectively) when binding. NB,
3123 this only applies when building 'shared'.
3124 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3126 *) Add attribute functions to EVP_PKEY structure. Modify
3127 PKCS12_create() to recognize a CSP name attribute and
3128 use it. Make -CSP option work again in pkcs12 utility.
3131 *) Add new functionality to the bn blinding code:
3132 - automatic re-creation of the BN_BLINDING parameters after
3133 a fixed number of uses (currently 32)
3134 - add new function for parameter creation
3135 - introduce flags to control the update behaviour of the
3136 BN_BLINDING parameters
3137 - hide BN_BLINDING structure
3138 Add a second BN_BLINDING slot to the RSA structure to improve
3139 performance when a single RSA object is shared among several
3143 *) Add support for DTLS.
3144 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3146 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3147 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3150 *) Remove buggy and incompletet DH cert support from
3151 ssl/ssl_rsa.c and ssl/s3_both.c
3154 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3155 the apps/openssl applications.
3158 *) Compile clean with "-Wall -Wmissing-prototypes
3159 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3160 DEBUG_SAFESTACK must also be set.
3163 *) Change ./Configure so that certain algorithms can be disabled by default.
3164 The new counterpiece to "no-xxx" is "enable-xxx".
3166 The patented RC5 and MDC2 algorithms will now be disabled unless
3167 "enable-rc5" and "enable-mdc2", respectively, are specified.
3169 (IDEA remains enabled despite being patented. This is because IDEA
3170 is frequently required for interoperability, and there is no license
3171 fee for non-commercial use. As before, "no-idea" can be used to
3172 avoid this algorithm.)
3176 *) Add processing of proxy certificates (see RFC 3820). This work was
3177 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3178 EGEE (Enabling Grids for E-science in Europe).
3181 *) RC4 performance overhaul on modern architectures/implementations, such
3182 as Intel P4, IA-64 and AMD64.
3185 *) New utility extract-section.pl. This can be used specify an alternative
3186 section number in a pod file instead of having to treat each file as
3187 a separate case in Makefile. This can be done by adding two lines to the
3190 =for comment openssl_section:XXX
3192 The blank line is mandatory.
3196 *) New arguments -certform, -keyform and -pass for s_client and s_server
3197 to allow alternative format key and certificate files and passphrase
3201 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3202 update associated structures and add various utility functions.
3204 Add new policy related verify parameters, include policy checking in
3205 standard verify code. Enhance 'smime' application with extra parameters
3206 to support policy checking and print out.
3209 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3210 Nehemiah processors. These extensions support AES encryption in hardware
3211 as well as RNG (though RNG support is currently disabled).
3212 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3214 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3217 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3218 [Andy Polyakov and a number of other people]
3220 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3221 implementation contributed by IBM.
3222 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3224 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3225 exponent rather than 'unsigned long'. There is a corresponding change to
3226 the new 'rsa_keygen' element of the RSA_METHOD structure.
3227 [Jelte Jansen, Geoff Thorpe]
3229 *) Functionality for creating the initial serial number file is now
3230 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3232 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3233 number file to 1, which is bound to cause problems. To avoid
3234 the problems while respecting compatibility between different 0.9.7
3235 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3236 CA.pl for serial number initialization. With the new release 0.9.8,
3237 we can fix the problem directly in the 'ca' utility.)
3240 *) Reduced header interdepencies by declaring more opaque objects in
3241 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3242 give fewer recursive includes, which could break lazy source code - so
3243 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3244 developers should define this symbol when building and using openssl to
3245 ensure they track the recommended behaviour, interfaces, [etc], but
3246 backwards-compatible behaviour prevails when this isn't defined.
3249 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3252 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3253 This will generate a random key of the appropriate length based on the
3254 cipher context. The EVP_CIPHER can provide its own random key generation
3255 routine to support keys of a specific form. This is used in the des and
3256 3des routines to generate a key of the correct parity. Update S/MIME
3257 code to use new functions and hence generate correct parity DES keys.
3258 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3259 valid (weak or incorrect parity).
3262 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3263 as looking them up. This is useful when the verified structure may contain
3264 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3265 present unless the new PKCS7_NO_CRL flag is asserted.
3268 *) Extend ASN1 oid configuration module. It now additionally accepts the
3271 shortName = some long name, 1.2.3.4
3274 *) Reimplemented the BN_CTX implementation. There is now no more static
3275 limitation on the number of variables it can handle nor the depth of the
3276 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3277 information can now expand as required, and rather than having a single
3278 static array of bignums, BN_CTX now uses a linked-list of such arrays
3279 allowing it to expand on demand whilst maintaining the usefulness of
3280 BN_CTX's "bundling".
3283 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3284 to allow all RSA operations to function using a single BN_CTX.
3287 *) Preliminary support for certificate policy evaluation and checking. This
3288 is initially intended to pass the tests outlined in "Conformance Testing
3289 of Relying Party Client Certificate Path Processing Logic" v1.07.
3292 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3293 remained unused and not that useful. A variety of other little bignum
3294 tweaks and fixes have also been made continuing on from the audit (see
3298 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3299 associated ASN1, EVP and SSL functions and old ASN1 macros.
3302 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3303 and this should never fail. So the return value from the use of
3304 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3305 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3308 *) BN_CTX_get() should return zero-valued bignums, providing the same
3309 initialised value as BN_new().
3310 [Geoff Thorpe, suggested by Ulf Möller]
3312 *) Support for inhibitAnyPolicy certificate extension.
3315 *) An audit of the BIGNUM code is underway, for which debugging code is
3316 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3317 is considered valid when processing BIGNUMs, and causes execution to
3318 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3319 further steps are taken to deliberately pollute unused data in BIGNUM
3320 structures to try and expose faulty code further on. For now, openssl will
3321 (in its default mode of operation) continue to tolerate the inconsistent
3322 forms that it has tolerated in the past, but authors and packagers should
3323 consider trying openssl and their own applications when compiled with
3324 these debugging symbols defined. It will help highlight potential bugs in
3325 their own code, and will improve the test coverage for OpenSSL itself. At
3326 some point, these tighter rules will become openssl's default to improve
3327 maintainability, though the assert()s and other overheads will remain only
3328 in debugging configurations. See bn.h for more details.
3329 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3331 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3332 that can only be obtained through BN_CTX_new() (which implicitly
3333 initialises it). The presence of this function only made it possible
3334 to overwrite an existing structure (and cause memory leaks).
3337 *) Because of the callback-based approach for implementing LHASH as a
3338 template type, lh_insert() adds opaque objects to hash-tables and
3339 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3340 to clean up those corresponding objects before destroying the hash table
3341 (and losing the object pointers). So some over-zealous constifications in
3342 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3343 objects as "const" and the lh_doall[_arg] callback wrappers are not
3344 prototyped to have "const" restrictions on the object pointers they are
3345 given (and so aren't required to cast them away any more).
3348 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3349 (speed) prefers to use its own implementation. The two implementations
3350 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3351 its object type properly exposed (MS_TM) instead of casting to/from "char
3352 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3353 aren't necessarily the greatest nomenclatures - but this is what was used
3354 internally to the implementation so I've used that for now.
3357 *) Ensure that deprecated functions do not get compiled when
3358 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3359 the self-tests were still using deprecated key-generation functions so
3360 these have been updated also.
3363 *) Reorganise PKCS#7 code to separate the digest location functionality
3364 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3365 New function PKCS7_set_digest() to set the digest type for PKCS#7
3366 digestedData type. Add additional code to correctly generate the
3367 digestedData type and add support for this type in PKCS7 initialization
3371 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3372 structure of type "other".
3375 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3376 sure the loop does correctly stop and breaking ("division by zero")
3377 modulus operations are not performed. The (pre-generated) prime
3378 table crypto/bn/bn_prime.h was already correct, but it could not be
3379 re-generated on some platforms because of the "division by zero"
3380 situation in the script.
3381 [Ralf S. Engelschall]
3383 *) Update support for ECC-based TLS ciphersuites according to
3384 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3385 SHA-1 now is only used for "small" curves (where the
3386 representation of a field element takes up to 24 bytes); for
3387 larger curves, the field element resulting from ECDH is directly
3388 used as premaster secret.
3389 [Douglas Stebila (Sun Microsystems Laboratories)]
3391 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3392 curve secp160r1 to the tests.
3393 [Douglas Stebila (Sun Microsystems Laboratories)]
3395 *) Add the possibility to load symbols globally with DSO.
3396 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3398 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3399 control of the error stack.
3402 *) Add support for STORE in ENGINE.
3405 *) Add the STORE type. The intention is to provide a common interface
3406 to certificate and key stores, be they simple file-based stores, or
3407 HSM-type store, or LDAP stores, or...
3408 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3411 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3412 pass a list of arguments to any function as well as provide a way
3413 for a function to pass data back to the caller.
3416 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3417 works like BUF_strdup() but can be used to duplicate a portion of
3418 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3422 *) Add the function sk_find_ex() which works like sk_find(), but will
3423 return an index to an element even if an exact match couldn't be
3424 found. The index is guaranteed to point at the element where the
3425 searched-for key would be inserted to preserve sorting order.
3428 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3429 takes an extra flags argument for optional functionality. Currently,
3430 the following flags are defined:
3432 OBJ_BSEARCH_VALUE_ON_NOMATCH
3433 This one gets OBJ_bsearch_ex() to return a pointer to the first
3434 element where the comparing function returns a negative or zero
3437 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3438 This one gets OBJ_bsearch_ex() to return a pointer to the first
3439 element where the comparing function returns zero. This is useful
3440 if there are more than one element where the comparing function
3444 *) Make it possible to create self-signed certificates with 'openssl ca'
3445 in such a way that the self-signed certificate becomes part of the
3446 CA database and uses the same mechanisms for serial number generation
3447 as all other certificate signing. The new flag '-selfsign' enables
3448 this functionality. Adapt CA.sh and CA.pl.in.
3451 *) Add functionality to check the public key of a certificate request
3452 against a given private. This is useful to check that a certificate
3453 request can be signed by that key (self-signing).
3456 *) Make it possible to have multiple active certificates with the same
3457 subject in the CA index file. This is done only if the keyword
3458 'unique_subject' is set to 'no' in the main CA section (default
3459 if 'CA_default') of the configuration file. The value is saved
3460 with the database itself in a separate index attribute file,
3461 named like the index file with '.attr' appended to the name.
3464 *) Generate muti valued AVAs using '+' notation in config files for
3468 *) Support for nameConstraints certificate extension.
3471 *) Support for policyConstraints certificate extension.
3474 *) Support for policyMappings certificate extension.
3477 *) Make sure the default DSA_METHOD implementation only uses its
3478 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3479 and change its own handlers to be NULL so as to remove unnecessary
3480 indirection. This lets alternative implementations fallback to the
3481 default implementation more easily.
3484 *) Support for directoryName in GeneralName related extensions
3488 *) Make it possible to link applications using Makefile.shared.
3489 Make that possible even when linking against static libraries!
3492 *) Support for single pass processing for S/MIME signing. This now
3493 means that S/MIME signing can be done from a pipe, in addition
3494 cleartext signing (multipart/signed type) is effectively streaming
3495 and the signed data does not need to be all held in memory.
3497 This is done with a new flag PKCS7_STREAM. When this flag is set
3498 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3499 is done after the data is output (and digests calculated) in
3500 SMIME_write_PKCS7().
3503 *) Add full support for -rpath/-R, both in shared libraries and
3504 applications, at least on the platforms where it's known how
3508 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3509 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3510 will now compute a table of multiples of the generator that
3511 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3512 faster (notably in the case of a single point multiplication,
3513 scalar * generator).
3514 [Nils Larsch, Bodo Moeller]
3516 *) IPv6 support for certificate extensions. The various extensions
3517 which use the IP:a.b.c.d can now take IPv6 addresses using the
3518 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3522 *) Added an ENGINE that implements RSA by performing private key
3523 exponentiations with the GMP library. The conversions to and from
3524 GMP's mpz_t format aren't optimised nor are any montgomery forms
3525 cached, and on x86 it appears OpenSSL's own performance has caught up.
3526 However there are likely to be other architectures where GMP could
3527 provide a boost. This ENGINE is not built in by default, but it can be
3528 specified at Configure time and should be accompanied by the necessary
3529 linker additions, eg;
3530 ./config -DOPENSSL_USE_GMP -lgmp
3533 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3534 testing availability of engines with "-t" - the old behaviour is
3535 produced by increasing the feature's verbosity with "-tt".
3538 *) ECDSA routines: under certain error conditions uninitialized BN objects
3539 could be freed. Solution: make sure initialization is performed early
3540 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3544 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3545 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3546 software implementations. For DSA and DH, parameter generation can
3547 also be overriden by providing the appropriate method callbacks.
3550 *) Change the "progress" mechanism used in key-generation and
3551 primality testing to functions that take a new BN_GENCB pointer in
3552 place of callback/argument pairs. The new API functions have "_ex"
3553 postfixes and the older functions are reimplemented as wrappers for
3554 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3555 declarations of the old functions to help (graceful) attempts to
3556 migrate to the new functions. Also, the new key-generation API
3557 functions operate on a caller-supplied key-structure and return
3558 success/failure rather than returning a key or NULL - this is to
3559 help make "keygen" another member function of RSA_METHOD etc.
3561 Example for using the new callback interface:
3563 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3567 BN_GENCB_set(&my_cb, my_callback, my_arg);
3569 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3570 /* For the meaning of a, b in calls to my_callback(), see the
3571 * documentation of the function that calls the callback.
3572 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3573 * my_callback should return 1 if it wants BN_is_prime_ex()
3574 * to continue, or 0 to stop.
3579 *) Change the ZLIB compression method to be stateful, and make it
3580 available to TLS with the number defined in
3581 draft-ietf-tls-compression-04.txt.
3584 *) Add the ASN.1 structures and functions for CertificatePair, which
3585 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3587 CertificatePair ::= SEQUENCE {
3588 forward [0] Certificate OPTIONAL,
3589 reverse [1] Certificate OPTIONAL,
3590 -- at least one of the pair shall be present -- }
3592 Also implement the PEM functions to read and write certificate
3593 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3595 This needed to be defined, mostly for the sake of the LDAP
3596 attribute crossCertificatePair, but may prove useful elsewhere as
3600 *) Make it possible to inhibit symlinking of shared libraries in
3601 Makefile.shared, for Cygwin's sake.
3604 *) Extend the BIGNUM API by creating a function
3605 void BN_set_negative(BIGNUM *a, int neg);
3606 and a macro that behave like
3607 int BN_is_negative(const BIGNUM *a);
3609 to avoid the need to access 'a->neg' directly in applications.
3612 *) Implement fast modular reduction for pseudo-Mersenne primes
3613 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3614 EC_GROUP_new_curve_GFp() will now automatically use this
3616 [Nils Larsch <nla@trustcenter.de>]
3618 *) Add new lock type (CRYPTO_LOCK_BN).
3621 *) Change the ENGINE framework to automatically load engines
3622 dynamically from specific directories unless they could be
3623 found to already be built in or loaded. Move all the
3624 current engines except for the cryptodev one to a new
3626 The engines in engines/ are built as shared libraries if
3627 the "shared" options was given to ./Configure or ./config.
3628 Otherwise, they are inserted in libcrypto.a.
3629 /usr/local/ssl/engines is the default directory for dynamic
3630 engines, but that can be overriden at configure time through
3631 the usual use of --prefix and/or --openssldir, and at run
3632 time with the environment variable OPENSSL_ENGINES.
3633 [Geoff Thorpe and Richard Levitte]
3635 *) Add Makefile.shared, a helper makefile to build shared
3636 libraries. Addapt Makefile.org.
3639 *) Add version info to Win32 DLLs.
3640 [Peter 'Luna' Runestig" <peter@runestig.com>]
3642 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3643 can be added using this API to created arbitrary PKCS#12
3644 files while avoiding the low level API.
3646 New options to PKCS12_create(), key or cert can be NULL and
3647 will then be omitted from the output file. The encryption
3648 algorithm NIDs can be set to -1 for no encryption, the mac
3649 iteration count can be set to 0 to omit the mac.
3651 Enhance pkcs12 utility by making the -nokeys and -nocerts
3652 options work when creating a PKCS#12 file. New option -nomac
3653 to omit the mac, NONE can be set for an encryption algorithm.
3654 New code is modified to use the enhanced PKCS12_create()
3655 instead of the low level API.
3658 *) Extend ASN1 encoder to support indefinite length constructed
3659 encoding. This can output sequences tags and octet strings in
3660 this form. Modify pk7_asn1.c to support indefinite length
3661 encoding. This is experimental and needs additional code to
3662 be useful, such as an ASN1 bio and some enhanced streaming
3665 Extend template encode functionality so that tagging is passed
3666 down to the template encoder.
3669 *) Let 'openssl req' fail if an argument to '-newkey' is not
3670 recognized instead of using RSA as a default.
3673 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3674 As these are not official, they are not included in "ALL";
3675 the "ECCdraft" ciphersuite group alias can be used to select them.
3676 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3678 *) Add ECDH engine support.
3679 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3681 *) Add ECDH in new directory crypto/ecdh/.
3682 [Douglas Stebila (Sun Microsystems Laboratories)]
3684 *) Let BN_rand_range() abort with an error after 100 iterations
3685 without success (which indicates a broken PRNG).
3688 *) Change BN_mod_sqrt() so that it verifies that the input value
3689 is really the square of the return value. (Previously,
3690 BN_mod_sqrt would show GIGO behaviour.)
3693 *) Add named elliptic curves over binary fields from X9.62, SECG,
3694 and WAP/WTLS; add OIDs that were still missing.
3696 [Sheueling Chang Shantz and Douglas Stebila
3697 (Sun Microsystems Laboratories)]
3699 *) Extend the EC library for elliptic curves over binary fields
3700 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3703 EC_GF2m_simple_method
3707 EC_GROUP_new_curve_GF2m
3708 EC_GROUP_set_curve_GF2m
3709 EC_GROUP_get_curve_GF2m
3710 EC_POINT_set_affine_coordinates_GF2m
3711 EC_POINT_get_affine_coordinates_GF2m
3712 EC_POINT_set_compressed_coordinates_GF2m
3714 Point compression for binary fields is disabled by default for
3715 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3718 As binary polynomials are represented as BIGNUMs, various members
3719 of the EC_GROUP and EC_POINT data structures can be shared
3720 between the implementations for prime fields and binary fields;
3721 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3722 are essentially identical to their ..._GFp counterparts.
3723 (For simplicity, the '..._GFp' prefix has been dropped from
3724 various internal method names.)
3726 An internal 'field_div' method (similar to 'field_mul' and
3727 'field_sqr') has been added; this is used only for binary fields.
3729 [Sheueling Chang Shantz and Douglas Stebila
3730 (Sun Microsystems Laboratories)]
3732 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3733 through methods ('mul', 'precompute_mult').
3735 The generic implementations (now internally called 'ec_wNAF_mul'
3736 and 'ec_wNAF_precomputed_mult') remain the default if these
3737 methods are undefined.
3739 [Sheueling Chang Shantz and Douglas Stebila
3740 (Sun Microsystems Laboratories)]
3742 *) New function EC_GROUP_get_degree, which is defined through
3743 EC_METHOD. For curves over prime fields, this returns the bit
3744 length of the modulus.
3746 [Sheueling Chang Shantz and Douglas Stebila
3747 (Sun Microsystems Laboratories)]
3749 *) New functions EC_GROUP_dup, EC_POINT_dup.
3750 (These simply call ..._new and ..._copy).
3752 [Sheueling Chang Shantz and Douglas Stebila
3753 (Sun Microsystems Laboratories)]
3755 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3756 Polynomials are represented as BIGNUMs (where the sign bit is not
3757 used) in the following functions [macros]:
3760 BN_GF2m_sub [= BN_GF2m_add]
3761 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3762 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3763 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3765 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3766 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3767 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3768 BN_GF2m_cmp [= BN_ucmp]
3770 (Note that only the 'mod' functions are actually for fields GF(2^m).
3771 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3773 For some functions, an the irreducible polynomial defining a
3774 field can be given as an 'unsigned int[]' with strictly
3775 decreasing elements giving the indices of those bits that are set;
3776 i.e., p[] represents the polynomial
3777 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3779 p[0] > p[1] > ... > p[k] = 0.
3780 This applies to the following functions:
3785 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3786 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3788 BN_GF2m_mod_sqrt_arr
3789 BN_GF2m_mod_solve_quad_arr
3793 Conversion can be performed by the following functions:
3798 bntest.c has additional tests for binary polynomial arithmetic.
3800 Two implementations for BN_GF2m_mod_div() are available.
3801 The default algorithm simply uses BN_GF2m_mod_inv() and
3802 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3803 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3804 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3806 [Sheueling Chang Shantz and Douglas Stebila
3807 (Sun Microsystems Laboratories)]
3809 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3810 functionality is disabled at compile-time.
3811 [Douglas Stebila <douglas.stebila@sun.com>]
3813 *) Change default behaviour of 'openssl asn1parse' so that more
3814 information is visible when viewing, e.g., a certificate:
3816 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3817 mode the content of non-printable OCTET STRINGs is output in a
3818 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3819 avoid the appearance of a printable string.
3820 [Nils Larsch <nla@trustcenter.de>]
3822 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3824 EC_GROUP_set_asn1_flag()
3825 EC_GROUP_get_asn1_flag()
3826 EC_GROUP_set_point_conversion_form()
3827 EC_GROUP_get_point_conversion_form()
3828 These control ASN1 encoding details:
3829 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3830 has been set to OPENSSL_EC_NAMED_CURVE.
3831 - Points are encoded in uncompressed form by default; options for
3832 asn1_for are as for point2oct, namely
3833 POINT_CONVERSION_COMPRESSED
3834 POINT_CONVERSION_UNCOMPRESSED
3835 POINT_CONVERSION_HYBRID
3837 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3840 EC_GROUP_get0_seed()
3841 EC_GROUP_get_seed_len()
3842 This is used only for ASN1 purposes (so far).
3843 [Nils Larsch <nla@trustcenter.de>]
3845 *) Add 'field_type' member to EC_METHOD, which holds the NID
3846 of the appropriate field type OID. The new function
3847 EC_METHOD_get_field_type() returns this value.
3848 [Nils Larsch <nla@trustcenter.de>]
3853 EC_POINT_point2hex()
3854 EC_POINT_hex2point()
3855 providing useful interfaces to EC_POINT_point2oct() and
3856 EC_POINT_oct2point().
3857 [Nils Larsch <nla@trustcenter.de>]
3859 *) Change internals of the EC library so that the functions
3860 EC_GROUP_set_generator()
3861 EC_GROUP_get_generator()
3862 EC_GROUP_get_order()
3863 EC_GROUP_get_cofactor()
3864 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3865 to methods, which would lead to unnecessary code duplication when
3866 adding different types of curves.
3867 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3869 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3870 arithmetic, and such that modified wNAFs are generated
3871 (which avoid length expansion in many cases).
3874 *) Add a function EC_GROUP_check_discriminant() (defined via
3875 EC_METHOD) that verifies that the curve discriminant is non-zero.
3877 Add a function EC_GROUP_check() that makes some sanity tests
3878 on a EC_GROUP, its generator and order. This includes
3879 EC_GROUP_check_discriminant().
3880 [Nils Larsch <nla@trustcenter.de>]
3882 *) Add ECDSA in new directory crypto/ecdsa/.
3884 Add applications 'openssl ecparam' and 'openssl ecdsa'
3885 (these are based on 'openssl dsaparam' and 'openssl dsa').
3887 ECDSA support is also included in various other files across the
3888 library. Most notably,
3889 - 'openssl req' now has a '-newkey ecdsa:file' option;
3890 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3891 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3892 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3893 them suitable for ECDSA where domain parameters must be
3894 extracted before the specific public key;
3895 - ECDSA engine support has been added.
3896 [Nils Larsch <nla@trustcenter.de>]
3898 *) Include some named elliptic curves, and add OIDs from X9.62,
3899 SECG, and WAP/WTLS. Each curve can be obtained from the new
3901 EC_GROUP_new_by_curve_name(),
3902 and the list of available named curves can be obtained with
3903 EC_get_builtin_curves().
3904 Also add a 'curve_name' member to EC_GROUP objects, which can be
3906 EC_GROUP_set_curve_name()
3907 EC_GROUP_get_curve_name()
3908 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3910 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3911 was actually never needed) and in BN_mul(). The removal in BN_mul()
3912 required a small change in bn_mul_part_recursive() and the addition
3913 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3914 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3915 bn_sub_words() and bn_add_words() except they take arrays with
3919 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3921 *) Cleanse PEM buffers before freeing them since they may contain
3923 [Benjamin Bennett <ben@psc.edu>]
3925 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3926 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3927 authentication-only ciphersuites.
3930 *) Since AES128 and AES256 share a single mask bit in the logic of
3931 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3932 kludge to work properly if AES128 is available and AES256 isn't.
3935 *) Expand security boundary to match 1.1.1 module.
3938 *) Remove redundant features: hash file source, editing of test vectors
3939 modify fipsld to use external fips_premain.c signature.
3942 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3943 run algorithm test programs.
3946 *) Make algorithm test programs more tolerant of whitespace.
3949 *) Have SSL/TLS server implementation tolerate "mismatched" record
3950 protocol version while receiving ClientHello even if the
3951 ClientHello is fragmented. (The server can't insist on the
3952 particular protocol version it has chosen before the ServerHello
3953 message has informed the client about his choice.)
3956 *) Load error codes if they are not already present instead of using a
3957 static variable. This allows them to be cleanly unloaded and reloaded.
3960 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3962 *) Introduce limits to prevent malicious keys being able to
3963 cause a denial of service. (CVE-2006-2940)
3964 [Steve Henson, Bodo Moeller]
3966 *) Fix ASN.1 parsing of certain invalid structures that can result
3967 in a denial of service. (CVE-2006-2937) [Steve Henson]
3969 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3970 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3972 *) Fix SSL client code which could crash if connecting to a
3973 malicious SSLv2 server. (CVE-2006-4343)
3974 [Tavis Ormandy and Will Drewry, Google Security Team]
3976 *) Change ciphersuite string processing so that an explicit
3977 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3978 will no longer include "AES128-SHA"), and any other similar
3979 ciphersuite (same bitmap) from *other* protocol versions (so that
3980 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3981 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3982 changes from 0.9.8b and 0.9.8d.
3985 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3987 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3988 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3990 *) Change the Unix randomness entropy gathering to use poll() when
3991 possible instead of select(), since the latter has some
3992 undesirable limitations.
3993 [Darryl Miles via Richard Levitte and Bodo Moeller]
3995 *) Disable rogue ciphersuites:
3997 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3998 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3999 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4001 The latter two were purportedly from
4002 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4005 Also deactive the remaining ciphersuites from
4006 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4007 unofficial, and the ID has long expired.
4010 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4011 dual-core machines) and other potential thread-safety issues.
4014 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4016 *) Adapt fipsld and the build system to link against the validated FIPS
4017 module in FIPS mode.
4020 *) Fixes for VC++ 2005 build under Windows.
4023 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4024 from a Windows bash shell such as MSYS. It is autodetected from the
4025 "config" script when run from a VC++ environment. Modify standard VC++
4026 build to use fipscanister.o from the GNU make build.
4029 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4031 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4032 The value now differs depending on if you build for FIPS or not.
4033 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4034 safely run with a non-FIPSed libcrypto, as it may crash because of
4035 the difference induced by this change.
4038 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4040 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4041 (part of SSL_OP_ALL). This option used to disable the
4042 countermeasure against man-in-the-middle protocol-version
4043 rollback in the SSL 2.0 server implementation, which is a bad
4044 idea. (CVE-2005-2969)
4046 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4047 for Information Security, National Institute of Advanced Industrial
4048 Science and Technology [AIST], Japan)]
4050 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4051 mainly for FIPS compliance and not fully integrated at this stage.
4054 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4055 the exponentiation using a fixed-length exponent. (Otherwise,
4056 the information leaked through timing could expose the secret key
4057 after many signatures; cf. Bleichenbacher's attack on DSA with
4061 *) Make a new fixed-window mod_exp implementation the default for
4062 RSA, DSA, and DH private-key operations so that the sequence of
4063 squares and multiplies and the memory access pattern are
4064 independent of the particular secret key. This will mitigate
4065 cache-timing and potential related attacks.
4067 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4068 and this is automatically used by BN_mod_exp_mont() if the new flag
4069 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4070 will use this BN flag for private exponents unless the flag
4071 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4072 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4074 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4076 *) Change the client implementation for SSLv23_method() and
4077 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4078 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4079 (Previously, the SSL 2.0 backwards compatible Client Hello
4080 message format would be used even with SSL_OP_NO_SSLv2.)
4083 *) Add support for smime-type MIME parameter in S/MIME messages which some
4087 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4088 a threadsafe manner. Modify rsa code to use new function and add calls
4089 to dsa and dh code (which had race conditions before).
4092 *) Include the fixed error library code in the C error file definitions
4093 instead of fixing them up at runtime. This keeps the error code
4094 structures constant.
4097 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4099 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4102 *) Fixes for newer kerberos headers. NB: the casts are needed because
4103 the 'length' field is signed on one version and unsigned on another
4104 with no (?) obvious way to tell the difference, without these VC++
4105 complains. Also the "definition" of FAR (blank) is no longer included
4106 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4107 some needed definitions.
4110 *) Undo Cygwin change.
4113 *) Added support for proxy certificates according to RFC 3820.
4114 Because they may be a security thread to unaware applications,
4115 they must be explicitely allowed in run-time. See
4116 docs/HOWTO/proxy_certificates.txt for further information.
4119 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4121 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4122 server and client random values. Previously
4123 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4124 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4126 This change has negligible security impact because:
4128 1. Server and client random values still have 24 bytes of pseudo random
4131 2. Server and client random values are sent in the clear in the initial
4134 3. The master secret is derived using the premaster secret (48 bytes in
4135 size for static RSA ciphersuites) as well as client server and random
4138 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4141 [Stephen Henson, reported by UK NISCC]
4143 *) Use Windows randomness collection on Cygwin.
4146 *) Fix hang in EGD/PRNGD query when communication socket is closed
4147 prematurely by EGD/PRNGD.
4148 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4150 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4153 *) Back-port of selected performance improvements from development
4154 branch, as well as improved support for PowerPC platforms.
4157 *) Add lots of checks for memory allocation failure, error codes to indicate
4158 failure and freeing up memory if a failure occurs.
4159 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4161 *) Add new -passin argument to dgst.
4164 *) Perform some character comparisons of different types in X509_NAME_cmp:
4165 this is needed for some certificates that reencode DNs into UTF8Strings
4166 (in violation of RFC3280) and can't or wont issue name rollover
4170 *) Make an explicit check during certificate validation to see that
4171 the CA setting in each certificate on the chain is correct. As a
4172 side effect always do the following basic checks on extensions,
4173 not just when there's an associated purpose to the check:
4175 - if there is an unhandled critical extension (unless the user
4176 has chosen to ignore this fault)
4177 - if the path length has been exceeded (if one is set at all)
4178 - that certain extensions fit the associated purpose (if one has
4182 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4184 *) Avoid a race condition when CRLs are checked in a multi threaded
4185 environment. This would happen due to the reordering of the revoked
4186 entries during signature checking and serial number lookup. Now the
4187 encoding is cached and the serial number sort performed under a lock.
4188 Add new STACK function sk_is_sorted().
4191 *) Add Delta CRL to the extension code.
4194 *) Various fixes to s3_pkt.c so alerts are sent properly.
4195 [David Holmes <d.holmes@f5.com>]
4197 *) Reduce the chances of duplicate issuer name and serial numbers (in
4198 violation of RFC3280) using the OpenSSL certificate creation utilities.
4199 This is done by creating a random 64 bit value for the initial serial
4200 number when a serial number file is created or when a self signed
4201 certificate is created using 'openssl req -x509'. The initial serial
4202 number file is created using 'openssl x509 -next_serial' in CA.pl
4203 rather than being initialized to 1.
4206 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4208 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4209 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4210 [Joe Orton, Steve Henson]
4212 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4214 [Joe Orton, Steve Henson]
4216 *) Make it possible to have multiple active certificates with the same
4217 subject in the CA index file. This is done only if the keyword
4218 'unique_subject' is set to 'no' in the main CA section (default
4219 if 'CA_default') of the configuration file. The value is saved
4220 with the database itself in a separate index attribute file,
4221 named like the index file with '.attr' appended to the name.
4224 *) X509 verify fixes. Disable broken certificate workarounds when
4225 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4226 keyUsage extension present. Don't accept CRLs with unhandled critical
4227 extensions: since verify currently doesn't process CRL extensions this
4228 rejects a CRL with *any* critical extensions. Add new verify error codes
4232 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4233 A clarification of RFC2560 will require the use of OCTET STRINGs and
4234 some implementations cannot handle the current raw format. Since OpenSSL
4235 copies and compares OCSP nonces as opaque blobs without any attempt at
4236 parsing them this should not create any compatibility issues.
4239 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4240 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4241 this HMAC (and other) operations are several times slower than OpenSSL
4245 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4246 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4248 *) Use the correct content when signing type "other".
4251 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4253 *) Fix various bugs revealed by running the NISCC test suite:
4255 Stop out of bounds reads in the ASN1 code when presented with
4256 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4258 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4260 If verify callback ignores invalid public key errors don't try to check
4261 certificate signature with the NULL public key.
4265 *) New -ignore_err option in ocsp application to stop the server
4266 exiting on the first error in a request.
4269 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4270 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4274 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4275 extra data after the compression methods not only for TLS 1.0
4276 but also for SSL 3.0 (as required by the specification).
4277 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4279 *) Change X509_certificate_type() to mark the key as exported/exportable
4280 when it's 512 *bits* long, not 512 bytes.
4283 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4284 blocks during encryption.
4287 *) Various fixes to base64 BIO and non blocking I/O. On write
4288 flushes were not handled properly if the BIO retried. On read
4289 data was not being buffered properly and had various logic bugs.
4290 This also affects blocking I/O when the data being decoded is a
4294 *) Various S/MIME bugfixes and compatibility changes:
4295 output correct application/pkcs7 MIME type if
4296 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4297 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4298 of files as .eml work). Correctly handle very long lines in MIME
4302 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4304 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4305 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4306 a protocol version number mismatch like a decryption error
4307 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4310 *) Turn on RSA blinding by default in the default implementation
4311 to avoid a timing attack. Applications that don't want it can call
4312 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4313 They would be ill-advised to do so in most cases.
4314 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4316 *) Change RSA blinding code so that it works when the PRNG is not
4317 seeded (in this case, the secret RSA exponent is abused as
4318 an unpredictable seed -- if it is not unpredictable, there
4319 is no point in blinding anyway). Make RSA blinding thread-safe
4320 by remembering the creator's thread ID in rsa->blinding and
4321 having all other threads use local one-time blinding factors
4322 (this requires more computation than sharing rsa->blinding, but
4323 avoids excessive locking; and if an RSA object is not shared
4324 between threads, blinding will still be very fast).
4327 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4328 ENGINE as defaults for all supported algorithms irrespective of
4329 the 'flags' parameter. 'flags' is now honoured, so applications
4330 should make sure they are passing it correctly.
4333 *) Target "mingw" now allows native Windows code to be generated in
4334 the Cygwin environment as well as with the MinGW compiler.
4337 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4339 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4340 via timing by performing a MAC computation even if incorrrect
4341 block cipher padding has been found. This is a countermeasure
4342 against active attacks where the attacker has to distinguish
4343 between bad padding and a MAC verification error. (CVE-2003-0078)
4345 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4346 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4347 Martin Vuagnoux (EPFL, Ilion)]
4349 *) Make the no-err option work as intended. The intention with no-err
4350 is not to have the whole error stack handling routines removed from
4351 libcrypto, it's only intended to remove all the function name and
4352 reason texts, thereby removing some of the footprint that may not
4353 be interesting if those errors aren't displayed anyway.
4355 NOTE: it's still possible for any application or module to have it's
4356 own set of error texts inserted. The routines are there, just not
4357 used by default when no-err is given.
4360 *) Add support for FreeBSD on IA64.
4361 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4363 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4364 Kerberos function mit_des_cbc_cksum(). Before this change,
4365 the value returned by DES_cbc_cksum() was like the one from
4366 mit_des_cbc_cksum(), except the bytes were swapped.
4367 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4369 *) Allow an application to disable the automatic SSL chain building.
4370 Before this a rather primitive chain build was always performed in
4371 ssl3_output_cert_chain(): an application had no way to send the
4372 correct chain if the automatic operation produced an incorrect result.
4374 Now the chain builder is disabled if either:
4376 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4378 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4380 The reasoning behind this is that an application would not want the
4381 auto chain building to take place if extra chain certificates are
4382 present and it might also want a means of sending no additional
4383 certificates (for example the chain has two certificates and the
4387 *) Add the possibility to build without the ENGINE framework.
4388 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4390 *) Under Win32 gmtime() can return NULL: check return value in
4391 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4394 *) DSA routines: under certain error conditions uninitialized BN objects
4395 could be freed. Solution: make sure initialization is performed early
4396 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4397 Nils Larsch <nla@trustcenter.de> via PR#459)
4400 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4401 checked on reconnect on the client side, therefore session resumption
4402 could still fail with a "ssl session id is different" error. This
4403 behaviour is masked when SSL_OP_ALL is used due to
4404 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4405 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4406 followup to PR #377.
4409 *) IA-32 assembler support enhancements: unified ELF targets, support
4410 for SCO/Caldera platforms, fix for Cygwin shared build.
4413 *) Add support for FreeBSD on sparc64. As a consequence, support for
4414 FreeBSD on non-x86 processors is separate from x86 processors on
4415 the config script, much like the NetBSD support.
4416 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4418 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4420 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4423 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4424 code (06) was taken as the first octet of the session ID and the last
4425 octet was ignored consequently. As a result SSLv2 client side session
4426 caching could not have worked due to the session ID mismatch between
4428 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4432 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4433 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4437 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4438 seems that in spite of existing for more than a year, many application
4439 author have done nothing to provide the necessary callbacks, which
4440 means that this particular engine will not work properly anywhere.
4441 This is a very unfortunate situation which forces us, in the name
4442 of usability, to give the hw_ncipher.c a static lock, which is part
4444 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4445 appear in 0.9.8 or later. We EXPECT application authors to have
4446 dealt properly with this when 0.9.8 is released (unless we actually
4447 make such changes in the libcrypto locking code that changes will
4448 have to be made anyway).
4451 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4452 octets have been read, EOF or an error occurs. Without this change
4453 some truncated ASN1 structures will not produce an error.
4456 *) Disable Heimdal support, since it hasn't been fully implemented.
4457 Still give the possibility to force the use of Heimdal, but with
4458 warnings and a request that patches get sent to openssl-dev.
4461 *) Add the VC-CE target, introduce the WINCE sysname, and add
4462 INSTALL.WCE and appropriate conditionals to make it build.
4463 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4465 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4466 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4467 edit numbers of the version.
4468 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4470 *) Introduce safe string copy and catenation functions
4471 (BUF_strlcpy() and BUF_strlcat()).
4472 [Ben Laurie (CHATS) and Richard Levitte]
4474 *) Avoid using fixed-size buffers for one-line DNs.
4475 [Ben Laurie (CHATS)]
4477 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4478 resizing buffers containing secrets, and use where appropriate.
4479 [Ben Laurie (CHATS)]
4481 *) Avoid using fixed size buffers for configuration file location.
4482 [Ben Laurie (CHATS)]
4484 *) Avoid filename truncation for various CA files.
4485 [Ben Laurie (CHATS)]
4487 *) Use sizeof in preference to magic numbers.
4488 [Ben Laurie (CHATS)]
4490 *) Avoid filename truncation in cert requests.
4491 [Ben Laurie (CHATS)]
4493 *) Add assertions to check for (supposedly impossible) buffer
4495 [Ben Laurie (CHATS)]
4497 *) Don't cache truncated DNS entries in the local cache (this could
4498 potentially lead to a spoofing attack).
4499 [Ben Laurie (CHATS)]
4501 *) Fix various buffers to be large enough for hex/decimal
4502 representations in a platform independent manner.
4503 [Ben Laurie (CHATS)]
4505 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4506 resizing buffers containing secrets, and use where appropriate.
4507 [Ben Laurie (CHATS)]
4509 *) Add BIO_indent() to avoid much slightly worrying code to do
4511 [Ben Laurie (CHATS)]
4513 *) Convert sprintf()/BIO_puts() to BIO_printf().
4514 [Ben Laurie (CHATS)]
4516 *) buffer_gets() could terminate with the buffer only half
4518 [Ben Laurie (CHATS)]
4520 *) Add assertions to prevent user-supplied crypto functions from
4521 overflowing internal buffers by having large block sizes, etc.
4522 [Ben Laurie (CHATS)]
4524 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4526 [Ben Laurie (CHATS)]
4528 *) Eliminate unused copy of key in RC4.
4529 [Ben Laurie (CHATS)]
4531 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4532 [Ben Laurie (CHATS)]
4534 *) Fix off-by-one error in EGD path.
4535 [Ben Laurie (CHATS)]
4537 *) If RANDFILE path is too long, ignore instead of truncating.
4538 [Ben Laurie (CHATS)]
4540 *) Eliminate unused and incorrectly sized X.509 structure
4542 [Ben Laurie (CHATS)]
4544 *) Eliminate unused and dangerous function knumber().
4545 [Ben Laurie (CHATS)]
4547 *) Eliminate unused and dangerous structure, KSSL_ERR.
4548 [Ben Laurie (CHATS)]
4550 *) Protect against overlong session ID context length in an encoded
4551 session object. Since these are local, this does not appear to be
4553 [Ben Laurie (CHATS)]
4555 *) Change from security patch (see 0.9.6e below) that did not affect
4556 the 0.9.6 release series:
4558 Remote buffer overflow in SSL3 protocol - an attacker could
4559 supply an oversized master key in Kerberos-enabled versions.
4561 [Ben Laurie (CHATS)]
4563 *) Change the SSL kerb5 codes to match RFC 2712.
4566 *) Make -nameopt work fully for req and add -reqopt switch.
4567 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4569 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4570 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4572 *) Make sure tests can be performed even if the corresponding algorithms
4573 have been removed entirely. This was also the last step to make
4574 OpenSSL compilable with DJGPP under all reasonable conditions.
4575 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4577 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4578 to allow version independent disabling of normally unselected ciphers,
4579 which may be activated as a side-effect of selecting a single cipher.
4581 (E.g., cipher list string "RSA" enables ciphersuites that are left
4582 out of "ALL" because they do not provide symmetric encryption.
4583 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4584 [Lutz Jaenicke, Bodo Moeller]
4586 *) Add appropriate support for separate platform-dependent build
4587 directories. The recommended way to make a platform-dependent
4588 build directory is the following (tested on Linux), maybe with
4591 # Place yourself outside of the OpenSSL source tree. In
4592 # this example, the environment variable OPENSSL_SOURCE
4593 # is assumed to contain the absolute OpenSSL source directory.
4594 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4595 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4596 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4597 mkdir -p `dirname $F`
4598 ln -s $OPENSSL_SOURCE/$F $F
4601 To be absolutely sure not to disturb the source tree, a "make clean"
4602 is a good thing. If it isn't successfull, don't worry about it,
4603 it probably means the source directory is very clean.
4606 *) Make sure any ENGINE control commands make local copies of string
4607 pointers passed to them whenever necessary. Otherwise it is possible
4608 the caller may have overwritten (or deallocated) the original string
4609 data when a later ENGINE operation tries to use the stored values.
4610 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4612 *) Improve diagnostics in file reading and command-line digests.
4613 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4615 *) Add AES modes CFB and OFB to the object database. Correct an
4616 error in AES-CFB decryption.
4619 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4620 allows existing EVP_CIPHER_CTX structures to be reused after
4621 calling EVP_*Final(). This behaviour is used by encryption
4622 BIOs and some applications. This has the side effect that
4623 applications must explicitly clean up cipher contexts with
4624 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4627 *) Check the values of dna and dnb in bn_mul_recursive before calling
4628 bn_mul_comba (a non zero value means the a or b arrays do not contain
4629 n2 elements) and fallback to bn_mul_normal if either is not zero.
4632 *) Fix escaping of non-ASCII characters when using the -subj option
4633 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4636 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4637 form for "surname", serialNumber has no short form.
4638 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4639 therefore remove "mail" short name for "internet 7".
4640 The OID for unique identifiers in X509 certificates is
4641 x500UniqueIdentifier, not uniqueIdentifier.
4642 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4645 *) Add an "init" command to the ENGINE config module and auto initialize
4646 ENGINEs. Without any "init" command the ENGINE will be initialized
4647 after all ctrl commands have been executed on it. If init=1 the
4648 ENGINE is initailized at that point (ctrls before that point are run
4649 on the uninitialized ENGINE and after on the initialized one). If
4650 init=0 then the ENGINE will not be iniatialized at all.
4653 *) Fix the 'app_verify_callback' interface so that the user-defined
4654 argument is actually passed to the callback: In the
4655 SSL_CTX_set_cert_verify_callback() prototype, the callback
4656 declaration has been changed from
4659 int (*cb)(X509_STORE_CTX *,void *);
4660 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4661 i=s->ctx->app_verify_callback(&ctx)
4662 has been changed into
4663 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4665 To update applications using SSL_CTX_set_cert_verify_callback(),
4666 a dummy argument can be added to their callback functions.
4667 [D. K. Smetters <smetters@parc.xerox.com>]
4669 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4670 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4672 *) Add and OPENSSL_LOAD_CONF define which will cause
4673 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4674 This allows older applications to transparently support certain
4675 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4676 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4677 load the config file and OPENSSL_add_all_algorithms_conf() which will
4678 always load it have also been added.
4681 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4682 Adjust NIDs and EVP layer.
4683 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4685 *) Config modules support in openssl utility.
4687 Most commands now load modules from the config file,
4688 though in a few (such as version) this isn't done
4689 because it couldn't be used for anything.
4691 In the case of ca and req the config file used is
4692 the same as the utility itself: that is the -config
4693 command line option can be used to specify an
4697 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4698 use "openssl_conf" if filename is NULL use default openssl config file.
4701 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4702 config section name. Add a new flag to tolerate a missing config file
4703 and move code to CONF_modules_load_file().
4706 *) Support for crypto accelerator cards from Accelerated Encryption
4707 Processing, www.aep.ie. (Use engine 'aep')
4708 The support was copied from 0.9.6c [engine] and adapted/corrected
4709 to work with the new engine framework.
4710 [AEP Inc. and Richard Levitte]
4712 *) Support for SureWare crypto accelerator cards from Baltimore
4713 Technologies. (Use engine 'sureware')
4714 The support was copied from 0.9.6c [engine] and adapted
4715 to work with the new engine framework.
4718 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4719 make the newer ENGINE framework commands for the CHIL engine work.
4720 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4722 *) Make it possible to produce shared libraries on ReliantUNIX.
4723 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4725 *) Add the configuration target debug-linux-ppro.
4726 Make 'openssl rsa' use the general key loading routines
4727 implemented in apps.c, and make those routines able to
4728 handle the key format FORMAT_NETSCAPE and the variant
4730 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4732 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4733 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4735 *) Add -keyform to rsautl, and document -engine.
4736 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4738 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4739 BIO_R_NO_SUCH_FILE error code rather than the generic
4740 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4743 *) Add new functions
4745 ERR_peek_last_error_line
4746 ERR_peek_last_error_line_data.
4747 These are similar to
4750 ERR_peek_error_line_data,
4751 but report on the latest error recorded rather than the first one
4752 still in the error queue.
4753 [Ben Laurie, Bodo Moeller]
4755 *) default_algorithms option in ENGINE config module. This allows things
4757 default_algorithms = ALL
4758 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4761 *) Prelminary ENGINE config module.
4764 *) New experimental application configuration code.
4767 *) Change the AES code to follow the same name structure as all other
4768 symmetric ciphers, and behave the same way. Move everything to
4769 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4770 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4772 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4773 [Ben Laurie and Theo de Raadt]
4775 *) Add option to output public keys in req command.
4776 [Massimiliano Pala madwolf@openca.org]
4778 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4779 (up to about 10% better than before for P-192 and P-224).
4782 *) New functions/macros
4784 SSL_CTX_set_msg_callback(ctx, cb)
4785 SSL_CTX_set_msg_callback_arg(ctx, arg)
4786 SSL_set_msg_callback(ssl, cb)
4787 SSL_set_msg_callback_arg(ssl, arg)
4789 to request calling a callback function
4791 void cb(int write_p, int version, int content_type,
4792 const void *buf, size_t len, SSL *ssl, void *arg)
4794 whenever a protocol message has been completely received
4795 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4796 protocol version according to which the SSL library interprets
4797 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4798 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4799 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4800 specification (change_cipher_spec(20), alert(21), handshake(22)).
4801 'buf' and 'len' point to the actual message, 'ssl' to the
4802 SSL object, and 'arg' is the application-defined value set by
4803 SSL[_CTX]_set_msg_callback_arg().
4805 'openssl s_client' and 'openssl s_server' have new '-msg' options
4806 to enable a callback that displays all protocol messages.
4809 *) Change the shared library support so shared libraries are built as
4810 soon as the corresponding static library is finished, and thereby get
4811 openssl and the test programs linked against the shared library.
4812 This still only happens when the keyword "shard" has been given to
4813 the configuration scripts.
4815 NOTE: shared library support is still an experimental thing, and
4816 backward binary compatibility is still not guaranteed.
4817 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4819 *) Add support for Subject Information Access extension.
4820 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4822 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4823 additional bytes when new memory had to be allocated, not just
4824 when reusing an existing buffer.
4827 *) New command line and configuration option 'utf8' for the req command.
4828 This allows field values to be specified as UTF8 strings.
4831 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4832 runs for the former and machine-readable output for the latter.
4835 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4836 of the e-mail address in the DN (i.e., it will go into a certificate
4837 extension only). The new configuration file option 'email_in_dn = no'
4838 has the same effect.
4839 [Massimiliano Pala madwolf@openca.org]
4841 *) Change all functions with names starting with des_ to be starting
4842 with DES_ instead. Add wrappers that are compatible with libdes,
4843 but are named _ossl_old_des_*. Finally, add macros that map the
4844 des_* symbols to the corresponding _ossl_old_des_* if libdes
4845 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4846 desired, the des_* symbols will be mapped to DES_*, with one
4849 Since we provide two compatibility mappings, the user needs to
4850 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4851 compatibility is desired. The default (i.e., when that macro
4852 isn't defined) is OpenSSL 0.9.6c compatibility.
4854 There are also macros that enable and disable the support of old
4855 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4856 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4857 are defined, the default will apply: to support the old des routines.
4859 In either case, one must include openssl/des.h to get the correct
4860 definitions. Do not try to just include openssl/des_old.h, that
4863 NOTE: This is a major break of an old API into a new one. Software
4864 authors are encouraged to switch to the DES_ style functions. Some
4865 time in the future, des_old.h and the libdes compatibility functions
4866 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4867 default), and then completely removed.
4870 *) Test for certificates which contain unsupported critical extensions.
4871 If such a certificate is found during a verify operation it is
4872 rejected by default: this behaviour can be overridden by either
4873 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4874 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4875 X509_supported_extension() has also been added which returns 1 if a
4876 particular extension is supported.
4879 *) Modify the behaviour of EVP cipher functions in similar way to digests
4880 to retain compatibility with existing code.
4883 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4884 compatibility with existing code. In particular the 'ctx' parameter does
4885 not have to be to be initialized before the call to EVP_DigestInit() and
4886 it is tidied up after a call to EVP_DigestFinal(). New function
4887 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4888 EVP_MD_CTX_copy() changed to not require the destination to be
4889 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4890 requires the destination to be valid.
4892 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4893 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4896 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4897 so that complete 'Handshake' protocol structures are kept in memory
4898 instead of overwriting 'msg_type' and 'length' with 'body' data.
4901 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4902 [Massimo Santin via Richard Levitte]
4904 *) Major restructuring to the underlying ENGINE code. This includes
4905 reduction of linker bloat, separation of pure "ENGINE" manipulation
4906 (initialisation, etc) from functionality dealing with implementations
4907 of specific crypto iterfaces. This change also introduces integrated
4908 support for symmetric ciphers and digest implementations - so ENGINEs
4909 can now accelerate these by providing EVP_CIPHER and EVP_MD
4910 implementations of their own. This is detailed in crypto/engine/README
4911 as it couldn't be adequately described here. However, there are a few
4912 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4913 were changed in the original introduction of ENGINE code have now
4914 reverted back - the hooking from this code to ENGINE is now a good
4915 deal more passive and at run-time, operations deal directly with
4916 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4917 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4918 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4919 they were not being used by the framework as there is no concept of a
4920 BIGNUM_METHOD and they could not be generalised to the new
4921 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4922 ENGINE_cpy() has been removed as it cannot be consistently defined in
4926 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4929 *) Change mkdef.pl to sort symbols that get the same entry number,
4930 and make sure the automatically generated functions ERR_load_*
4931 become part of libeay.num as well.
4934 *) New function SSL_renegotiate_pending(). This returns true once
4935 renegotiation has been requested (either SSL_renegotiate() call
4936 or HelloRequest/ClientHello receveived from the peer) and becomes
4937 false once a handshake has been completed.
4938 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4939 sends a HelloRequest, but does not ensure that a handshake takes
4940 place. SSL_renegotiate_pending() is useful for checking if the
4941 client has followed the request.)
4944 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4945 By default, clients may request session resumption even during
4946 renegotiation (if session ID contexts permit); with this option,
4947 session resumption is possible only in the first handshake.
4949 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4950 more bits available for options that should not be part of
4951 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4954 *) Add some demos for certificate and certificate request creation.
4957 *) Make maximum certificate chain size accepted from the peer application
4958 settable (SSL*_get/set_max_cert_list()), as proposed by
4959 "Douglas E. Engert" <deengert@anl.gov>.
4962 *) Add support for shared libraries for Unixware-7
4963 (Boyd Lynn Gerber <gerberb@zenez.com>).
4966 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4967 be done prior to destruction. Use this to unload error strings from
4968 ENGINEs that load their own error strings. NB: This adds two new API
4969 functions to "get" and "set" this destroy handler in an ENGINE.
4972 *) Alter all existing ENGINE implementations (except "openssl" and
4973 "openbsd") to dynamically instantiate their own error strings. This
4974 makes them more flexible to be built both as statically-linked ENGINEs
4975 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4976 Also, add stub code to each that makes building them as self-contained
4977 shared-libraries easier (see README.ENGINE).
4980 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4981 implementations into applications that are completely implemented in
4982 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4983 commands that can be used to configure what shared-library to load and
4984 to control aspects of the way it is handled. Also, made an update to
4985 the README.ENGINE file that brings its information up-to-date and
4986 provides some information and instructions on the "dynamic" ENGINE
4987 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4990 *) Make it possible to unload ranges of ERR strings with a new
4991 "ERR_unload_strings" function.
4994 *) Add a copy() function to EVP_MD.
4997 *) Make EVP_MD routines take a context pointer instead of just the
4998 md_data void pointer.
5001 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5002 that the digest can only process a single chunk of data
5003 (typically because it is provided by a piece of
5004 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5005 is only going to provide a single chunk of data, and hence the
5006 framework needn't accumulate the data for oneshot drivers.
5009 *) As with "ERR", make it possible to replace the underlying "ex_data"
5010 functions. This change also alters the storage and management of global
5011 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5012 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5013 index counters. The API functions that use this state have been changed
5014 to take a "class_index" rather than pointers to the class's local STACK
5015 and counter, and there is now an API function to dynamically create new
5016 classes. This centralisation allows us to (a) plug a lot of the
5017 thread-safety problems that existed, and (b) makes it possible to clean
5018 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5019 such data would previously have always leaked in application code and
5020 workarounds were in place to make the memory debugging turn a blind eye
5021 to it. Application code that doesn't use this new function will still
5022 leak as before, but their memory debugging output will announce it now
5023 rather than letting it slide.
5025 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5026 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5027 has a return value to indicate success or failure.
5030 *) Make it possible to replace the underlying "ERR" functions such that the
5031 global state (2 LHASH tables and 2 locks) is only used by the "default"
5032 implementation. This change also adds two functions to "get" and "set"
5033 the implementation prior to it being automatically set the first time
5034 any other ERR function takes place. Ie. an application can call "get",
5035 pass the return value to a module it has just loaded, and that module
5036 can call its own "set" function using that value. This means the
5037 module's "ERR" operations will use (and modify) the error state in the
5038 application and not in its own statically linked copy of OpenSSL code.
5041 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5042 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5043 the operation, and provides a more encapsulated way for external code
5044 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5045 to use these functions rather than manually incrementing the counts.
5047 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5050 *) Add EVP test program.
5053 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5056 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5057 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5058 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5059 These allow a CRL to be built without having to access X509_CRL fields
5060 directly. Modify 'ca' application to use new functions.
5063 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5064 bug workarounds. Rollback attack detection is a security feature.
5065 The problem will only arise on OpenSSL servers when TLSv1 is not
5066 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5067 Software authors not wanting to support TLSv1 will have special reasons
5068 for their choice and can explicitly enable this option.
5069 [Bodo Moeller, Lutz Jaenicke]
5071 *) Rationalise EVP so it can be extended: don't include a union of
5072 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5073 (similar to those existing for EVP_CIPHER_CTX).
5078 EVP_MD_CTX_init(&md); /* new function call */
5079 EVP_DigestInit(&md, EVP_sha1());
5080 EVP_DigestUpdate(&md, in, len);
5081 EVP_DigestFinal(&md, out, NULL);
5082 EVP_MD_CTX_cleanup(&md); /* new function call */
5086 *) Make DES key schedule conform to the usual scheme, as well as
5087 correcting its structure. This means that calls to DES functions
5088 now have to pass a pointer to a des_key_schedule instead of a
5089 plain des_key_schedule (which was actually always a pointer
5092 des_key_schedule ks;
5094 des_set_key_checked(..., &ks);
5095 des_ncbc_encrypt(..., &ks, ...);
5097 (Note that a later change renames 'des_...' into 'DES_...'.)
5100 *) Initial reduction of linker bloat: the use of some functions, such as
5101 PEM causes large amounts of unused functions to be linked in due to
5102 poor organisation. For example pem_all.c contains every PEM function
5103 which has a knock on effect of linking in large amounts of (unused)
5104 ASN1 code. Grouping together similar functions and splitting unrelated
5105 functions prevents this.
5108 *) Cleanup of EVP macros.
5111 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5112 correct _ecb suffix.
5115 *) Add initial OCSP responder support to ocsp application. The
5116 revocation information is handled using the text based index
5117 use by the ca application. The responder can either handle
5118 requests generated internally, supplied in files (for example
5119 via a CGI script) or using an internal minimal server.
5122 *) Add configuration choices to get zlib compression for TLS.
5125 *) Changes to Kerberos SSL for RFC 2712 compliance:
5126 1. Implemented real KerberosWrapper, instead of just using
5127 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5128 2. Implemented optional authenticator field of KerberosWrapper.
5130 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5131 and authenticator structs; see crypto/krb5/.
5133 Generalized Kerberos calls to support multiple Kerberos libraries.
5134 [Vern Staats <staatsvr@asc.hpc.mil>,
5135 Jeffrey Altman <jaltman@columbia.edu>
5136 via Richard Levitte]
5138 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5139 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5140 values for each of the key sizes rather than having just
5141 parameters (and 'speed' generating keys each time).
5144 *) Speed up EVP routines.
5147 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5148 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5149 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5150 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5152 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5153 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5154 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5157 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5159 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5162 *) Added the OS2-EMX target.
5163 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5165 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5166 to support NCONF routines in extension code. New function CONF_set_nconf()
5167 to allow functions which take an NCONF to also handle the old LHASH
5168 structure: this means that the old CONF compatible routines can be
5169 retained (in particular wrt extensions) without having to duplicate the
5170 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5173 *) Enhance the general user interface with mechanisms for inner control
5174 and with possibilities to have yes/no kind of prompts.
5177 *) Change all calls to low level digest routines in the library and
5178 applications to use EVP. Add missing calls to HMAC_cleanup() and
5179 don't assume HMAC_CTX can be copied using memcpy().
5180 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5182 *) Add the possibility to control engines through control names but with
5183 arbitrary arguments instead of just a string.
5184 Change the key loaders to take a UI_METHOD instead of a callback
5185 function pointer. NOTE: this breaks binary compatibility with earlier
5186 versions of OpenSSL [engine].
5187 Adapt the nCipher code for these new conditions and add a card insertion
5191 *) Enhance the general user interface with mechanisms to better support
5192 dialog box interfaces, application-defined prompts, the possibility
5193 to use defaults (for example default passwords from somewhere else)
5194 and interrupts/cancellations.
5197 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5198 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5201 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5202 tidy up some unnecessarily weird code in 'sk_new()').
5203 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5205 *) Change the key loading routines for ENGINEs to use the same kind
5206 callback (pem_password_cb) as all other routines that need this
5210 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5211 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5212 than this minimum value is recommended.
5215 *) New random seeder for OpenVMS, using the system process statistics
5216 that are easily reachable.
5219 *) Windows apparently can't transparently handle global
5220 variables defined in DLLs. Initialisations such as:
5222 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5224 wont compile. This is used by the any applications that need to
5225 declare their own ASN1 modules. This was fixed by adding the option
5226 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5227 needed for static libraries under Win32.
5230 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5231 setting of purpose and trust fields. New X509_STORE trust and
5232 purpose functions and tidy up setting in other SSL functions.
5235 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5236 structure. These are inherited by X509_STORE_CTX when it is
5237 initialised. This allows various defaults to be set in the
5238 X509_STORE structure (such as flags for CRL checking and custom
5239 purpose or trust settings) for functions which only use X509_STORE_CTX
5240 internally such as S/MIME.
5242 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5243 trust settings if they are not set in X509_STORE. This allows X509_STORE
5244 purposes and trust (in S/MIME for example) to override any set by default.
5246 Add command line options for CRL checking to smime, s_client and s_server
5250 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5251 are set then the CRL is looked up in the X509_STORE structure and
5252 its validity and signature checked, then if the certificate is found
5253 in the CRL the verify fails with a revoked error.
5255 Various new CRL related callbacks added to X509_STORE_CTX structure.
5257 Command line options added to 'verify' application to support this.
5259 This needs some additional work, such as being able to handle multiple
5260 CRLs with different times, extension based lookup (rather than just
5261 by subject name) and ultimately more complete V2 CRL extension
5265 *) Add a general user interface API (crypto/ui/). This is designed
5266 to replace things like des_read_password and friends (backward
5267 compatibility functions using this new API are provided).
5268 The purpose is to remove prompting functions from the DES code
5269 section as well as provide for prompting through dialog boxes in
5270 a window system and the like.
5273 *) Add "ex_data" support to ENGINE so implementations can add state at a
5274 per-structure level rather than having to store it globally.
5277 *) Make it possible for ENGINE structures to be copied when retrieved by
5278 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5279 This causes the "original" ENGINE structure to act like a template,
5280 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5281 operational state can be localised to each ENGINE structure, despite the
5282 fact they all share the same "methods". New ENGINE structures returned in
5283 this case have no functional references and the return value is the single
5284 structural reference. This matches the single structural reference returned
5285 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5289 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5290 needs to match any other type at all we need to manually clear the
5294 *) Changes to the "openssl engine" utility to include;
5295 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5296 about an ENGINE's available control commands.
5297 - executing control commands from command line arguments using the
5298 '-pre' and '-post' switches. '-post' is only used if '-t' is
5299 specified and the ENGINE is successfully initialised. The syntax for
5300 the individual commands are colon-separated, for example;
5301 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5304 *) New dynamic control command support for ENGINEs. ENGINEs can now
5305 declare their own commands (numbers), names (strings), descriptions,
5306 and input types for run-time discovery by calling applications. A
5307 subset of these commands are implicitly classed as "executable"
5308 depending on their input type, and only these can be invoked through
5309 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5310 can be based on user input, config files, etc). The distinction is
5311 that "executable" commands cannot return anything other than a boolean
5312 result and can only support numeric or string input, whereas some
5313 discoverable commands may only be for direct use through
5314 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5315 pointers, or other custom uses. The "executable" commands are to
5316 support parameterisations of ENGINE behaviour that can be
5317 unambiguously defined by ENGINEs and used consistently across any
5318 OpenSSL-based application. Commands have been added to all the
5319 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5320 control over shared-library paths without source code alterations.
5323 *) Changed all ENGINE implementations to dynamically allocate their
5324 ENGINEs rather than declaring them statically. Apart from this being
5325 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5326 this also allows the implementations to compile without using the
5327 internal engine_int.h header.
5330 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5331 'const' value. Any code that should be able to modify a RAND_METHOD
5332 should already have non-const pointers to it (ie. they should only
5333 modify their own ones).
5336 *) Made a variety of little tweaks to the ENGINE code.
5337 - "atalla" and "ubsec" string definitions were moved from header files
5338 to C code. "nuron" string definitions were placed in variables
5339 rather than hard-coded - allowing parameterisation of these values
5340 later on via ctrl() commands.
5341 - Removed unused "#if 0"'d code.
5342 - Fixed engine list iteration code so it uses ENGINE_free() to release
5343 structural references.
5344 - Constified the RAND_METHOD element of ENGINE structures.
5345 - Constified various get/set functions as appropriate and added
5346 missing functions (including a catch-all ENGINE_cpy that duplicates
5347 all ENGINE values onto a new ENGINE except reference counts/state).
5348 - Removed NULL parameter checks in get/set functions. Setting a method
5349 or function to NULL is a way of cancelling out a previously set
5350 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5351 and doesn't justify the extra error symbols and code.
5352 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5353 flags from engine_int.h to engine.h.
5354 - Changed prototypes for ENGINE handler functions (init(), finish(),
5355 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5358 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5359 to the algorithm using long division. The binary algorithm can be
5360 used only if the modulus is odd. On 32-bit systems, it is faster
5361 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5362 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5363 up to 450 bits. In 64-bit environments, the binary algorithm
5364 appears to be advantageous for much longer moduli; here we use it
5365 for moduli up to 2048 bits.
5368 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5369 could not support the combine flag in choice fields.
5372 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5373 extensions from a certificate request to the certificate.
5376 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5377 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5378 file: this allows the display of the certificate about to be
5379 signed to be customised, to allow certain fields to be included
5380 or excluded and extension details. The old system didn't display
5381 multicharacter strings properly, omitted fields not in the policy
5382 and couldn't display additional details such as extensions.
5385 *) Function EC_POINTs_mul for multiple scalar multiplication
5386 of an arbitrary number of elliptic curve points
5387 \sum scalars[i]*points[i],
5388 optionally including the generator defined for the EC_GROUP:
5389 scalar*generator + \sum scalars[i]*points[i].
5391 EC_POINT_mul is a simple wrapper function for the typical case
5392 that the point list has just one item (besides the optional
5396 *) First EC_METHODs for curves over GF(p):
5398 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5399 operations and provides various method functions that can also
5400 operate with faster implementations of modular arithmetic.
5402 EC_GFp_mont_method() reuses most functions that are part of
5403 EC_GFp_simple_method, but uses Montgomery arithmetic.
5405 [Bodo Moeller; point addition and point doubling
5406 implementation directly derived from source code provided by
5407 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5409 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5410 crypto/ec/ec_lib.c):
5412 Curves are EC_GROUP objects (with an optional group generator)
5413 based on EC_METHODs that are built into the library.
5415 Points are EC_POINT objects based on EC_GROUP objects.
5417 Most of the framework would be able to handle curves over arbitrary
5418 finite fields, but as there are no obvious types for fields other
5419 than GF(p), some functions are limited to that for now.
5422 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5423 that the file contains a complete HTTP response.
5426 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5427 change the def and num file printf format specifier from "%-40sXXX"
5428 to "%-39s XXX". The latter will always guarantee a space after the
5429 field while the former will cause them to run together if the field
5430 is 40 of more characters long.
5433 *) Constify the cipher and digest 'method' functions and structures
5434 and modify related functions to take constant EVP_MD and EVP_CIPHER
5438 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5439 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5442 *) Modify EVP_Digest*() routines so they now return values. Although the
5443 internal software routines can never fail additional hardware versions
5447 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5449 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5450 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5453 ERR_R_NESTED_ASN1_ERROR
5455 ERR_R_MISSING_ASN1_EOS
5456 were 4 .. 9, conflicting with
5457 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5459 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5460 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5462 Add new error code 'ERR_R_INTERNAL_ERROR'.
5465 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5469 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5470 sets the subject name for a new request or supersedes the
5471 subject name in a given request. Formats that can be parsed are
5472 'CN=Some Name, OU=myOU, C=IT'
5474 'CN=Some Name/OU=myOU/C=IT'.
5476 Add options '-batch' and '-verbose' to 'openssl req'.
5477 [Massimiliano Pala <madwolf@hackmasters.net>]
5479 *) Introduce the possibility to access global variables through
5480 functions on platform were that's the best way to handle exporting
5481 global variables in shared libraries. To enable this functionality,
5482 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5483 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5484 is normally done by Configure or something similar).
5486 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5487 in the source file (foo.c) like this:
5489 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5490 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5492 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5493 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5495 OPENSSL_DECLARE_GLOBAL(int,foo);
5496 #define foo OPENSSL_GLOBAL_REF(foo)
5497 OPENSSL_DECLARE_GLOBAL(double,bar);
5498 #define bar OPENSSL_GLOBAL_REF(bar)
5500 The #defines are very important, and therefore so is including the
5501 header file everywhere where the defined globals are used.
5503 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5504 of ASN.1 items, but that structure is a bit different.
5506 The largest change is in util/mkdef.pl which has been enhanced with
5507 better and easier to understand logic to choose which symbols should
5508 go into the Windows .def files as well as a number of fixes and code
5509 cleanup (among others, algorithm keywords are now sorted
5510 lexicographically to avoid constant rewrites).
5513 *) In BN_div() keep a copy of the sign of 'num' before writing the
5514 result to 'rm' because if rm==num the value will be overwritten
5515 and produce the wrong result if 'num' is negative: this caused
5516 problems with BN_mod() and BN_nnmod().
5519 *) Function OCSP_request_verify(). This checks the signature on an
5520 OCSP request and verifies the signer certificate. The signer
5521 certificate is just checked for a generic purpose and OCSP request
5525 *) Add OCSP_check_validity() function to check the validity of OCSP
5526 responses. OCSP responses are prepared in real time and may only
5527 be a few seconds old. Simply checking that the current time lies
5528 between thisUpdate and nextUpdate max reject otherwise valid responses
5529 caused by either OCSP responder or client clock inaccuracy. Instead
5530 we allow thisUpdate and nextUpdate to fall within a certain period of
5531 the current time. The age of the response can also optionally be
5532 checked. Two new options -validity_period and -status_age added to
5536 *) If signature or public key algorithm is unrecognized print out its
5537 OID rather that just UNKNOWN.
5540 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5541 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5542 ID to be generated from the issuer certificate alone which can then be
5543 passed to OCSP_id_issuer_cmp().
5546 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5547 ASN1 modules to export functions returning ASN1_ITEM pointers
5548 instead of the ASN1_ITEM structures themselves. This adds several
5549 new macros which allow the underlying ASN1 function/structure to
5550 be accessed transparently. As a result code should not use ASN1_ITEM
5551 references directly (such as &X509_it) but instead use the relevant
5552 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5553 use of the new ASN1 code on platforms where exporting structures
5554 is problematical (for example in shared libraries) but exporting
5555 functions returning pointers to structures is not.
5558 *) Add support for overriding the generation of SSL/TLS session IDs.
5559 These callbacks can be registered either in an SSL_CTX or per SSL.
5560 The purpose of this is to allow applications to control, if they wish,
5561 the arbitrary values chosen for use as session IDs, particularly as it
5562 can be useful for session caching in multiple-server environments. A
5563 command-line switch for testing this (and any client code that wishes
5564 to use such a feature) has been added to "s_server".
5565 [Geoff Thorpe, Lutz Jaenicke]
5567 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5568 of the form '#if defined(...) || defined(...) || ...' and
5569 '#if !defined(...) && !defined(...) && ...'. This also avoids
5570 the growing number of special cases it was previously handling.
5573 *) Make all configuration macros available for application by making
5574 sure they are available in opensslconf.h, by giving them names starting
5575 with "OPENSSL_" to avoid conflicts with other packages and by making
5576 sure e_os2.h will cover all platform-specific cases together with
5578 Additionally, it is now possible to define configuration/platform-
5579 specific names (called "system identities"). In the C code, these
5580 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5581 macro with the name beginning with "OPENSSL_SYS_", which is determined
5582 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5586 *) New option -set_serial to 'req' and 'x509' this allows the serial
5587 number to use to be specified on the command line. Previously self
5588 signed certificates were hard coded with serial number 0 and the
5589 CA options of 'x509' had to use a serial number in a file which was
5593 *) New options to 'ca' utility to support V2 CRL entry extensions.
5594 Currently CRL reason, invalidity date and hold instruction are
5595 supported. Add new CRL extensions to V3 code and some new objects.
5598 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5599 disable standard block padding (aka PKCS#5 padding) in the EVP
5600 API, which was previously mandatory. This means that the data is
5601 not padded in any way and so the total length much be a multiple
5602 of the block size, otherwise an error occurs.
5605 *) Initial (incomplete) OCSP SSL support.
5608 *) New function OCSP_parse_url(). This splits up a URL into its host,
5609 port and path components: primarily to parse OCSP URLs. New -url
5610 option to ocsp utility.
5613 *) New nonce behavior. The return value of OCSP_check_nonce() now
5614 reflects the various checks performed. Applications can decide
5615 whether to tolerate certain situations such as an absent nonce
5616 in a response when one was present in a request: the ocsp application
5617 just prints out a warning. New function OCSP_add1_basic_nonce()
5618 this is to allow responders to include a nonce in a response even if
5619 the request is nonce-less.
5622 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5623 skipped when using openssl x509 multiple times on a single input file,
5624 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5627 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5628 set string type: to handle setting ASN1_TIME structures. Fix ca
5629 utility to correctly initialize revocation date of CRLs.
5632 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5633 the clients preferred ciphersuites and rather use its own preferences.
5634 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5635 Internet Explorer by ensuring unchanged hash method during stepup.
5636 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5639 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5640 to aes and add a new 'exist' option to print out symbols that don't
5644 *) Additional options to ocsp utility to allow flags to be set and
5645 additional certificates supplied.
5648 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5649 OCSP client a number of certificate to only verify the response
5653 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5654 handle the new API. Currently only ECB, CBC modes supported. Add new
5657 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5658 Encryption Standard (AES) Ciphersuites for Transport Layer
5659 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5660 not enabled by default and were not part of the "ALL" ciphersuite
5661 alias because they were not yet official; they could be
5662 explicitly requested by specifying the "AESdraft" ciphersuite
5663 group alias. In the final release of OpenSSL 0.9.7, the group
5664 alias is called "AES" and is part of "ALL".)
5665 [Ben Laurie, Steve Henson, Bodo Moeller]
5667 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5668 request to response.
5671 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5672 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5673 extract information from a certificate request. OCSP_response_create()
5674 creates a response and optionally adds a basic response structure.
5675 OCSP_basic_add1_status() adds a complete single response to a basic
5676 response and returns the OCSP_SINGLERESP structure just added (to allow
5677 extensions to be included for example). OCSP_basic_add1_cert() adds a
5678 certificate to a basic response and OCSP_basic_sign() signs a basic
5679 response with various flags. New helper functions ASN1_TIME_check()
5680 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5681 (converts ASN1_TIME to GeneralizedTime).
5684 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5685 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5686 structure from a certificate. X509_pubkey_digest() digests the public_key
5687 contents: this is used in various key identifiers.
5690 *) Make sk_sort() tolerate a NULL argument.
5691 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5693 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5694 passed by the function are trusted implicitly. If any of them signed the
5695 response then it is assumed to be valid and is not verified.
5698 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5699 to data. This was previously part of the PKCS7 ASN1 code. This
5700 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5701 [Steve Henson, reported by Kenneth R. Robinette
5702 <support@securenetterm.com>]
5704 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5705 routines: without these tracing memory leaks is very painful.
5706 Fix leaks in PKCS12 and PKCS7 routines.
5709 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5710 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5711 effectively meant GeneralizedTime would never be used. Now it
5712 is initialised to -1 but X509_time_adj() now has to check the value
5713 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5714 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5715 [Steve Henson, reported by Kenneth R. Robinette
5716 <support@securenetterm.com>]
5718 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5719 result in a zero length in the ASN1_INTEGER structure which was
5720 not consistent with the structure when d2i_ASN1_INTEGER() was used
5721 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5722 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5723 where it did not print out a minus for negative ASN1_INTEGER.
5726 *) Add summary printout to ocsp utility. The various functions which
5727 convert status values to strings have been renamed to:
5728 OCSP_response_status_str(), OCSP_cert_status_str() and
5729 OCSP_crl_reason_str() and are no longer static. New options
5730 to verify nonce values and to disable verification. OCSP response
5731 printout format cleaned up.
5734 *) Add additional OCSP certificate checks. These are those specified
5735 in RFC2560. This consists of two separate checks: the CA of the
5736 certificate being checked must either be the OCSP signer certificate
5737 or the issuer of the OCSP signer certificate. In the latter case the
5738 OCSP signer certificate must contain the OCSP signing extended key
5739 usage. This check is performed by attempting to match the OCSP
5740 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5741 in the OCSP_CERTID structures of the response.
5744 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5745 and related routines. This uses the standard OpenSSL certificate
5746 verify routines to perform initial checks (just CA validity) and
5747 to obtain the certificate chain. Then additional checks will be
5748 performed on the chain. Currently the root CA is checked to see
5749 if it is explicitly trusted for OCSP signing. This is used to set
5750 a root CA as a global signing root: that is any certificate that
5751 chains to that CA is an acceptable OCSP signing certificate.
5754 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5755 extensions from a separate configuration file.
5756 As when reading extensions from the main configuration file,
5757 the '-extensions ...' option may be used for specifying the
5759 [Massimiliano Pala <madwolf@comune.modena.it>]
5761 *) New OCSP utility. Allows OCSP requests to be generated or
5762 read. The request can be sent to a responder and the output
5763 parsed, outputed or printed in text form. Not complete yet:
5764 still needs to check the OCSP response validity.
5767 *) New subcommands for 'openssl ca':
5768 'openssl ca -status <serial>' prints the status of the cert with
5769 the given serial number (according to the index file).
5770 'openssl ca -updatedb' updates the expiry status of certificates
5772 [Massimiliano Pala <madwolf@comune.modena.it>]
5774 *) New '-newreq-nodes' command option to CA.pl. This is like
5775 '-newreq', but calls 'openssl req' with the '-nodes' option
5776 so that the resulting key is not encrypted.
5777 [Damien Miller <djm@mindrot.org>]
5779 *) New configuration for the GNU Hurd.
5780 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5782 *) Initial code to implement OCSP basic response verify. This
5783 is currently incomplete. Currently just finds the signer's
5784 certificate and verifies the signature on the response.
5787 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5788 value of OPENSSLDIR. This is available via the new '-d' option
5789 to 'openssl version', and is also included in 'openssl version -a'.
5792 *) Allowing defining memory allocation callbacks that will be given
5793 file name and line number information in additional arguments
5794 (a const char* and an int). The basic functionality remains, as
5795 well as the original possibility to just replace malloc(),
5796 realloc() and free() by functions that do not know about these
5797 additional arguments. To register and find out the current
5798 settings for extended allocation functions, the following
5799 functions are provided:
5801 CRYPTO_set_mem_ex_functions
5802 CRYPTO_set_locked_mem_ex_functions
5803 CRYPTO_get_mem_ex_functions
5804 CRYPTO_get_locked_mem_ex_functions
5806 These work the same way as CRYPTO_set_mem_functions and friends.
5807 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5808 extended allocation function is enabled.
5809 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5810 a conventional allocation function is enabled.
5811 [Richard Levitte, Bodo Moeller]
5813 *) Finish off removing the remaining LHASH function pointer casts.
5814 There should no longer be any prototype-casting required when using
5815 the LHASH abstraction, and any casts that remain are "bugs". See
5816 the callback types and macros at the head of lhash.h for details
5817 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5820 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5821 If /dev/[u]random devices are not available or do not return enough
5822 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5824 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5825 /etc/entropy will be queried once each in this sequence, quering stops
5826 when enough entropy was collected without querying more sockets.
5829 *) Change the Unix RAND_poll() variant to be able to poll several
5830 random devices, as specified by DEVRANDOM, until a sufficient amount
5831 of data has been collected. We spend at most 10 ms on each file
5832 (select timeout) and read in non-blocking mode. DEVRANDOM now
5833 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5834 (previously it was just the string "/dev/urandom"), so on typical
5835 platforms the 10 ms delay will never occur.
5836 Also separate out the Unix variant to its own file, rand_unix.c.
5837 For VMS, there's a currently-empty rand_vms.c.
5840 *) Move OCSP client related routines to ocsp_cl.c. These
5841 provide utility functions which an application needing
5842 to issue a request to an OCSP responder and analyse the
5843 response will typically need: as opposed to those which an
5844 OCSP responder itself would need which will be added later.
5846 OCSP_request_sign() signs an OCSP request with an API similar
5847 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5848 response. OCSP_response_get1_basic() extracts basic response
5849 from response. OCSP_resp_find_status(): finds and extracts status
5850 information from an OCSP_CERTID structure (which will be created
5851 when the request structure is built). These are built from lower
5852 level functions which work on OCSP_SINGLERESP structures but
5853 wont normally be used unless the application wishes to examine
5854 extensions in the OCSP response for example.
5856 Replace nonce routines with a pair of functions.
5857 OCSP_request_add1_nonce() adds a nonce value and optionally
5858 generates a random value. OCSP_check_nonce() checks the
5859 validity of the nonce in an OCSP response.
5862 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5863 This doesn't copy the supplied OCSP_CERTID and avoids the
5864 need to free up the newly created id. Change return type
5865 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5866 This can then be used to add extensions to the request.
5867 Deleted OCSP_request_new(), since most of its functionality
5868 is now in OCSP_REQUEST_new() (and the case insensitive name
5869 clash) apart from the ability to set the request name which
5870 will be added elsewhere.
5873 *) Update OCSP API. Remove obsolete extensions argument from
5874 various functions. Extensions are now handled using the new
5875 OCSP extension code. New simple OCSP HTTP function which
5876 can be used to send requests and parse the response.
5879 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5880 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5881 uses the special reorder version of SET OF to sort the attributes
5882 and reorder them to match the encoded order. This resolves a long
5883 standing problem: a verify on a PKCS7 structure just after signing
5884 it used to fail because the attribute order did not match the
5885 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5886 it uses the received order. This is necessary to tolerate some broken
5887 software that does not order SET OF. This is handled by encoding
5888 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5889 to produce the required SET OF.
5892 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5893 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5894 files to get correct declarations of the ASN.1 item variables.
5897 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5898 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5899 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5900 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5901 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5902 ASN1_ITEM and no wrapper functions.
5905 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5906 replace the old function pointer based I/O routines. Change most of
5907 the *_d2i_bio() and *_d2i_fp() functions to use these.
5910 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5911 lines, recognice more "algorithms" that can be deselected, and make
5912 it complain about algorithm deselection that isn't recognised.
5915 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5916 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5917 to use new functions. Add NO_ASN1_OLD which can be set to remove
5918 some old style ASN1 functions: this can be used to determine if old
5919 code will still work when these eventually go away.
5922 *) New extension functions for OCSP structures, these follow the
5923 same conventions as certificates and CRLs.
5926 *) New function X509V3_add1_i2d(). This automatically encodes and
5927 adds an extension. Its behaviour can be customised with various
5928 flags to append, replace or delete. Various wrappers added for
5929 certifcates and CRLs.
5932 *) Fix to avoid calling the underlying ASN1 print routine when
5933 an extension cannot be parsed. Correct a typo in the
5934 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5937 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5938 entries for variables.
5941 *) Add functionality to apps/openssl.c for detecting locking
5942 problems: As the program is single-threaded, all we have
5943 to do is register a locking callback using an array for
5944 storing which locks are currently held by the program.
5947 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5948 SSL_get_ex_data_X509_STORE_idx(), which is used in
5949 ssl_verify_cert_chain() and thus can be called at any time
5950 during TLS/SSL handshakes so that thread-safety is essential.
5951 Unfortunately, the ex_data design is not at all suited
5952 for multi-threaded use, so it probably should be abolished.
5955 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5956 [Broadcom, tweaked and integrated by Geoff Thorpe]
5958 *) Move common extension printing code to new function
5959 X509V3_print_extensions(). Reorganise OCSP print routines and
5960 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5963 *) New function X509_signature_print() to remove duplication in some
5967 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5968 set (this was treated exactly the same as SET OF previously). This
5969 is used to reorder the STACK representing the structure to match the
5970 encoding. This will be used to get round a problem where a PKCS7
5971 structure which was signed could not be verified because the STACK
5972 order did not reflect the encoded order.
5975 *) Reimplement the OCSP ASN1 module using the new code.
5978 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5979 for its ASN1 operations. The old style function pointers still exist
5980 for now but they will eventually go away.
5983 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5984 completely replaces the old ASN1 functionality with a table driven
5985 encoder and decoder which interprets an ASN1_ITEM structure describing
5986 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5987 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5988 has also been converted to the new form.
5991 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5992 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5993 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5994 for negative moduli.
5997 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5998 of not touching the result's sign bit.
6001 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6005 *) Changed the LHASH code to use prototypes for callbacks, and created
6006 macros to declare and implement thin (optionally static) functions
6007 that provide type-safety and avoid function pointer casting for the
6008 type-specific callbacks.
6011 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6013 [Veers Staats <staatsvr@asc.hpc.mil>,
6014 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6016 *) Reformat the FAQ so the different questions and answers can be divided
6017 in sections depending on the subject.
6020 *) Have the zlib compression code load ZLIB.DLL dynamically under
6024 *) New function BN_mod_sqrt for computing square roots modulo a prime
6025 (using the probabilistic Tonelli-Shanks algorithm unless
6026 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6027 be handled deterministically).
6028 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6030 *) Make BN_mod_inverse faster by explicitly handling small quotients
6031 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6032 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6035 *) New function BN_kronecker.
6038 *) Fix BN_gcd so that it works on negative inputs; the result is
6039 positive unless both parameters are zero.
6040 Previously something reasonably close to an infinite loop was
6041 possible because numbers could be growing instead of shrinking
6042 in the implementation of Euclid's algorithm.
6045 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6046 sign of the number in question.
6048 Fix BN_is_word(a,w) to work correctly for w == 0.
6050 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6051 because its test if the absolute value of 'a' equals 'w'.
6052 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6053 it exists mostly for use in the implementations of BN_is_zero(),
6054 BN_is_one(), and BN_is_word().
6057 *) New function BN_swap.
6060 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6061 the exponentiation functions are more likely to produce reasonable
6062 results on negative inputs.
6065 *) Change BN_mod_mul so that the result is always non-negative.
6066 Previously, it could be negative if one of the factors was negative;
6067 I don't think anyone really wanted that behaviour.
6070 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6071 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6072 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6073 and add new functions:
6082 BN_mod_lshift1_quick
6086 These functions always generate non-negative results.
6088 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6089 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6091 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6092 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6093 be reduced modulo m.
6094 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6097 The following entry accidentily appeared in the CHANGES file
6098 distributed with OpenSSL 0.9.7. The modifications described in
6099 it do *not* apply to OpenSSL 0.9.7.
6101 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6102 was actually never needed) and in BN_mul(). The removal in BN_mul()
6103 required a small change in bn_mul_part_recursive() and the addition
6104 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6105 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6106 bn_sub_words() and bn_add_words() except they take arrays with
6111 *) In 'openssl passwd', verify passwords read from the terminal
6112 unless the '-salt' option is used (which usually means that
6113 verification would just waste user's time since the resulting
6114 hash is going to be compared with some given password hash)
6115 or the new '-noverify' option is used.
6117 This is an incompatible change, but it does not affect
6118 non-interactive use of 'openssl passwd' (passwords on the command
6119 line, '-stdin' option, '-in ...' option) and thus should not
6123 *) Remove all references to RSAref, since there's no more need for it.
6126 *) Make DSO load along a path given through an environment variable
6127 (SHLIB_PATH) with shl_load().
6130 *) Constify the ENGINE code as a result of BIGNUM constification.
6131 Also constify the RSA code and most things related to it. In a
6132 few places, most notable in the depth of the ASN.1 code, ugly
6133 casts back to non-const were required (to be solved at a later
6137 *) Make it so the openssl application has all engines loaded by default.
6140 *) Constify the BIGNUM routines a little more.
6143 *) Add the following functions:
6145 ENGINE_load_cswift()
6147 ENGINE_load_atalla()
6149 ENGINE_load_builtin_engines()
6151 That way, an application can itself choose if external engines that
6152 are built-in in OpenSSL shall ever be used or not. The benefit is
6153 that applications won't have to be linked with libdl or other dso
6154 libraries unless it's really needed.
6156 Changed 'openssl engine' to load all engines on demand.
6157 Changed the engine header files to avoid the duplication of some
6158 declarations (they differed!).
6161 *) 'openssl engine' can now list capabilities.
6164 *) Better error reporting in 'openssl engine'.
6167 *) Never call load_dh_param(NULL) in s_server.
6170 *) Add engine application. It can currently list engines by name and
6171 identity, and test if they are actually available.
6174 *) Improve RPM specification file by forcing symbolic linking and making
6175 sure the installed documentation is also owned by root.root.
6176 [Damien Miller <djm@mindrot.org>]
6178 *) Give the OpenSSL applications more possibilities to make use of
6179 keys (public as well as private) handled by engines.
6182 *) Add OCSP code that comes from CertCo.
6185 *) Add VMS support for the Rijndael code.
6188 *) Added untested support for Nuron crypto accelerator.
6191 *) Add support for external cryptographic devices. This code was
6192 previously distributed separately as the "engine" branch.
6193 [Geoff Thorpe, Richard Levitte]
6195 *) Rework the filename-translation in the DSO code. It is now possible to
6196 have far greater control over how a "name" is turned into a filename
6197 depending on the operating environment and any oddities about the
6198 different shared library filenames on each system.
6201 *) Support threads on FreeBSD-elf in Configure.
6204 *) Fix for SHA1 assembly problem with MASM: it produces
6205 warnings about corrupt line number information when assembling
6206 with debugging information. This is caused by the overlapping
6208 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6211 NCONF_get_number() has no error checking at all. As a replacement,
6212 NCONF_get_number_e() is defined (_e for "error checking") and is
6213 promoted strongly. The old NCONF_get_number is kept around for
6214 binary backward compatibility.
6215 Make it possible for methods to load from something other than a BIO,
6216 by providing a function pointer that is given a name instead of a BIO.
6217 For example, this could be used to load configuration data from an
6221 *) Fix for non blocking accept BIOs. Added new I/O special reason
6222 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6223 with non blocking I/O was not possible because no retry code was
6224 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6228 *) Added the beginnings of Rijndael support.
6231 *) Fix for bug in DirectoryString mask setting. Add support for
6232 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6233 to allow certificate printing to more controllable, additional
6234 'certopt' option to 'x509' to allow new printing options to be
6238 *) Clean old EAY MD5 hack from e_os.h.
6241 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6243 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6244 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6245 [Joe Orton, Steve Henson]
6247 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6249 *) Fix additional bug revealed by the NISCC test suite:
6251 Stop bug triggering large recursion when presented with
6252 certain ASN.1 tags (CVE-2003-0851)
6255 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6257 *) Fix various bugs revealed by running the NISCC test suite:
6259 Stop out of bounds reads in the ASN1 code when presented with
6260 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6262 If verify callback ignores invalid public key errors don't try to check
6263 certificate signature with the NULL public key.
6267 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6268 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6272 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6273 extra data after the compression methods not only for TLS 1.0
6274 but also for SSL 3.0 (as required by the specification).
6275 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6277 *) Change X509_certificate_type() to mark the key as exported/exportable
6278 when it's 512 *bits* long, not 512 bytes.
6281 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6283 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6284 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6285 a protocol version number mismatch like a decryption error
6286 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6289 *) Turn on RSA blinding by default in the default implementation
6290 to avoid a timing attack. Applications that don't want it can call
6291 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6292 They would be ill-advised to do so in most cases.
6293 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6295 *) Change RSA blinding code so that it works when the PRNG is not
6296 seeded (in this case, the secret RSA exponent is abused as
6297 an unpredictable seed -- if it is not unpredictable, there
6298 is no point in blinding anyway). Make RSA blinding thread-safe
6299 by remembering the creator's thread ID in rsa->blinding and
6300 having all other threads use local one-time blinding factors
6301 (this requires more computation than sharing rsa->blinding, but
6302 avoids excessive locking; and if an RSA object is not shared
6303 between threads, blinding will still be very fast).
6306 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6308 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6309 via timing by performing a MAC computation even if incorrrect
6310 block cipher padding has been found. This is a countermeasure
6311 against active attacks where the attacker has to distinguish
6312 between bad padding and a MAC verification error. (CVE-2003-0078)
6314 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6315 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6316 Martin Vuagnoux (EPFL, Ilion)]
6318 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6320 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6321 memory from it's contents. This is done with a counter that will
6322 place alternating values in each byte. This can be used to solve
6323 two issues: 1) the removal of calls to memset() by highly optimizing
6324 compilers, and 2) cleansing with other values than 0, since those can
6325 be read through on certain media, for example a swap space on disk.
6328 *) Bugfix: client side session caching did not work with external caching,
6329 because the session->cipher setting was not restored when reloading
6330 from the external cache. This problem was masked, when
6331 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6332 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6335 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6336 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6337 [Zeev Lieber <zeev-l@yahoo.com>]
6339 *) Undo an undocumented change introduced in 0.9.6e which caused
6340 repeated calls to OpenSSL_add_all_ciphers() and
6341 OpenSSL_add_all_digests() to be ignored, even after calling
6345 *) Change the default configuration reader to deal with last line not
6346 being properly terminated.
6349 *) Change X509_NAME_cmp() so it applies the special rules on handling
6350 DN values that are of type PrintableString, as well as RDNs of type
6351 emailAddress where the value has the type ia5String.
6352 [stefank@valicert.com via Richard Levitte]
6354 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6355 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6356 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6357 the bitwise-OR of the two for use by the majority of applications
6358 wanting this behaviour, and update the docs. The documented
6359 behaviour and actual behaviour were inconsistent and had been
6360 changing anyway, so this is more a bug-fix than a behavioural
6362 [Geoff Thorpe, diagnosed by Nadav Har'El]
6364 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6365 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6368 *) Fix initialization code race conditions in
6369 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6370 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6371 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6372 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6373 ssl2_get_cipher_by_char(),
6374 ssl3_get_cipher_by_char().
6375 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6377 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6378 the cached sessions are flushed, as the remove_cb() might use ex_data
6379 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6380 (see [openssl.org #212]).
6381 [Geoff Thorpe, Lutz Jaenicke]
6383 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6384 length, instead of the encoding length to d2i_ASN1_OBJECT.
6387 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6389 *) [In 0.9.6g-engine release:]
6390 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6391 [Lynn Gazis <lgazis@rainbow.com>]
6393 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6395 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6396 and get fix the header length calculation.
6397 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6398 Alon Kantor <alonk@checkpoint.com> (and others),
6401 *) Use proper error handling instead of 'assertions' in buffer
6402 overflow checks added in 0.9.6e. This prevents DoS (the
6403 assertions could call abort()).
6404 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6406 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6408 *) Add various sanity checks to asn1_get_length() to reject
6409 the ASN1 length bytes if they exceed sizeof(long), will appear
6410 negative or the content length exceeds the length of the
6412 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6414 *) Fix cipher selection routines: ciphers without encryption had no flags
6415 for the cipher strength set and where therefore not handled correctly
6416 by the selection routines (PR #130).
6419 *) Fix EVP_dsa_sha macro.
6423 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6424 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6425 that was added in OpenSSL 0.9.6d.
6427 As the countermeasure turned out to be incompatible with some
6428 broken SSL implementations, the new option is part of SSL_OP_ALL.
6429 SSL_OP_ALL is usually employed when compatibility with weird SSL
6430 implementations is desired (e.g. '-bugs' option to 's_client' and
6431 's_server'), so the new option is automatically set in many
6435 *) Changes in security patch:
6437 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6438 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6439 Air Force Materiel Command, USAF, under agreement number
6442 *) Add various sanity checks to asn1_get_length() to reject
6443 the ASN1 length bytes if they exceed sizeof(long), will appear
6444 negative or the content length exceeds the length of the
6445 supplied buffer. (CVE-2002-0659)
6446 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6448 *) Assertions for various potential buffer overflows, not known to
6450 [Ben Laurie (CHATS)]
6452 *) Various temporary buffers to hold ASCII versions of integers were
6453 too small for 64 bit platforms. (CVE-2002-0655)
6454 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6456 *) Remote buffer overflow in SSL3 protocol - an attacker could
6457 supply an oversized session ID to a client. (CVE-2002-0656)
6458 [Ben Laurie (CHATS)]
6460 *) Remote buffer overflow in SSL2 protocol - an attacker could
6461 supply an oversized client master key. (CVE-2002-0656)
6462 [Ben Laurie (CHATS)]
6464 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6466 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6467 encoded as NULL) with id-dsa-with-sha1.
6468 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6470 *) Check various X509_...() return values in apps/req.c.
6471 [Nils Larsch <nla@trustcenter.de>]
6473 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6474 an end-of-file condition would erronously be flagged, when the CRLF
6475 was just at the end of a processed block. The bug was discovered when
6476 processing data through a buffering memory BIO handing the data to a
6477 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6478 <ptsekov@syntrex.com> and Nedelcho Stanev.
6481 *) Implement a countermeasure against a vulnerability recently found
6482 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6483 before application data chunks to avoid the use of known IVs
6484 with data potentially chosen by the attacker.
6487 *) Fix length checks in ssl3_get_client_hello().
6490 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6491 to prevent ssl3_read_internal() from incorrectly assuming that
6492 ssl3_read_bytes() found application data while handshake
6493 processing was enabled when in fact s->s3->in_read_app_data was
6494 merely automatically cleared during the initial handshake.
6495 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6497 *) Fix object definitions for Private and Enterprise: they were not
6498 recognized in their shortname (=lowercase) representation. Extend
6499 obj_dat.pl to issue an error when using undefined keywords instead
6500 of silently ignoring the problem (Svenning Sorensen
6501 <sss@sss.dnsalias.net>).
6504 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6505 generators, i.e. generators other than 2 and 5. (Previously, the
6506 code did not properly initialise the 'add' and 'rem' values to
6507 BN_generate_prime().)
6509 In the new general case, we do not insist that 'generator' is
6510 actually a primitive root: This requirement is rather pointless;
6511 a generator of the order-q subgroup is just as good, if not
6515 *) Map new X509 verification errors to alerts. Discovered and submitted by
6516 Tom Wu <tom@arcot.com>.
6519 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6520 returning non-zero before the data has been completely received
6521 when using non-blocking I/O.
6522 [Bodo Moeller; problem pointed out by John Hughes]
6524 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6525 [Ben Laurie, Lutz Jaenicke]
6527 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6528 Yoram Zahavi <YoramZ@gilian.com>).
6531 *) Add information about CygWin 1.3 and on, and preserve proper
6532 configuration for the versions before that.
6533 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6535 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6536 check whether we deal with a copy of a session and do not delete from
6537 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6538 <izhar@checkpoint.com>.
6541 *) Do not store session data into the internal session cache, if it
6542 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6543 flag is set). Proposed by Aslam <aslam@funk.com>.
6546 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6550 *) [In 0.9.6d-engine release:]
6551 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6552 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6554 *) Add the configuration target linux-s390x.
6555 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6557 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6558 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6559 variable as an indication that a ClientHello message has been
6560 received. As the flag value will be lost between multiple
6561 invocations of ssl3_accept when using non-blocking I/O, the
6562 function may not be aware that a handshake has actually taken
6563 place, thus preventing a new session from being added to the
6566 To avoid this problem, we now set s->new_session to 2 instead of
6567 using a local variable.
6568 [Lutz Jaenicke, Bodo Moeller]
6570 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6571 if the SSL_R_LENGTH_MISMATCH error is detected.
6572 [Geoff Thorpe, Bodo Moeller]
6574 *) New 'shared_ldflag' column in Configure platform table.
6577 *) Fix EVP_CIPHER_mode macro.
6578 ["Dan S. Camper" <dan@bti.net>]
6580 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6581 type, we must throw them away by setting rr->length to 0.
6582 [D P Chang <dpc@qualys.com>]
6584 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6586 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6587 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6588 worked incorrectly for those cases where range = 10..._2 and
6589 3*range is two bits longer than range.)
6592 *) Only add signing time to PKCS7 structures if it is not already
6596 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6597 OBJ_ld_ce should be OBJ_id_ce.
6598 Also some ip-pda OIDs in crypto/objects/objects.txt were
6599 incorrect (cf. RFC 3039).
6600 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6602 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6603 returns early because it has nothing to do.
6604 [Andy Schneider <andy.schneider@bjss.co.uk>]
6606 *) [In 0.9.6c-engine release:]
6607 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6608 [Andy Schneider <andy.schneider@bjss.co.uk>]
6610 *) [In 0.9.6c-engine release:]
6611 Add support for Cryptographic Appliance's keyserver technology.
6612 (Use engine 'keyclient')
6613 [Cryptographic Appliances and Geoff Thorpe]
6615 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6616 is called via tools/c89.sh because arguments have to be
6617 rearranged (all '-L' options must appear before the first object
6619 [Richard Shapiro <rshapiro@abinitio.com>]
6621 *) [In 0.9.6c-engine release:]
6622 Add support for Broadcom crypto accelerator cards, backported
6624 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6626 *) [In 0.9.6c-engine release:]
6627 Add support for SureWare crypto accelerator cards from
6628 Baltimore Technologies. (Use engine 'sureware')
6629 [Baltimore Technologies and Mark Cox]
6631 *) [In 0.9.6c-engine release:]
6632 Add support for crypto accelerator cards from Accelerated
6633 Encryption Processing, www.aep.ie. (Use engine 'aep')
6634 [AEP Inc. and Mark Cox]
6636 *) Add a configuration entry for gcc on UnixWare.
6637 [Gary Benson <gbenson@redhat.com>]
6639 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6640 messages are stored in a single piece (fixed-length part and
6641 variable-length part combined) and fix various bugs found on the way.
6644 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6645 instead. BIO_gethostbyname() does not know what timeouts are
6646 appropriate, so entries would stay in cache even when they have
6648 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6650 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6651 faced with a pathologically small ClientHello fragment that does
6652 not contain client_version: Instead of aborting with an error,
6653 simply choose the highest available protocol version (i.e.,
6654 TLS 1.0 unless it is disabled). In practice, ClientHello
6655 messages are never sent like this, but this change gives us
6656 strictly correct behaviour at least for TLS.
6659 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6660 never resets s->method to s->ctx->method when called from within
6661 one of the SSL handshake functions.
6662 [Bodo Moeller; problem pointed out by Niko Baric]
6664 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6665 (sent using the client's version number) if client_version is
6666 smaller than the protocol version in use. Also change
6667 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6668 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6669 the client will at least see that alert.
6672 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6676 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6677 client receives HelloRequest while in a handshake.
6678 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6680 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6681 should end in 'break', not 'goto end' which circuments various
6682 cleanups done in state SSL_ST_OK. But session related stuff
6683 must be disabled for SSL_ST_OK in the case that we just sent a
6686 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6687 before just sending a HelloRequest.
6688 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6690 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6691 reveal whether illegal block cipher padding was found or a MAC
6692 verification error occured. (Neither SSLerr() codes nor alerts
6693 are directly visible to potential attackers, but the information
6694 may leak via logfiles.)
6696 Similar changes are not required for the SSL 2.0 implementation
6697 because the number of padding bytes is sent in clear for SSL 2.0,
6698 and the extra bytes are just ignored. However ssl/s2_pkt.c
6699 failed to verify that the purported number of padding bytes is in
6703 *) Add OpenUNIX-8 support including shared libraries
6704 (Boyd Lynn Gerber <gerberb@zenez.com>).
6707 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6708 'wristwatch attack' using huge encoding parameters (cf.
6709 James H. Manger's CRYPTO 2001 paper). Note that the
6710 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6711 encoding parameters and hence was not vulnerable.
6714 *) BN_sqr() bug fix.
6715 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6717 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6718 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6719 followed by modular reduction.
6720 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6722 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6723 equivalent based on BN_pseudo_rand() instead of BN_rand().
6726 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6727 This function was broken, as the check for a new client hello message
6728 to handle SGC did not allow these large messages.
6729 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6732 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6735 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6736 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6739 *) Rework the configuration and shared library support for Tru64 Unix.
6740 The configuration part makes use of modern compiler features and
6741 still retains old compiler behavior for those that run older versions
6742 of the OS. The shared library support part includes a variant that
6743 uses the RPATH feature, and is available through the special
6744 configuration target "alpha-cc-rpath", which will never be selected
6746 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6748 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6749 with the same message size as in ssl3_get_certificate_request().
6750 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6751 messages might inadvertently be reject as too long.
6752 [Petr Lampa <lampa@fee.vutbr.cz>]
6754 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6757 *) Modified SSL library such that the verify_callback that has been set
6758 specificly for an SSL object with SSL_set_verify() is actually being
6759 used. Before the change, a verify_callback set with this function was
6760 ignored and the verify_callback() set in the SSL_CTX at the time of
6761 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6762 to allow the necessary settings.
6765 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6766 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6767 done automatically (in contradiction to the requirements of the C
6768 standard). This made problems when used from OpenSSH.
6771 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6772 dh->length and always used
6774 BN_rand_range(priv_key, dh->p).
6776 BN_rand_range() is not necessary for Diffie-Hellman, and this
6777 specific range makes Diffie-Hellman unnecessarily inefficient if
6778 dh->length (recommended exponent length) is much smaller than the
6779 length of dh->p. We could use BN_rand_range() if the order of
6780 the subgroup was stored in the DH structure, but we only have
6785 BN_rand(priv_key, l, ...)
6787 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6793 RSA_eay_public_encrypt
6794 RSA_eay_private_decrypt
6795 RSA_eay_private_encrypt (signing)
6796 RSA_eay_public_decrypt (signature verification)
6798 (default implementations for RSA_public_encrypt,
6799 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6800 always reject numbers >= n.
6803 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6804 to synchronize access to 'locking_thread'. This is necessary on
6805 systems where access to 'locking_thread' (an 'unsigned long'
6806 variable) is not atomic.
6809 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6810 *before* setting the 'crypto_lock_rand' flag. The previous code had
6811 a race condition if 0 is a valid thread ID.
6812 [Travis Vitek <vitek@roguewave.com>]
6814 *) Add support for shared libraries under Irix.
6815 [Albert Chin-A-Young <china@thewrittenword.com>]
6817 *) Add configuration option to build on Linux on both big-endian and
6819 [Ralf Baechle <ralf@uni-koblenz.de>]
6821 *) Add the possibility to create shared libraries on HP-UX.
6824 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6826 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6827 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6828 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6829 PRNG state recovery was possible based on the output of
6830 one PRNG request appropriately sized to gain knowledge on
6831 'md' followed by enough consecutive 1-byte PRNG requests
6832 to traverse all of 'state'.
6834 1. When updating 'md_local' (the current thread's copy of 'md')
6835 during PRNG output generation, hash all of the previous
6836 'md_local' value, not just the half used for PRNG output.
6838 2. Make the number of bytes from 'state' included into the hash
6839 independent from the number of PRNG bytes requested.
6841 The first measure alone would be sufficient to avoid
6842 Markku-Juhani's attack. (Actually it had never occurred
6843 to me that the half of 'md_local' used for chaining was the
6844 half from which PRNG output bytes were taken -- I had always
6845 assumed that the secret half would be used.) The second
6846 measure makes sure that additional data from 'state' is never
6847 mixed into 'md_local' in small portions; this heuristically
6848 further strengthens the PRNG.
6851 *) Fix crypto/bn/asm/mips3.s.
6854 *) When only the key is given to "enc", the IV is undefined. Print out
6855 an error message in this case.
6858 *) Handle special case when X509_NAME is empty in X509 printing routines.
6861 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6862 positive and less than q.
6865 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6866 used: it isn't thread safe and the add_lock_callback should handle
6868 [Paul Rose <Paul.Rose@bridge.com>]
6870 *) Verify that incoming data obeys the block size in
6871 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6875 [Ulf Möller, Bodo Möller]
6877 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6878 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6879 when fixing the server behaviour for backwards-compatible 'client
6880 hello' messages. (Note that the attack is impractical against
6881 SSL 3.0 and TLS 1.0 anyway because length and version checking
6882 means that the probability of guessing a valid ciphertext is
6883 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6886 Before 0.9.5, the countermeasure (hide the error by generating a
6887 random 'decryption result') did not work properly because
6888 ERR_clear_error() was missing, meaning that SSL_get_error() would
6889 detect the supposedly ignored error.
6891 Both problems are now fixed.
6894 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6895 (previously it was 1024).
6898 *) Fix for compatibility mode trust settings: ignore trust settings
6899 unless some valid trust or reject settings are present.
6902 *) Fix for blowfish EVP: its a variable length cipher.
6905 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6906 parameters in DSA public key structures and return an error in the
6907 DSA routines if parameters are absent.
6910 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6911 in the current directory if neither $RANDFILE nor $HOME was set.
6912 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6913 caused some confusion to Windows users who haven't defined $HOME.
6914 Thus RAND_file_name() is changed again: e_os.h can define a
6915 DEFAULT_HOME, which will be used if $HOME is not set.
6916 For Windows, we use "C:"; on other platforms, we still require
6917 environment variables.
6919 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6920 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6921 having multiple threads call RAND_poll() concurrently.
6924 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6925 combination of a flag and a thread ID variable.
6926 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6927 flag), *other* threads can enter ssleay_add_bytes without obeying
6928 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6929 that they do not hold after the first thread unsets add_do_not_lock).
6932 *) Change bctest again: '-x' expressions are not available in all
6936 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6938 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6939 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6941 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6942 the default extension for executables, if any. Also, make the perl
6943 scripts that use symlink() to test if it really exists and use "cp"
6944 if it doesn't. All this made OpenSSL compilable and installable in
6948 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6949 If SEQUENCE is length is indefinite just set c->slen to the total
6950 amount of data available.
6951 [Steve Henson, reported by shige@FreeBSD.org]
6952 [This change does not apply to 0.9.7.]
6954 *) Change bctest to avoid here-documents inside command substitution
6955 (workaround for FreeBSD /bin/sh bug).
6956 For compatibility with Ultrix, avoid shell functions (introduced
6957 in the bctest version that searches along $PATH).
6960 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6961 with des_encrypt() defined on some operating systems, like Solaris
6965 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6966 On the Importance of Eliminating Errors in Cryptographic
6967 Computations, J. Cryptology 14 (2001) 2, 101-119,
6968 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6971 *) MIPS assembler BIGNUM division bug fix.
6974 *) Disabled incorrect Alpha assembler code.
6977 *) Fix PKCS#7 decode routines so they correctly update the length
6978 after reading an EOC for the EXPLICIT tag.
6980 [This change does not apply to 0.9.7.]
6982 *) Fix bug in PKCS#12 key generation routines. This was triggered
6983 if a 3DES key was generated with a 0 initial byte. Include
6984 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6985 (but broken) behaviour.
6988 *) Enhance bctest to search for a working bc along $PATH and print
6990 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6992 *) Fix memory leaks in err.c: free err_data string if necessary;
6993 don't write to the wrong index in ERR_set_error_data.
6996 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7000 *) Replace rdtsc with _emit statements for VC++ version 5.
7001 [Jeremy Cooper <jeremy@baymoo.org>]
7003 *) Make it possible to reuse SSLv2 sessions.
7006 *) In copy_email() check for >= 0 as a return value for
7007 X509_NAME_get_index_by_NID() since 0 is a valid index.
7008 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7010 *) Avoid coredump with unsupported or invalid public keys by checking if
7011 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7012 PKCS7_verify() fails with non detached data.
7015 *) Don't use getenv in library functions when run as setuid/setgid.
7016 New function OPENSSL_issetugid().
7019 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7020 due to incorrect handling of multi-threading:
7022 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7024 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7026 3. Count how many times MemCheck_off() has been called so that
7027 nested use can be treated correctly. This also avoids
7028 inband-signalling in the previous code (which relied on the
7029 assumption that thread ID 0 is impossible).
7032 *) Add "-rand" option also to s_client and s_server.
7035 *) Fix CPU detection on Irix 6.x.
7036 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7037 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7039 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7042 [This change does not apply to 0.9.7.]
7044 *) Use the cached encoding of an X509_NAME structure rather than
7045 copying it. This is apparently the reason for the libsafe "errors"
7046 but the code is actually correct.
7049 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7050 Bleichenbacher's DSA attack.
7051 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7052 to be set and top=0 forces the highest bit to be set; top=-1 is new
7053 and leaves the highest bit random.
7054 [Ulf Moeller, Bodo Moeller]
7056 *) In the NCONF_...-based implementations for CONF_... queries
7057 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7058 a temporary CONF structure with the data component set to NULL
7059 (which gives segmentation faults in lh_retrieve).
7060 Instead, use NULL for the CONF pointer in CONF_get_string and
7061 CONF_get_number (which may use environment variables) and directly
7062 return NULL from CONF_get_section.
7065 *) Fix potential buffer overrun for EBCDIC.
7068 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7069 keyUsage if basicConstraints absent for a CA.
7072 *) Make SMIME_write_PKCS7() write mail header values with a format that
7073 is more generally accepted (no spaces before the semicolon), since
7074 some programs can't parse those values properly otherwise. Also make
7075 sure BIO's that break lines after each write do not create invalid
7079 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7080 macros previously used would not encode an empty SEQUENCE OF
7081 and break the signature.
7083 [This change does not apply to 0.9.7.]
7085 *) Zero the premaster secret after deriving the master secret in
7089 *) Add some EVP_add_digest_alias registrations (as found in
7090 OpenSSL_add_all_digests()) to SSL_library_init()
7091 aka OpenSSL_add_ssl_algorithms(). This provides improved
7092 compatibility with peers using X.509 certificates
7093 with unconventional AlgorithmIdentifier OIDs.
7096 *) Fix for Irix with NO_ASM.
7097 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7099 *) ./config script fixes.
7100 [Ulf Moeller, Richard Levitte]
7102 *) Fix 'openssl passwd -1'.
7105 *) Change PKCS12_key_gen_asc() so it can cope with non null
7106 terminated strings whose length is passed in the passlen
7107 parameter, for example from PEM callbacks. This was done
7108 by adding an extra length parameter to asc2uni().
7109 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7111 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7112 call failed, free the DSA structure.
7115 *) Fix to uni2asc() to cope with zero length Unicode strings.
7116 These are present in some PKCS#12 files.
7119 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7120 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7121 when writing a 32767 byte record.
7122 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7124 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7125 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7127 (RSA objects have a reference count access to which is protected
7128 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7129 so they are meant to be shared between threads.)
7130 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7131 "Reddie, Steven" <Steven.Reddie@ca.com>]
7133 *) Fix a deadlock in CRYPTO_mem_leaks().
7136 *) Use better test patterns in bntest.
7139 *) rand_win.c fix for Borland C.
7142 *) BN_rshift bugfix for n == 0.
7145 *) Add a 'bctest' script that checks for some known 'bc' bugs
7146 so that 'make test' does not abort just because 'bc' is broken.
7149 *) Store verify_result within SSL_SESSION also for client side to
7150 avoid potential security hole. (Re-used sessions on the client side
7151 always resulted in verify_result==X509_V_OK, not using the original
7152 result of the server certificate verification.)
7155 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7156 SSL3_RT_APPLICATION_DATA, return 0.
7157 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7161 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7162 releases, have been re-implemented by renaming the previous
7163 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7164 and ssl3_read_internal, respectively, and adding 'peek' parameters
7165 to them. The new ssl[23]_{read,peek} functions are calls to
7166 ssl[23]_read_internal with the 'peek' flag set appropriately.
7167 A 'peek' parameter has also been added to ssl3_read_bytes, which
7168 does the actual work for ssl3_read_internal.
7171 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7172 the method-specific "init()" handler. Also clean up ex_data after
7173 calling the method-specific "finish()" handler. Previously, this was
7174 happening the other way round.
7177 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7178 The previous value, 12, was not always sufficient for BN_mod_exp().
7181 *) Make sure that shared libraries get the internal name engine with
7182 the full version number and not just 0. This should mark the
7183 shared libraries as not backward compatible. Of course, this should
7184 be changed again when we can guarantee backward binary compatibility.
7187 *) Fix typo in get_cert_by_subject() in by_dir.c
7188 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7190 *) Rework the system to generate shared libraries:
7192 - Make note of the expected extension for the shared libraries and
7193 if there is a need for symbolic links from for example libcrypto.so.0
7194 to libcrypto.so.0.9.7. There is extended info in Configure for
7197 - Make as few rebuilds of the shared libraries as possible.
7199 - Still avoid linking the OpenSSL programs with the shared libraries.
7201 - When installing, install the shared libraries separately from the
7205 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7207 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7208 and not in SSL_clear because the latter is also used by the
7209 accept/connect functions; previously, the settings made by
7210 SSL_set_read_ahead would be lost during the handshake.
7211 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7213 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7214 Previously, it would create entries for disableed algorithms no
7218 *) Added several new manual pages for SSL_* function.
7221 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7223 *) In ssl23_get_client_hello, generate an error message when faced
7224 with an initial SSL 3.0/TLS record that is too small to contain the
7225 first two bytes of the ClientHello message, i.e. client_version.
7226 (Note that this is a pathologic case that probably has never happened
7227 in real life.) The previous approach was to use the version number
7228 from the record header as a substitute; but our protocol choice
7229 should not depend on that one because it is not authenticated
7230 by the Finished messages.
7233 *) More robust randomness gathering functions for Windows.
7234 [Jeffrey Altman <jaltman@columbia.edu>]
7236 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7237 not set then we don't setup the error code for issuer check errors
7238 to avoid possibly overwriting other errors which the callback does
7239 handle. If an application does set the flag then we assume it knows
7240 what it is doing and can handle the new informational codes
7244 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7245 a general "ANY" type, as such it should be able to decode anything
7246 including tagged types. However it didn't check the class so it would
7247 wrongly interpret tagged types in the same way as their universal
7248 counterpart and unknown types were just rejected. Changed so that the
7249 tagged and unknown types are handled in the same way as a SEQUENCE:
7250 that is the encoding is stored intact. There is also a new type
7251 "V_ASN1_OTHER" which is used when the class is not universal, in this
7252 case we have no idea what the actual type is so we just lump them all
7256 *) On VMS, stdout may very well lead to a file that is written to
7257 in a record-oriented fashion. That means that every write() will
7258 write a separate record, which will be read separately by the
7259 programs trying to read from it. This can be very confusing.
7261 The solution is to put a BIO filter in the way that will buffer
7262 text until a linefeed is reached, and then write everything a
7263 line at a time, so every record written will be an actual line,
7264 not chunks of lines and not (usually doesn't happen, but I've
7265 seen it once) several lines in one record. BIO_f_linebuffer() is
7268 Currently, it's a VMS-only method, because that's where it has
7269 been tested well enough.
7272 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7273 it can return incorrect results.
7274 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7275 but it was in 0.9.6-beta[12].)
7278 *) Disable the check for content being present when verifying detached
7279 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7280 include zero length content when signing messages.
7283 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7284 BIO_ctrl (for BIO pairs).
7287 *) Add DSO method for VMS.
7290 *) Bug fix: Montgomery multiplication could produce results with the
7294 *) Add RPM specification openssl.spec and modify it to build three
7295 packages. The default package contains applications, application
7296 documentation and run-time libraries. The devel package contains
7297 include files, static libraries and function documentation. The
7298 doc package contains the contents of the doc directory. The original
7299 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7302 *) Add a large number of documentation files for many SSL routines.
7303 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7305 *) Add a configuration entry for Sony News 4.
7306 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7308 *) Don't set the two most significant bits to one when generating a
7309 random number < q in the DSA library.
7312 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7313 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7314 the underlying transport is blocking) if a handshake took place.
7315 (The default behaviour is needed by applications such as s_client
7316 and s_server that use select() to determine when to use SSL_read;
7317 but for applications that know in advance when to expect data, it
7318 just makes things more complicated.)
7321 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7325 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7326 work better on such systems.
7327 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7329 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7330 Update PKCS12_parse() so it copies the friendlyName and the
7331 keyid to the certificates aux info.
7334 *) Fix bug in PKCS7_verify() which caused an infinite loop
7335 if there was more than one signature.
7336 [Sven Uszpelkat <su@celocom.de>]
7338 *) Major change in util/mkdef.pl to include extra information
7339 about each symbol, as well as presentig variables as well
7340 as functions. This change means that there's n more need
7341 to rebuild the .num files when some algorithms are excluded.
7344 *) Allow the verify time to be set by an application,
7345 rather than always using the current time.
7348 *) Phase 2 verify code reorganisation. The certificate
7349 verify code now looks up an issuer certificate by a
7350 number of criteria: subject name, authority key id
7351 and key usage. It also verifies self signed certificates
7352 by the same criteria. The main comparison function is
7353 X509_check_issued() which performs these checks.
7355 Lot of changes were necessary in order to support this
7356 without completely rewriting the lookup code.
7358 Authority and subject key identifier are now cached.
7360 The LHASH 'certs' is X509_STORE has now been replaced
7361 by a STACK_OF(X509_OBJECT). This is mainly because an
7362 LHASH can't store or retrieve multiple objects with
7363 the same hash value.
7365 As a result various functions (which were all internal
7366 use only) have changed to handle the new X509_STORE
7367 structure. This will break anything that messed round
7368 with X509_STORE internally.
7370 The functions X509_STORE_add_cert() now checks for an
7371 exact match, rather than just subject name.
7373 The X509_STORE API doesn't directly support the retrieval
7374 of multiple certificates matching a given criteria, however
7375 this can be worked round by performing a lookup first
7376 (which will fill the cache with candidate certificates)
7377 and then examining the cache for matches. This is probably
7378 the best we can do without throwing out X509_LOOKUP
7379 entirely (maybe later...).
7381 The X509_VERIFY_CTX structure has been enhanced considerably.
7383 All certificate lookup operations now go via a get_issuer()
7384 callback. Although this currently uses an X509_STORE it
7385 can be replaced by custom lookups. This is a simple way
7386 to bypass the X509_STORE hackery necessary to make this
7387 work and makes it possible to use more efficient techniques
7388 in future. A very simple version which uses a simple
7389 STACK for its trusted certificate store is also provided
7390 using X509_STORE_CTX_trusted_stack().
7392 The verify_cb() and verify() callbacks now have equivalents
7393 in the X509_STORE_CTX structure.
7395 X509_STORE_CTX also has a 'flags' field which can be used
7396 to customise the verify behaviour.
7399 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7400 excludes S/MIME capabilities.
7403 *) When a certificate request is read in keep a copy of the
7404 original encoding of the signed data and use it when outputing
7405 again. Signatures then use the original encoding rather than
7406 a decoded, encoded version which may cause problems if the
7407 request is improperly encoded.
7410 *) For consistency with other BIO_puts implementations, call
7411 buffer_write(b, ...) directly in buffer_puts instead of calling
7414 In BIO_puts, increment b->num_write as in BIO_write.
7415 [Peter.Sylvester@EdelWeb.fr]
7417 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7418 BN_zero, we may not return a BIGNUM with an array consisting of
7422 *) Avoid calling abort() from within the library when problems are
7423 detected, except if preprocessor symbols have been defined
7424 (such as REF_CHECK, BN_DEBUG etc.).
7427 *) New openssl application 'rsautl'. This utility can be
7428 used for low level RSA operations. DER public key
7429 BIO/fp routines also added.
7432 *) New Configure entry and patches for compiling on QNX 4.
7433 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7435 *) A demo state-machine implementation was sponsored by
7436 Nuron (http://www.nuron.com/) and is now available in
7437 demos/state_machine.
7440 *) New options added to the 'dgst' utility for signature
7441 generation and verification.
7444 *) Unrecognized PKCS#7 content types are now handled via a
7445 catch all ASN1_TYPE structure. This allows unsupported
7446 types to be stored as a "blob" and an application can
7447 encode and decode it manually.
7450 *) Fix various signed/unsigned issues to make a_strex.c
7452 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7454 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7455 length if passed a buffer. ASN1_INTEGER_to_BN failed
7456 if passed a NULL BN and its argument was negative.
7457 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7459 *) Modification to PKCS#7 encoding routines to output definite
7460 length encoding. Since currently the whole structures are in
7461 memory there's not real point in using indefinite length
7462 constructed encoding. However if OpenSSL is compiled with
7463 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7466 *) Added BIO_vprintf() and BIO_vsnprintf().
7469 *) Added more prefixes to parse for in the the strings written
7470 through a logging bio, to cover all the levels that are available
7471 through syslog. The prefixes are now:
7473 PANIC, EMERG, EMR => LOG_EMERG
7474 ALERT, ALR => LOG_ALERT
7475 CRIT, CRI => LOG_CRIT
7476 ERROR, ERR => LOG_ERR
7477 WARNING, WARN, WAR => LOG_WARNING
7478 NOTICE, NOTE, NOT => LOG_NOTICE
7479 INFO, INF => LOG_INFO
7480 DEBUG, DBG => LOG_DEBUG
7482 and as before, if none of those prefixes are present at the
7483 beginning of the string, LOG_ERR is chosen.
7485 On Win32, the LOG_* levels are mapped according to this:
7487 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7488 LOG_WARNING => EVENTLOG_WARNING_TYPE
7489 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7493 *) Made it possible to reconfigure with just the configuration
7494 argument "reconf" or "reconfigure". The command line arguments
7495 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7496 and are retrieved from there when reconfiguring.
7500 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7502 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7505 *) The obj_dat.pl script was messing up the sorting of object
7506 names. The reason was that it compared the quoted version
7507 of strings as a result "OCSP" > "OCSP Signing" because
7508 " > SPACE. Changed script to store unquoted versions of
7509 names and add quotes on output. It was also omitting some
7510 names from the lookup table if they were given a default
7511 value (that is if SN is missing it is given the same
7512 value as LN and vice versa), these are now added on the
7513 grounds that if an object has a name we should be able to
7514 look it up. Finally added warning output when duplicate
7515 short or long names are found.
7518 *) Changes needed for Tandem NSK.
7519 [Scott Uroff <scott@xypro.com>]
7521 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7522 RSA_padding_check_SSLv23(), special padding was never detected
7523 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7524 version rollback attacks was not effective.
7526 In s23_clnt.c, don't use special rollback-attack detection padding
7527 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7528 client; similarly, in s23_srvr.c, don't do the rollback check if
7529 SSL 2.0 is the only protocol enabled in the server.
7532 *) Make it possible to get hexdumps of unprintable data with 'openssl
7533 asn1parse'. By implication, the functions ASN1_parse_dump() and
7534 BIO_dump_indent() are added.
7537 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7538 these print out strings and name structures based on various
7539 flags including RFC2253 support and proper handling of
7540 multibyte characters. Added options to the 'x509' utility
7541 to allow the various flags to be set.
7544 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7545 Also change the functions X509_cmp_current_time() and
7546 X509_gmtime_adj() work with an ASN1_TIME structure,
7547 this will enable certificates using GeneralizedTime in validity
7548 dates to be checked.
7551 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7552 negative public key encodings) on by default,
7553 NO_NEG_PUBKEY_BUG can be set to disable it.
7556 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7557 content octets. An i2c_ASN1_OBJECT is unnecessary because
7558 the encoding can be trivially obtained from the structure.
7561 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7562 not read locks (CRYPTO_r_[un]lock).
7565 *) A first attempt at creating official support for shared
7566 libraries through configuration. I've kept it so the
7567 default is static libraries only, and the OpenSSL programs
7568 are always statically linked for now, but there are
7569 preparations for dynamic linking in place.
7570 This has been tested on Linux and Tru64.
7573 *) Randomness polling function for Win9x, as described in:
7574 Peter Gutmann, Software Generation of Practically Strong
7578 *) Fix so PRNG is seeded in req if using an already existing
7582 *) New options to smime application. -inform and -outform
7583 allow alternative formats for the S/MIME message including
7584 PEM and DER. The -content option allows the content to be
7585 specified separately. This should allow things like Netscape
7586 form signing output easier to verify.
7589 *) Fix the ASN1 encoding of tags using the 'long form'.
7592 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7593 STRING types. These convert content octets to and from the
7594 underlying type. The actual tag and length octets are
7595 already assumed to have been read in and checked. These
7596 are needed because all other string types have virtually
7597 identical handling apart from the tag. By having versions
7598 of the ASN1 functions that just operate on content octets
7599 IMPLICIT tagging can be handled properly. It also allows
7600 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7601 and ASN1_INTEGER are identical apart from the tag.
7604 *) Change the handling of OID objects as follows:
7606 - New object identifiers are inserted in objects.txt, following
7607 the syntax given in objects.README.
7608 - objects.pl is used to process obj_mac.num and create a new
7610 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7613 This is currently kind of a hack, and the perl code in objects.pl
7614 isn't very elegant, but it works as I intended. The simplest way
7615 to check that it worked correctly is to look in obj_dat.h and
7616 check the array nid_objs and make sure the objects haven't moved
7617 around (this is important!). Additions are OK, as well as
7618 consistent name changes.
7621 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7624 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7625 The given file adds to whatever has already been seeded into the
7626 random pool through the RANDFILE configuration file option or
7627 environment variable, or the default random state file.
7630 *) mkstack.pl now sorts each macro group into lexical order.
7631 Previously the output order depended on the order the files
7632 appeared in the directory, resulting in needless rewriting
7636 *) Patches to make OpenSSL compile under Win32 again. Mostly
7637 work arounds for the VC++ problem that it treats func() as
7638 func(void). Also stripped out the parts of mkdef.pl that
7639 added extra typesafe functions: these no longer exist.
7642 *) Reorganisation of the stack code. The macros are now all
7643 collected in safestack.h . Each macro is defined in terms of
7644 a "stack macro" of the form SKM_<name>(type, a, b). The
7645 DEBUG_SAFESTACK is now handled in terms of function casts,
7646 this has the advantage of retaining type safety without the
7647 use of additional functions. If DEBUG_SAFESTACK is not defined
7648 then the non typesafe macros are used instead. Also modified the
7649 mkstack.pl script to handle the new form. Needs testing to see
7650 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7651 the default if no major problems. Similar behaviour for ASN1_SET_OF
7652 and PKCS12_STACK_OF.
7655 *) When some versions of IIS use the 'NET' form of private key the
7656 key derivation algorithm is different. Normally MD5(password) is
7657 used as a 128 bit RC4 key. In the modified case
7658 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7659 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7660 as the old Netscape_RSA functions except they have an additional
7661 'sgckey' parameter which uses the modified algorithm. Also added
7662 an -sgckey command line option to the rsa utility. Thanks to
7663 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7664 algorithm to openssl-dev.
7667 *) The evp_local.h macros were using 'c.##kname' which resulted in
7668 invalid expansion on some systems (SCO 5.0.5 for example).
7669 Corrected to 'c.kname'.
7670 [Phillip Porch <root@theporch.com>]
7672 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7673 a STACK of email addresses from a certificate or request, these look
7674 in the subject name and the subject alternative name extensions and
7675 omit any duplicate addresses.
7678 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7679 This makes DSA verification about 2 % faster.
7682 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7683 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7684 plus overhead for 1024 bit moduli).
7685 This makes exponentiations about 0.5 % faster for 1024 bit
7686 exponents (as measured by "openssl speed rsa2048").
7689 *) Rename memory handling macros to avoid conflicts with other
7691 Malloc => OPENSSL_malloc
7692 Malloc_locked => OPENSSL_malloc_locked
7693 Realloc => OPENSSL_realloc
7694 Free => OPENSSL_free
7697 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7698 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7701 *) CygWin32 support.
7702 [John Jarvie <jjarvie@newsguy.com>]
7704 *) The type-safe stack code has been rejigged. It is now only compiled
7705 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7706 by default all type-specific stack functions are "#define"d back to
7707 standard stack functions. This results in more streamlined output
7708 but retains the type-safety checking possibilities of the original
7712 *) The STACK code has been cleaned up, and certain type declarations
7713 that didn't make a lot of sense have been brought in line. This has
7714 also involved a cleanup of sorts in safestack.h to more correctly
7715 map type-safe stack functions onto their plain stack counterparts.
7716 This work has also resulted in a variety of "const"ifications of
7717 lots of the code, especially "_cmp" operations which should normally
7718 be prototyped with "const" parameters anyway.
7721 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7722 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7723 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7724 where all of 'md' is used each time the PRNG is used, but 'state'
7725 is used only indexed by a cyclic counter. As entropy may not be
7726 well distributed from the beginning, 'md' is important as a
7727 chaining variable. However, the output function chains only half
7728 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7729 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7730 in all of 'state' being rewritten, with the new values depending
7731 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7734 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7735 the handshake is continued after ssl_verify_cert_chain();
7736 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7737 can lead to 'unexplainable' connection aborts later.
7738 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7740 *) Major EVP API cipher revision.
7741 Add hooks for extra EVP features. This allows various cipher
7742 parameters to be set in the EVP interface. Support added for variable
7743 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7744 setting of RC2 and RC5 parameters.
7746 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7749 Remove lots of duplicated code from the EVP library. For example *every*
7750 cipher init() function handles the 'iv' in the same way according to the
7751 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7752 for CFB and OFB modes they zero ctx->num.
7754 New functionality allows removal of S/MIME code RC2 hack.
7756 Most of the routines have the same form and so can be declared in terms
7759 By shifting this to the top level EVP_CipherInit() it can be removed from
7760 all individual ciphers. If the cipher wants to handle IVs or keys
7761 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7764 Change lots of functions like EVP_EncryptUpdate() to now return a
7765 value: although software versions of the algorithms cannot fail
7766 any installed hardware versions can.
7769 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7770 this option is set, tolerate broken clients that send the negotiated
7771 protocol version number instead of the requested protocol version
7775 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7776 i.e. non-zero for export ciphersuites, zero otherwise.
7777 Previous versions had this flag inverted, inconsistent with
7778 rsa_tmp_cb (..._TMP_RSA_CB).
7779 [Bodo Moeller; problem reported by Amit Chopra]
7781 *) Add missing DSA library text string. Work around for some IIS
7782 key files with invalid SEQUENCE encoding.
7785 *) Add a document (doc/standards.txt) that list all kinds of standards
7786 and so on that are implemented in OpenSSL.
7789 *) Enhance c_rehash script. Old version would mishandle certificates
7790 with the same subject name hash and wouldn't handle CRLs at all.
7791 Added -fingerprint option to crl utility, to support new c_rehash
7795 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7798 *) Fix for SSL server purpose checking. Server checking was
7799 rejecting certificates which had extended key usage present
7800 but no ssl client purpose.
7801 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7803 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7804 is a little unclear about how a blank password is handled.
7805 Since the password in encoded as a BMPString with terminating
7806 double NULL a zero length password would end up as just the
7807 double NULL. However no password at all is different and is
7808 handled differently in the PKCS#12 key generation code. NS
7809 treats a blank password as zero length. MSIE treats it as no
7810 password on export: but it will try both on import. We now do
7811 the same: PKCS12_parse() tries zero length and no password if
7812 the password is set to "" or NULL (NULL is now a valid password:
7813 it wasn't before) as does the pkcs12 application.
7816 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7817 perror when PEM_read_bio_X509_REQ fails, the error message must
7818 be obtained from the error queue.
7821 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7822 it in ERR_remove_state if appropriate, and change ERR_get_state
7823 accordingly to avoid race conditions (this is necessary because
7824 thread_hash is no longer constant once set).
7827 *) Bugfix for linux-elf makefile.one.
7830 *) RSA_get_default_method() will now cause a default
7831 RSA_METHOD to be chosen if one doesn't exist already.
7832 Previously this was only set during a call to RSA_new()
7833 or RSA_new_method(NULL) meaning it was possible for
7834 RSA_get_default_method() to return NULL.
7837 *) Added native name translation to the existing DSO code
7838 that will convert (if the flag to do so is set) filenames
7839 that are sufficiently small and have no path information
7840 into a canonical native form. Eg. "blah" converted to
7841 "libblah.so" or "blah.dll" etc.
7844 *) New function ERR_error_string_n(e, buf, len) which is like
7845 ERR_error_string(e, buf), but writes at most 'len' bytes
7846 including the 0 terminator. For ERR_error_string_n, 'buf'
7848 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7850 *) CONF library reworked to become more general. A new CONF
7851 configuration file reader "class" is implemented as well as a
7852 new functions (NCONF_*, for "New CONF") to handle it. The now
7853 old CONF_* functions are still there, but are reimplemented to
7854 work in terms of the new functions. Also, a set of functions
7855 to handle the internal storage of the configuration data is
7856 provided to make it easier to write new configuration file
7857 reader "classes" (I can definitely see something reading a
7858 configuration file in XML format, for example), called _CONF_*,
7859 or "the configuration storage API"...
7861 The new configuration file reading functions are:
7863 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7864 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7866 NCONF_default, NCONF_WIN32
7868 NCONF_dump_fp, NCONF_dump_bio
7870 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7871 NCONF_new creates a new CONF object. This works in the same way
7872 as other interfaces in OpenSSL, like the BIO interface.
7873 NCONF_dump_* dump the internal storage of the configuration file,
7874 which is useful for debugging. All other functions take the same
7875 arguments as the old CONF_* functions wth the exception of the
7876 first that must be a `CONF *' instead of a `LHASH *'.
7878 To make it easer to use the new classes with the old CONF_* functions,
7879 the function CONF_set_default_method is provided.
7882 *) Add '-tls1' option to 'openssl ciphers', which was already
7883 mentioned in the documentation but had not been implemented.
7884 (This option is not yet really useful because even the additional
7885 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7888 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7889 OpenSSL-based applications) load shared libraries and bind to
7890 them in a portable way.
7891 [Geoff Thorpe, with contributions from Richard Levitte]
7893 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7895 *) Make sure _lrotl and _lrotr are only used with MSVC.
7897 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7898 (the default implementation of RAND_status).
7900 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7901 to '-clrext' (= clear extensions), as intended and documented.
7902 [Bodo Moeller; inconsistency pointed out by Michael Attili
7905 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7906 was larger than the MD block size.
7907 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7909 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7910 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7911 using the passed key: if the passed key was a private key the result
7912 of X509_print(), for example, would be to print out all the private key
7916 *) des_quad_cksum() byte order bug fix.
7917 [Ulf Möller, using the problem description in krb4-0.9.7, where
7918 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7920 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7922 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7924 *) For easily testing in shell scripts whether some command
7925 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7926 returns with exit code 0 iff no command of the given name is available.
7927 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7928 the output goes to stdout and nothing is printed to stderr.
7929 Additional arguments are always ignored.
7931 Since for each cipher there is a command of the same name,
7932 the 'no-cipher' compilation switches can be tested this way.
7934 ('openssl no-XXX' is not able to detect pseudo-commands such
7935 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7938 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7941 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7942 is set; it will be thrown away anyway because each handshake creates
7944 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7945 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7946 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7947 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7950 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7951 'Q' and 'R' lose their special meanings (quit/renegotiate).
7952 This is part of what -quiet does; unlike -quiet, -ign_eof
7953 does not suppress any output.
7956 *) Add compatibility options to the purpose and trust code. The
7957 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7958 accepts a certificate or CA, this was the previous behaviour,
7959 with all the associated security issues.
7961 X509_TRUST_COMPAT is the old trust behaviour: only and
7962 automatically trust self signed roots in certificate store. A
7963 new trust setting X509_TRUST_DEFAULT is used to specify that
7964 a purpose has no associated trust setting and it should instead
7965 use the value in the default purpose.
7968 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7969 and fix a memory leak.
7972 *) In util/mkerr.pl (which implements 'make errors'), preserve
7973 reason strings from the previous version of the .c file, as
7974 the default to have only downcase letters (and digits) in
7975 automatically generated reasons codes is not always appropriate.
7978 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7979 using strerror. Previously, ERR_reason_error_string() returned
7980 library names as reason strings for SYSerr; but SYSerr is a special
7981 case where small numbers are errno values, not library numbers.
7984 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7985 converts DSA parameters into DH parameters. (When creating parameters,
7986 DSA_generate_parameters is used.)
7989 *) Include 'length' (recommended exponent length) in C code generated
7990 by 'openssl dhparam -C'.
7993 *) The second argument to set_label in perlasm was already being used
7994 so couldn't be used as a "file scope" flag. Moved to third argument
7998 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7999 instead of RAND_bytes for encryption IVs and salts.
8002 *) Include RAND_status() into RAND_METHOD instead of implementing
8003 it only for md_rand.c Otherwise replacing the PRNG by calling
8004 RAND_set_rand_method would be impossible.
8007 *) Don't let DSA_generate_key() enter an infinite loop if the random
8008 number generation fails.
8011 *) New 'rand' application for creating pseudo-random output.
8014 *) Added configuration support for Linux/IA64
8015 [Rolf Haberrecker <rolf@suse.de>]
8017 *) Assembler module support for Mingw32.
8020 *) Shared library support for HPUX (in shlib/).
8021 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8023 *) Shared library support for Solaris gcc.
8024 [Lutz Behnke <behnke@trustcenter.de>]
8026 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8028 *) PKCS7_encrypt() was adding text MIME headers twice because they
8029 were added manually and by SMIME_crlf_copy().
8032 *) In bntest.c don't call BN_rand with zero bits argument.
8033 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8035 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8036 case was implemented. This caused BN_div_recp() to fail occasionally.
8039 *) Add an optional second argument to the set_label() in the perl
8040 assembly language builder. If this argument exists and is set
8041 to 1 it signals that the assembler should use a symbol whose
8042 scope is the entire file, not just the current function. This
8043 is needed with MASM which uses the format label:: for this scope.
8044 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8046 *) Change the ASN1 types so they are typedefs by default. Before
8047 almost all types were #define'd to ASN1_STRING which was causing
8048 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8052 *) Change names of new functions to the new get1/get0 naming
8053 convention: After 'get1', the caller owns a reference count
8054 and has to call ..._free; 'get0' returns a pointer to some
8055 data structure without incrementing reference counters.
8056 (Some of the existing 'get' functions increment a reference
8057 counter, some don't.)
8058 Similarly, 'set1' and 'add1' functions increase reference
8059 counters or duplicate objects.
8062 *) Allow for the possibility of temp RSA key generation failure:
8063 the code used to assume it always worked and crashed on failure.
8066 *) Fix potential buffer overrun problem in BIO_printf().
8067 [Ulf Möller, using public domain code by Patrick Powell; problem
8068 pointed out by David Sacerdote <das33@cornell.edu>]
8070 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8071 RAND_egd() and RAND_status(). In the command line application,
8072 the EGD socket can be specified like a seed file using RANDFILE
8076 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8077 Some CAs (e.g. Verisign) distribute certificates in this form.
8080 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8081 list to exclude them. This means that no special compilation option
8082 is needed to use anonymous DH: it just needs to be included in the
8086 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8087 EVP_MD_type. The old functionality is available in a new macro called
8088 EVP_MD_md(). Change code that uses it and update docs.
8091 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8092 where the 'void *' argument is replaced by a function pointer argument.
8093 Previously 'void *' was abused to point to functions, which works on
8094 many platforms, but is not correct. As these functions are usually
8095 called by macros defined in OpenSSL header files, most source code
8096 should work without changes.
8099 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8100 sections with information on -D... compiler switches used for
8101 compiling the library so that applications can see them. To enable
8102 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8103 must be defined. E.g.,
8104 #define OPENSSL_ALGORITHM_DEFINES
8105 #include <openssl/opensslconf.h>
8106 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8107 [Richard Levitte, Ulf and Bodo Möller]
8109 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8113 *) Change the 'other' type in certificate aux info to a STACK_OF
8114 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8115 the required ASN1 format: arbitrary types determined by an OID.
8118 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8119 argument to 'req'. This is not because the function is newer or
8120 better than others it just uses the work 'NEW' in the certificate
8121 request header lines. Some software needs this.
8124 *) Reorganise password command line arguments: now passwords can be
8125 obtained from various sources. Delete the PEM_cb function and make
8126 it the default behaviour: i.e. if the callback is NULL and the
8127 usrdata argument is not NULL interpret it as a null terminated pass
8128 phrase. If usrdata and the callback are NULL then the pass phrase
8129 is prompted for as usual.
8132 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8133 the support is automatically enabled. The resulting binaries will
8134 autodetect the card and use it if present.
8135 [Ben Laurie and Compaq Inc.]
8137 *) Work around for Netscape hang bug. This sends certificate request
8138 and server done in one record. Since this is perfectly legal in the
8139 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8140 the bugs/SSLv3 entry for more info.
8143 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8146 *) Add -rand argument to smime and pkcs12 applications and read/write
8150 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8153 *) Add command line password options to the remaining applications.
8156 *) Bug fix for BN_div_recp() for numerators with an even number of
8160 *) More tests in bntest.c, and changed test_bn output.
8163 *) ./config recognizes MacOS X now.
8166 *) Bug fix for BN_div() when the first words of num and divsor are
8167 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8170 *) Add support for various broken PKCS#8 formats, and command line
8171 options to produce them.
8174 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8175 get temporary BIGNUMs from a BN_CTX.
8178 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8182 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8183 include a #define from the old name to the new. The original intent
8184 was that statically linked binaries could for example just call
8185 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8186 link with digests. This never worked becayse SSLeay_add_all_digests()
8187 and SSLeay_add_all_ciphers() were in the same source file so calling
8188 one would link with the other. They are now in separate source files.
8191 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8194 *) Use a less unusual form of the Miller-Rabin primality test (it used
8195 a binary algorithm for exponentiation integrated into the Miller-Rabin
8196 loop, our standard modexp algorithms are faster).
8199 *) Support for the EBCDIC character set completed.
8200 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8202 *) Source code cleanups: use const where appropriate, eliminate casts,
8203 use void * instead of char * in lhash.
8206 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8207 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8208 this the server could overwrite ephemeral keys that the client
8212 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8213 using 50 iterations of the Rabin-Miller test.
8215 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8216 iterations of the Rabin-Miller test as required by the appendix
8217 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8218 As BN_is_prime_fasttest includes trial division, DSA parameter
8219 generation becomes much faster.
8221 This implies a change for the callback functions in DSA_is_prime
8222 and DSA_generate_parameters: The callback function is called once
8223 for each positive witness in the Rabin-Miller test, not just
8224 occasionally in the inner loop; and the parameters to the
8225 callback function now provide an iteration count for the outer
8226 loop rather than for the current invocation of the inner loop.
8227 DSA_generate_parameters additionally can call the callback
8228 function with an 'iteration count' of -1, meaning that a
8229 candidate has passed the trial division test (when q is generated
8230 from an application-provided seed, trial division is skipped).
8233 *) New function BN_is_prime_fasttest that optionally does trial
8234 division before starting the Rabin-Miller test and has
8235 an additional BN_CTX * argument (whereas BN_is_prime always
8236 has to allocate at least one BN_CTX).
8237 'callback(1, -1, cb_arg)' is called when a number has passed the
8238 trial division stage.
8241 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8245 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8248 *) New function BN_pseudo_rand().
8251 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8252 bignum version of BN_from_montgomery() with the working code from
8253 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8257 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8258 made it impossible to use the same SSL_SESSION data structure in
8259 SSL2 clients in multiple threads.
8262 *) The return value of RAND_load_file() no longer counts bytes obtained
8263 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8264 to seed the PRNG (previously an explicit byte count was required).
8265 [Ulf Möller, Bodo Möller]
8267 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8268 used (char *) instead of (void *) and had casts all over the place.
8271 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8274 *) Retain source code compatibility for BN_prime_checks macro:
8275 BN_is_prime(..., BN_prime_checks, ...) now uses
8276 BN_prime_checks_for_size to determine the appropriate number of
8277 Rabin-Miller iterations.
8280 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8281 DH_CHECK_P_NOT_SAFE_PRIME.
8282 (Check if this is true? OpenPGP calls them "strong".)
8285 *) Merge the functionality of "dh" and "gendh" programs into a new program
8286 "dhparam". The old programs are retained for now but will handle DH keys
8287 (instead of parameters) in future.
8290 *) Make the ciphers, s_server and s_client programs check the return values
8291 when a new cipher list is set.
8294 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8295 ciphers. Before when the 56bit ciphers were enabled the sorting was
8298 The syntax for the cipher sorting has been extended to support sorting by
8299 cipher-strength (using the strength_bits hard coded in the tables).
8300 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8302 Fix a bug in the cipher-command parser: when supplying a cipher command
8303 string with an "undefined" symbol (neither command nor alphanumeric
8304 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8305 an error is flagged.
8307 Due to the strength-sorting extension, the code of the
8308 ssl_create_cipher_list() function was completely rearranged. I hope that
8309 the readability was also increased :-)
8310 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8312 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8313 for the first serial number and places 2 in the serial number file. This
8314 avoids problems when the root CA is created with serial number zero and
8315 the first user certificate has the same issuer name and serial number
8319 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8320 the new code. Add documentation for this stuff.
8323 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8324 X509_*() to X509at_*() on the grounds that they don't handle X509
8325 structures and behave in an analagous way to the X509v3 functions:
8326 they shouldn't be called directly but wrapper functions should be used
8329 So we also now have some wrapper functions that call the X509at functions
8330 when passed certificate requests. (TO DO: similar things can be done with
8331 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8332 things. Some of these need some d2i or i2d and print functionality
8333 because they handle more complex structures.)
8336 *) Add missing #ifndefs that caused missing symbols when building libssl
8337 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8338 NO_RSA in ssl/s2*.c.
8339 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8341 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8342 has a return value which indicates the quality of the random data
8343 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8344 error queue. New function RAND_pseudo_bytes() generates output that is
8345 guaranteed to be unique but not unpredictable. RAND_add is like
8346 RAND_seed, but takes an extra argument for an entropy estimate
8347 (RAND_seed always assumes full entropy).
8350 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8351 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8352 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8353 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8354 false-positive rate of at most 2^-80 for random input.
8357 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8360 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8361 in the 0.9.5 release), this returns the chain
8362 from an X509_CTX structure with a dup of the stack and all
8363 the X509 reference counts upped: so the stack will exist
8364 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8367 Also make SSL_SESSION_print() print out the verify return
8371 *) Add manpage for the pkcs12 command. Also change the default
8372 behaviour so MAC iteration counts are used unless the new
8373 -nomaciter option is used. This improves file security and
8374 only older versions of MSIE (4.0 for example) need it.
8377 *) Honor the no-xxx Configure options when creating .DEF files.
8380 *) Add PKCS#10 attributes to field table: challengePassword,
8381 unstructuredName and unstructuredAddress. These are taken from
8382 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8383 international characters are used.
8385 More changes to X509_ATTRIBUTE code: allow the setting of types
8386 based on strings. Remove the 'loc' parameter when adding
8387 attributes because these will be a SET OF encoding which is sorted
8391 *) Initial changes to the 'req' utility to allow request generation
8392 automation. This will allow an application to just generate a template
8393 file containing all the field values and have req construct the
8396 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8397 used all over the place including certificate requests and PKCS#7
8398 structures. They are currently handled manually where necessary with
8399 some primitive wrappers for PKCS#7. The new functions behave in a
8400 manner analogous to the X509 extension functions: they allow
8401 attributes to be looked up by NID and added.
8403 Later something similar to the X509V3 code would be desirable to
8404 automatically handle the encoding, decoding and printing of the
8405 more complex types. The string types like challengePassword can
8406 be handled by the string table functions.
8408 Also modified the multi byte string table handling. Now there is
8409 a 'global mask' which masks out certain types. The table itself
8410 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8411 is useful when for example there is only one permissible type
8412 (as in countryName) and using the mask might result in no valid
8416 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8417 SSL_get_peer_finished to allow applications to obtain the latest
8418 Finished messages sent to the peer or expected from the peer,
8419 respectively. (SSL_get_peer_finished is usually the Finished message
8420 actually received from the peer, otherwise the protocol will be aborted.)
8422 As the Finished message are message digests of the complete handshake
8423 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8424 be used for external authentication procedures when the authentication
8425 provided by SSL/TLS is not desired or is not enough.
8428 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8429 the host supports BWX extension and if Compaq C is present on the
8430 $PATH. Just exploiting of the BWX extension results in 20-30%
8431 performance kick for some algorithms, e.g. DES and RC4 to mention
8432 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8436 *) Add support for MS "fast SGC". This is arguably a violation of the
8437 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8438 weak crypto and after checking the certificate is SGC a second one
8439 with strong crypto. MS SGC stops the first handshake after receiving
8440 the server certificate message and sends a second client hello. Since
8441 a server will typically do all the time consuming operations before
8442 expecting any further messages from the client (server key exchange
8443 is the most expensive) there is little difference between the two.
8445 To get OpenSSL to support MS SGC we have to permit a second client
8446 hello message after we have sent server done. In addition we have to
8447 reset the MAC if we do get this second client hello.
8450 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8451 if a DER encoded private key is RSA or DSA traditional format. Changed
8452 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8453 format DER encoded private key. Newer code should use PKCS#8 format which
8454 has the key type encoded in the ASN1 structure. Added DER private key
8455 support to pkcs8 application.
8458 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8459 ciphersuites has been selected (as required by the SSL 3/TLS 1
8460 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8461 is set, we interpret this as a request to violate the specification
8462 (the worst that can happen is a handshake failure, and 'correct'
8463 behaviour would result in a handshake failure anyway).
8466 *) In SSL_CTX_add_session, take into account that there might be multiple
8467 SSL_SESSION structures with the same session ID (e.g. when two threads
8468 concurrently obtain them from an external cache).
8469 The internal cache can handle only one SSL_SESSION with a given ID,
8470 so if there's a conflict, we now throw out the old one to achieve
8474 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8475 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8476 some routines that use cipher OIDs: some ciphers do not have OIDs
8477 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8481 *) Simplify the trust setting structure and code. Now we just have
8482 two sequences of OIDs for trusted and rejected settings. These will
8483 typically have values the same as the extended key usage extension
8484 and any application specific purposes.
8486 The trust checking code now has a default behaviour: it will just
8487 check for an object with the same NID as the passed id. Functions can
8488 be provided to override either the default behaviour or the behaviour
8489 for a given id. SSL client, server and email already have functions
8490 in place for compatibility: they check the NID and also return "trusted"
8491 if the certificate is self signed.
8494 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8495 traditional format into an EVP_PKEY structure.
8498 *) Add a password callback function PEM_cb() which either prompts for
8499 a password if usr_data is NULL or otherwise assumes it is a null
8500 terminated password. Allow passwords to be passed on command line
8501 environment or config files in a few more utilities.
8504 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8505 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8506 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8507 Update documentation.
8510 *) Support for ASN1 "NULL" type. This could be handled before by using
8511 ASN1_TYPE but there wasn't any function that would try to read a NULL
8512 and produce an error if it couldn't. For compatibility we also have
8513 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8514 don't allocate anything because they don't need to.
8517 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8519 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8521 *) Rebuild of the memory allocation routines used by OpenSSL code and
8522 possibly others as well. The purpose is to make an interface that
8523 provide hooks so anyone can build a separate set of allocation and
8524 deallocation routines to be used by OpenSSL, for example memory
8525 pool implementations, or something else, which was previously hard
8526 since Malloc(), Realloc() and Free() were defined as macros having
8527 the values malloc, realloc and free, respectively (except for Win32
8528 compilations). The same is provided for memory debugging code.
8529 OpenSSL already comes with functionality to find memory leaks, but
8530 this gives people a chance to debug other memory problems.
8532 With these changes, a new set of functions and macros have appeared:
8534 CRYPTO_set_mem_debug_functions() [F]
8535 CRYPTO_get_mem_debug_functions() [F]
8536 CRYPTO_dbg_set_options() [F]
8537 CRYPTO_dbg_get_options() [F]
8538 CRYPTO_malloc_debug_init() [M]
8540 The memory debug functions are NULL by default, unless the library
8541 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8542 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8543 gives the standard debugging functions that come with OpenSSL) or
8544 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8545 provided by the library user) must be used. When the standard
8546 debugging functions are used, CRYPTO_dbg_set_options can be used to
8547 request additional information:
8548 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8549 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8551 Also, things like CRYPTO_set_mem_functions will always give the
8552 expected result (the new set of functions is used for allocation
8553 and deallocation) at all times, regardless of platform and compiler
8556 To finish it up, some functions that were never use in any other
8557 way than through macros have a new API and new semantic:
8560 CRYPTO_dbg_realloc()
8563 All macros of value have retained their old syntax.
8564 [Richard Levitte and Bodo Moeller]
8566 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8567 ordering of SMIMECapabilities wasn't in "strength order" and there
8568 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8572 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8573 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8574 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8576 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8577 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8578 functionality to handle multipart/signed properly) and a utility
8579 called 'smime' to call all this stuff. This is based on code I
8580 originally wrote for Celo who have kindly allowed it to be
8581 included in OpenSSL.
8584 *) Add variants des_set_key_checked and des_set_key_unchecked of
8585 des_set_key (aka des_key_sched). Global variable des_check_key
8586 decides which of these is called by des_set_key; this way
8587 des_check_key behaves as it always did, but applications and
8588 the library itself, which was buggy for des_check_key == 1,
8589 have a cleaner way to pick the version they need.
8592 *) New function PKCS12_newpass() which changes the password of a
8596 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8597 dynamic mix. In both cases the ids can be used as an index into the
8598 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8599 functions so they accept a list of the field values and the
8600 application doesn't need to directly manipulate the X509_TRUST
8604 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8608 *) Modify the way the V3 extension code looks up extensions. This now
8609 works in a similar way to the object code: we have some "standard"
8610 extensions in a static table which is searched with OBJ_bsearch()
8611 and the application can add dynamic ones if needed. The file
8612 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8613 updated whenever a new extension is added to the core code and kept
8614 in ext_nid order. There is a simple program 'tabtest.c' which checks
8615 this. New extensions are not added too often so this file can readily
8616 be maintained manually.
8618 There are two big advantages in doing things this way. The extensions
8619 can be looked up immediately and no longer need to be "added" using
8620 X509V3_add_standard_extensions(): this function now does nothing.
8621 [Side note: I get *lots* of email saying the extension code doesn't
8622 work because people forget to call this function]
8623 Also no dynamic allocation is done unless new extensions are added:
8624 so if we don't add custom extensions there is no need to call
8625 X509V3_EXT_cleanup().
8628 *) Modify enc utility's salting as follows: make salting the default. Add a
8629 magic header, so unsalted files fail gracefully instead of just decrypting
8630 to garbage. This is because not salting is a big security hole, so people
8631 should be discouraged from doing it.
8634 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8635 digest to be passed on the command line but it only used this
8636 parameter when signing a certificate. Modified so all relevant
8637 operations are affected by the digest parameter including the
8638 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8639 DSA key was used because it didn't fix the digest.
8642 *) Initial certificate chain verify code. Currently tests the untrusted
8643 certificates for consistency with the verify purpose (which is set
8644 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8646 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8647 this is because it will reject chains with invalid extensions whereas
8648 every previous version of OpenSSL and SSLeay made no checks at all.
8650 Trust code: checks the root CA for the relevant trust settings. Trust
8651 settings have an initial value consistent with the verify purpose: e.g.
8652 if the verify purpose is for SSL client use it expects the CA to be
8653 trusted for SSL client use. However the default value can be changed to
8654 permit custom trust settings: one example of this would be to only trust
8655 certificates from a specific "secure" set of CAs.
8657 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8658 which should be used for version portability: especially since the
8659 verify structure is likely to change more often now.
8661 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8662 to set them. If not set then assume SSL clients will verify SSL servers
8665 Two new options to the verify program: -untrusted allows a set of
8666 untrusted certificates to be passed in and -purpose which sets the
8667 intended purpose of the certificate. If a purpose is set then the
8668 new chain verify code is used to check extension consistency.
8671 *) Support for the authority information access extension.
8674 *) Modify RSA and DSA PEM read routines to transparently handle
8675 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8676 public keys in a format compatible with certificate
8677 SubjectPublicKeyInfo structures. Unfortunately there were already
8678 functions called *_PublicKey_* which used various odd formats so
8679 these are retained for compatibility: however the DSA variants were
8680 never in a public release so they have been deleted. Changed dsa/rsa
8681 utilities to handle the new format: note no releases ever handled public
8682 keys so we should be OK.
8684 The primary motivation for this change is to avoid the same fiasco
8685 that dogs private keys: there are several incompatible private key
8686 formats some of which are standard and some OpenSSL specific and
8687 require various evil hacks to allow partial transparent handling and
8688 even then it doesn't work with DER formats. Given the option anything
8689 other than PKCS#8 should be dumped: but the other formats have to
8690 stay in the name of compatibility.
8692 With public keys and the benefit of hindsight one standard format
8693 is used which works with EVP_PKEY, RSA or DSA structures: though
8694 it clearly returns an error if you try to read the wrong kind of key.
8696 Added a -pubkey option to the 'x509' utility to output the public key.
8697 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8698 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8699 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8700 that do the same as the EVP_PKEY_assign_*() except they up the
8701 reference count of the added key (they don't "swallow" the
8705 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8706 CRLs would fail if the file contained no certificates or no CRLs:
8707 added a new function to read in both types and return the number
8708 read: this means that if none are read it will be an error. The
8709 DER versions of the certificate and CRL reader would always fail
8710 because it isn't possible to mix certificates and CRLs in DER format
8711 without choking one or the other routine. Changed this to just read
8712 a certificate: this is the best we can do. Also modified the code
8713 in apps/verify.c to take notice of return codes: it was previously
8714 attempting to read in certificates from NULL pointers and ignoring
8715 any errors: this is one reason why the cert and CRL reader seemed
8716 to work. It doesn't check return codes from the default certificate
8717 routines: these may well fail if the certificates aren't installed.
8720 *) Code to support otherName option in GeneralName.
8723 *) First update to verify code. Change the verify utility
8724 so it warns if it is passed a self signed certificate:
8725 for consistency with the normal behaviour. X509_verify
8726 has been modified to it will now verify a self signed
8727 certificate if *exactly* the same certificate appears
8728 in the store: it was previously impossible to trust a
8729 single self signed certificate. This means that:
8730 openssl verify ss.pem
8731 now gives a warning about a self signed certificate but
8732 openssl verify -CAfile ss.pem ss.pem
8736 *) For servers, store verify_result in SSL_SESSION data structure
8737 (and add it to external session representation).
8738 This is needed when client certificate verifications fails,
8739 but an application-provided verification callback (set by
8740 SSL_CTX_set_cert_verify_callback) allows accepting the session
8741 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8742 but returns 1): When the session is reused, we have to set
8743 ssl->verify_result to the appropriate error code to avoid
8745 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8747 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8748 case in PKCS7_dataInit() where the signed PKCS7 structure
8749 didn't contain any existing data because it was being created.
8750 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8752 *) Add a salt to the key derivation routines in enc.c. This
8753 forms the first 8 bytes of the encrypted file. Also add a
8754 -S option to allow a salt to be input on the command line.
8757 *) New function X509_cmp(). Oddly enough there wasn't a function
8758 to compare two certificates. We do this by working out the SHA1
8759 hash and comparing that. X509_cmp() will be needed by the trust
8763 *) SSL_get1_session() is like SSL_get_session(), but increments
8764 the reference count in the SSL_SESSION returned.
8765 [Geoff Thorpe <geoff@eu.c2.net>]
8767 *) Fix for 'req': it was adding a null to request attributes.
8768 Also change the X509_LOOKUP and X509_INFO code to handle
8769 certificate auxiliary information.
8772 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8776 *) Add the possibility to add extra information to the memory leak
8777 detecting output, to form tracebacks, showing from where each
8778 allocation was originated: CRYPTO_push_info("constant string") adds
8779 the string plus current file name and line number to a per-thread
8780 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8781 is like calling CYRPTO_pop_info() until the stack is empty.
8782 Also updated memory leak detection code to be multi-thread-safe.
8785 *) Add options -text and -noout to pkcs7 utility and delete the
8786 encryption options which never did anything. Update docs.
8789 *) Add options to some of the utilities to allow the pass phrase
8790 to be included on either the command line (not recommended on
8791 OSes like Unix) or read from the environment. Update the
8792 manpages and fix a few bugs.
8795 *) Add a few manpages for some of the openssl commands.
8798 *) Fix the -revoke option in ca. It was freeing up memory twice,
8799 leaking and not finding already revoked certificates.
8802 *) Extensive changes to support certificate auxiliary information.
8803 This involves the use of X509_CERT_AUX structure and X509_AUX
8804 functions. An X509_AUX function such as PEM_read_X509_AUX()
8805 can still read in a certificate file in the usual way but it
8806 will also read in any additional "auxiliary information". By
8807 doing things this way a fair degree of compatibility can be
8808 retained: existing certificates can have this information added
8809 using the new 'x509' options.
8811 Current auxiliary information includes an "alias" and some trust
8812 settings. The trust settings will ultimately be used in enhanced
8813 certificate chain verification routines: currently a certificate
8814 can only be trusted if it is self signed and then it is trusted
8818 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8819 The problem was that one of the replacement routines had not been working
8820 since SSLeay releases. For now the offending routine has been replaced
8821 with non-optimised assembler. Even so, this now gives around 95%
8822 performance improvement for 1024 bit RSA signs.
8825 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8826 handling. Most clients have the effective key size in bits equal to
8827 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8828 A few however don't do this and instead use the size of the decrypted key
8829 to determine the RC2 key length and the AlgorithmIdentifier to determine
8830 the effective key length. In this case the effective key length can still
8831 be 40 bits but the key length can be 168 bits for example. This is fixed
8832 by manually forcing an RC2 key into the EVP_PKEY structure because the
8833 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8834 the key length and effective key length are equal.
8837 *) Add a bunch of functions that should simplify the creation of
8838 X509_NAME structures. Now you should be able to do:
8839 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8840 and have it automatically work out the correct field type and fill in
8841 the structures. The more adventurous can try:
8842 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8843 and it will (hopefully) work out the correct multibyte encoding.
8846 *) Change the 'req' utility to use the new field handling and multibyte
8847 copy routines. Before the DN field creation was handled in an ad hoc
8848 way in req, ca, and x509 which was rather broken and didn't support
8849 BMPStrings or UTF8Strings. Since some software doesn't implement
8850 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8851 using the dirstring_type option. See the new comment in the default
8852 openssl.cnf for more info.
8855 *) Make crypto/rand/md_rand.c more robust:
8856 - Assure unique random numbers after fork().
8857 - Make sure that concurrent threads access the global counter and
8858 md serializably so that we never lose entropy in them
8859 or use exactly the same state in multiple threads.
8860 Access to the large state is not always serializable because
8861 the additional locking could be a performance killer, and
8862 md should be large enough anyway.
8865 *) New file apps/app_rand.c with commonly needed functionality
8866 for handling the random seed file.
8868 Use the random seed file in some applications that previously did not:
8870 dsaparam -genkey (which also ignored its '-rand' option),
8873 x509 (when signing).
8874 Except on systems with /dev/urandom, it is crucial to have a random
8875 seed file at least for key creation, DSA signing, and for DH exchanges;
8876 for RSA signatures we could do without one.
8878 gendh and gendsa (unlike genrsa) used to read only the first byte
8879 of each file listed in the '-rand' option. The function as previously
8880 found in genrsa is now in app_rand.c and is used by all programs
8881 that support '-rand'.
8884 *) In RAND_write_file, use mode 0600 for creating files;
8885 don't just chmod when it may be too late.
8888 *) Report an error from X509_STORE_load_locations
8889 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8892 *) New function ASN1_mbstring_copy() this copies a string in either
8893 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8894 into an ASN1_STRING type. A mask of permissible types is passed
8895 and it chooses the "minimal" type to use or an error if not type
8899 *) Add function equivalents to the various macros in asn1.h. The old
8900 macros are retained with an M_ prefix. Code inside the library can
8901 use the M_ macros. External code (including the openssl utility)
8902 should *NOT* in order to be "shared library friendly".
8905 *) Add various functions that can check a certificate's extensions
8906 to see if it usable for various purposes such as SSL client,
8907 server or S/MIME and CAs of these types. This is currently
8908 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8909 verification. Also added a -purpose flag to x509 utility to
8910 print out all the purposes.
8913 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8917 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8918 for, obtain and decode and extension and obtain its critical flag.
8919 This allows all the necessary extension code to be handled in a
8920 single function call.
8923 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8924 platforms. See crypto/rc4/rc4_enc.c for further details.
8927 *) New -noout option to asn1parse. This causes no output to be produced
8928 its main use is when combined with -strparse and -out to extract data
8929 from a file (which may not be in ASN.1 format).
8932 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8933 when producing the local key id.
8934 [Richard Levitte <levitte@stacken.kth.se>]
8936 *) New option -dhparam in s_server. This allows a DH parameter file to be
8937 stated explicitly. If it is not stated then it tries the first server
8938 certificate file. The previous behaviour hard coded the filename
8942 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8943 a public key to be input or output. For example:
8944 openssl rsa -in key.pem -pubout -out pubkey.pem
8945 Also added necessary DSA public key functions to handle this.
8948 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8949 in the message. This was handled by allowing
8950 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8951 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8953 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8954 to the end of the strings whereas this didn't. This would cause problems
8955 if strings read with d2i_ASN1_bytes() were later modified.
8956 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8958 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8959 data and it contains EOF it will end up returning an error. This is
8960 caused by input 46 bytes long. The cause is due to the way base64
8961 BIOs find the start of base64 encoded data. They do this by trying a
8962 trial decode on each line until they find one that works. When they
8963 do a flag is set and it starts again knowing it can pass all the
8964 data directly through the decoder. Unfortunately it doesn't reset
8965 the context it uses. This means that if EOF is reached an attempt
8966 is made to pass two EOFs through the context and this causes the
8967 resulting error. This can also cause other problems as well. As is
8968 usual with these problems it takes *ages* to find and the fix is
8969 trivial: move one line.
8970 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8972 *) Ugly workaround to get s_client and s_server working under Windows. The
8973 old code wouldn't work because it needed to select() on sockets and the
8974 tty (for keypresses and to see if data could be written). Win32 only
8975 supports select() on sockets so we select() with a 1s timeout on the
8976 sockets and then see if any characters are waiting to be read, if none
8977 are present then we retry, we also assume we can always write data to
8978 the tty. This isn't nice because the code then blocks until we've
8979 received a complete line of data and it is effectively polling the
8980 keyboard at 1s intervals: however it's quite a bit better than not
8981 working at all :-) A dedicated Windows application might handle this
8982 with an event loop for example.
8985 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8986 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8987 will be called when RSA_sign() and RSA_verify() are used. This is useful
8988 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8989 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8990 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8991 This necessitated the support of an extra signature type NID_md5_sha1
8992 for SSL signatures and modifications to the SSL library to use it instead
8993 of calling RSA_public_decrypt() and RSA_private_encrypt().
8996 *) Add new -verify -CAfile and -CApath options to the crl program, these
8997 will lookup a CRL issuers certificate and verify the signature in a
8998 similar way to the verify program. Tidy up the crl program so it
8999 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9000 less strict. It will now permit CRL extensions even if it is not
9001 a V2 CRL: this will allow it to tolerate some broken CRLs.
9004 *) Initialize all non-automatic variables each time one of the openssl
9005 sub-programs is started (this is necessary as they may be started
9006 multiple times from the "OpenSSL>" prompt).
9007 [Lennart Bang, Bodo Moeller]
9009 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9010 removing all other RSA functionality (this is what NO_RSA does). This
9011 is so (for example) those in the US can disable those operations covered
9012 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9016 *) Non-copying interface to BIO pairs.
9017 (still largely untested)
9020 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9021 ASCII string. This was handled independently in various places before.
9024 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9025 UTF8 strings a character at a time.
9028 *) Use client_version from client hello to select the protocol
9029 (s23_srvr.c) and for RSA client key exchange verification
9030 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9033 *) Add various utility functions to handle SPKACs, these were previously
9034 handled by poking round in the structure internals. Added new function
9035 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9036 print, verify and generate SPKACs. Based on an original idea from
9037 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9040 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9043 *) Allow the config file extension section to be overwritten on the
9044 command line. Based on an original idea from Massimiliano Pala
9045 <madwolf@comune.modena.it>. The new option is called -extensions
9046 and can be applied to ca, req and x509. Also -reqexts to override
9047 the request extensions in req and -crlexts to override the crl extensions
9051 *) Add new feature to the SPKAC handling in ca. Now you can include
9052 the same field multiple times by preceding it by "XXXX." for example:
9055 this is the same syntax as used in the req config file.
9058 *) Allow certificate extensions to be added to certificate requests. These
9059 are specified in a 'req_extensions' option of the req section of the
9060 config file. They can be printed out with the -text option to req but
9061 are otherwise ignored at present.
9064 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9065 data read consists of only the final block it would not decrypted because
9066 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9067 A misplaced 'break' also meant the decrypted final block might not be
9068 copied until the next read.
9071 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9072 a few extra parameters to the DH structure: these will be useful if
9073 for example we want the value of 'q' or implement X9.42 DH.
9076 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9077 provides hooks that allow the default DSA functions or functions on a
9078 "per key" basis to be replaced. This allows hardware acceleration and
9079 hardware key storage to be handled without major modification to the
9080 library. Also added low level modexp hooks and CRYPTO_EX structure and
9081 associated functions.
9084 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9085 as "read only": it can't be written to and the buffer it points to will
9086 not be freed. Reading from a read only BIO is much more efficient than
9087 a normal memory BIO. This was added because there are several times when
9088 an area of memory needs to be read from a BIO. The previous method was
9089 to create a memory BIO and write the data to it, this results in two
9090 copies of the data and an O(n^2) reading algorithm. There is a new
9091 function BIO_new_mem_buf() which creates a read only memory BIO from
9092 an area of memory. Also modified the PKCS#7 routines to use read only
9096 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9097 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9098 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9099 but a retry condition occured while trying to read the rest.
9102 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9103 NID_pkcs7_encrypted by default: this was wrong since this should almost
9104 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9105 the encrypted data type: this is a more sensible place to put it and it
9106 allows the PKCS#12 code to be tidied up that duplicated this
9110 *) Changed obj_dat.pl script so it takes its input and output files on
9111 the command line. This should avoid shell escape redirection problems
9115 *) Initial support for certificate extension requests, these are included
9116 in things like Xenroll certificate requests. Included functions to allow
9117 extensions to be obtained and added.
9120 *) -crlf option to s_client and s_server for sending newlines as
9121 CRLF (as required by many protocols).
9124 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9126 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9127 [Ralf S. Engelschall]
9129 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9130 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9132 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9136 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9137 DH parameters/keys (q is lost during that conversion, but the resulting
9138 DH parameters contain its length).
9140 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9141 much faster than DH_generate_parameters (which creates parameters
9142 where p = 2*q + 1), and also the smaller q makes DH computations
9143 much more efficient (160-bit exponentiation instead of 1024-bit
9144 exponentiation); so this provides a convenient way to support DHE
9145 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9146 utter importance to use
9147 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9149 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9150 when such DH parameters are used, because otherwise small subgroup
9151 attacks may become possible!
9154 *) Avoid memory leak in i2d_DHparams.
9157 *) Allow the -k option to be used more than once in the enc program:
9158 this allows the same encrypted message to be read by multiple recipients.
9161 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9162 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9163 it will always use the numerical form of the OID, even if it has a short
9167 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9168 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9169 otherwise bn_mod_exp was called. In the case of hardware keys for example
9170 no private key components need be present and it might store extra data
9171 in the RSA structure, which cannot be accessed from bn_mod_exp.
9172 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9173 private key operations.
9176 *) Added support for SPARC Linux.
9179 *) pem_password_cb function type incompatibly changed from
9180 typedef int pem_password_cb(char *buf, int size, int rwflag);
9182 ....(char *buf, int size, int rwflag, void *userdata);
9183 so that applications can pass data to their callbacks:
9184 The PEM[_ASN1]_{read,write}... functions and macros now take an
9185 additional void * argument, which is just handed through whenever
9186 the password callback is called.
9187 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9189 New function SSL_CTX_set_default_passwd_cb_userdata.
9191 Compatibility note: As many C implementations push function arguments
9192 onto the stack in reverse order, the new library version is likely to
9193 interoperate with programs that have been compiled with the old
9194 pem_password_cb definition (PEM_whatever takes some data that
9195 happens to be on the stack as its last argument, and the callback
9196 just ignores this garbage); but there is no guarantee whatsoever that
9199 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9200 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9201 problems not only on Windows, but also on some Unix platforms.
9202 To avoid problematic command lines, these definitions are now in an
9203 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9204 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9207 *) MIPS III/IV assembler module is reimplemented.
9210 *) More DES library cleanups: remove references to srand/rand and
9211 delete an unused file.
9214 *) Add support for the the free Netwide assembler (NASM) under Win32,
9215 since not many people have MASM (ml) and it can be hard to obtain.
9216 This is currently experimental but it seems to work OK and pass all
9217 the tests. Check out INSTALL.W32 for info.
9220 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9221 without temporary keys kept an extra copy of the server key,
9222 and connections with temporary keys did not free everything in case
9226 *) New function RSA_check_key and new openssl rsa option -check
9227 for verifying the consistency of RSA keys.
9228 [Ulf Moeller, Bodo Moeller]
9230 *) Various changes to make Win32 compile work:
9231 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9232 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9233 comparison" warnings.
9234 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9237 *) Add a debugging option to PKCS#5 v2 key generation function: when
9238 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9239 derived keys are printed to stderr.
9242 *) Copy the flags in ASN1_STRING_dup().
9243 [Roman E. Pavlov <pre@mo.msk.ru>]
9245 *) The x509 application mishandled signing requests containing DSA
9246 keys when the signing key was also DSA and the parameters didn't match.
9248 It was supposed to omit the parameters when they matched the signing key:
9249 the verifying software was then supposed to automatically use the CA's
9250 parameters if they were absent from the end user certificate.
9252 Omitting parameters is no longer recommended. The test was also
9253 the wrong way round! This was probably due to unusual behaviour in
9254 EVP_cmp_parameters() which returns 1 if the parameters match.
9255 This meant that parameters were omitted when they *didn't* match and
9256 the certificate was useless. Certificates signed with 'ca' didn't have
9258 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9260 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9261 The interface is as follows:
9262 Applications can use
9263 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9264 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9265 "off" is now the default.
9266 The library internally uses
9267 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9268 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9269 to disable memory-checking temporarily.
9271 Some inconsistent states that previously were possible (and were
9272 even the default) are now avoided.
9274 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9275 with each memory chunk allocated; this is occasionally more helpful
9276 than just having a counter.
9278 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9280 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9284 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9285 which largely parallels "options", but is for changing API behaviour,
9286 whereas "options" are about protocol behaviour.
9287 Initial "mode" flags are:
9289 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9290 a single record has been written.
9291 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9292 retries use the same buffer location.
9293 (But all of the contents must be
9297 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9300 *) Fix problems with no-hmac etc.
9301 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9303 *) New functions RSA_get_default_method(), RSA_set_method() and
9304 RSA_get_method(). These allows replacement of RSA_METHODs without having
9305 to mess around with the internals of an RSA structure.
9308 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9309 Also really enable memory leak checks in openssl.c and in some
9311 [Chad C. Mulligan, Bodo Moeller]
9313 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9314 up the length of negative integers. This has now been simplified to just
9315 store the length when it is first determined and use it later, rather
9316 than trying to keep track of where data is copied and updating it to
9318 [Steve Henson, reported by Brien Wheeler
9319 <bwheeler@authentica-security.com>]
9321 *) Add a new function PKCS7_signatureVerify. This allows the verification
9322 of a PKCS#7 signature but with the signing certificate passed to the
9323 function itself. This contrasts with PKCS7_dataVerify which assumes the
9324 certificate is present in the PKCS#7 structure. This isn't always the
9325 case: certificates can be omitted from a PKCS#7 structure and be
9326 distributed by "out of band" means (such as a certificate database).
9329 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9330 function prototypes in pem.h, also change util/mkdef.pl to add the
9331 necessary function names.
9334 *) mk1mf.pl (used by Windows builds) did not properly read the
9335 options set by Configure in the top level Makefile, and Configure
9336 was not even able to write more than one option correctly.
9337 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9340 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9341 file to be loaded from a BIO or FILE pointer. The BIO version will
9342 for example allow memory BIOs to contain config info.
9345 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9346 Whoever hopes to achieve shared-library compatibility across versions
9347 must use this, not the compile-time macro.
9348 (Exercise 0.9.4: Which is the minimum library version required by
9350 Note: All this applies only to multi-threaded programs, others don't
9354 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9355 through a BIO pair triggered the default case, i.e.
9356 SSLerr(...,SSL_R_UNKNOWN_STATE).
9359 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9360 can use the SSL library even if none of the specific BIOs is
9364 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9365 for the encoded length.
9366 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9368 *) Add initial documentation of the X509V3 functions.
9371 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9372 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9373 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9374 secure PKCS#8 private key format with a high iteration count.
9377 *) Fix determination of Perl interpreter: A perl or perl5
9378 _directory_ in $PATH was also accepted as the interpreter.
9379 [Ralf S. Engelschall]
9381 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9382 wrong with it but it was very old and did things like calling
9383 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9387 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9388 to use the new extension code.
9391 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9392 with macros. This should make it easier to change their form, add extra
9393 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9397 *) Add to configuration table a new entry that can specify an alternative
9398 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9399 according to Mark Crispin <MRC@Panda.COM>.
9403 *) DES CBC did not update the IV. Weird.
9406 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9407 Changing the behaviour of the former might break existing programs --
9408 where IV updating is needed, des_ncbc_encrypt can be used.
9411 *) When bntest is run from "make test" it drives bc to check its
9412 calculations, as well as internally checking them. If an internal check
9413 fails, it needs to cause bc to give a non-zero result or make test carries
9414 on without noticing the failure. Fixed.
9417 *) DES library cleanups.
9420 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9421 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9422 ciphers. NOTE: although the key derivation function has been verified
9423 against some published test vectors it has not been extensively tested
9424 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9428 *) Instead of "mkdir -p", which is not fully portable, use new
9429 Perl script "util/mkdir-p.pl".
9432 *) Rewrite the way password based encryption (PBE) is handled. It used to
9433 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9434 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9435 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9436 the 'parameter' field of the AlgorithmIdentifier is passed to the
9437 underlying key generation function so it must do its own ASN1 parsing.
9438 This has also changed the EVP_PBE_CipherInit() function which now has a
9439 'parameter' argument instead of literal salt and iteration count values
9440 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9443 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9444 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9445 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9446 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9447 value was just used as a "magic string" and not used directly its
9448 value doesn't matter.
9451 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9455 *) "linux-sparc64" configuration (ultrapenguin).
9456 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9457 "linux-sparc" configuration.
9458 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9460 *) config now generates no-xxx options for missing ciphers.
9463 *) Support the EBCDIC character set (work in progress).
9464 File ebcdic.c not yet included because it has a different license.
9465 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9467 *) Support BS2000/OSD-POSIX.
9468 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9470 *) Make callbacks for key generation use void * instead of char *.
9473 *) Make S/MIME samples compile (not yet tested).
9476 *) Additional typesafe stacks.
9479 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9483 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9485 *) New configuration variant "sco5-gcc".
9487 *) Updated some demos.
9488 [Sean O Riordain, Wade Scholine]
9490 *) Add missing BIO_free at exit of pkcs12 application.
9493 *) Fix memory leak in conf.c.
9496 *) Updates for Win32 to assembler version of MD5.
9499 *) Set #! path to perl in apps/der_chop to where we found it
9500 instead of using a fixed path.
9503 *) SHA library changes for irix64-mips4-cc.
9506 *) Improvements for VMS support.
9510 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9512 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9513 This also avoids the problems with SC4.2 and unpatched SC5.
9514 [Andy Polyakov <appro@fy.chalmers.se>]
9516 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9517 These are required because of the typesafe stack would otherwise break
9518 existing code. If old code used a structure member which used to be STACK
9519 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9520 sk_num or sk_value it would produce an error because the num, data members
9521 are not present in STACK_OF. Now it just produces a warning. sk_set
9522 replaces the old method of assigning a value to sk_value
9523 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9524 that does this will no longer work (and should use sk_set instead) but
9525 this could be regarded as a "questionable" behaviour anyway.
9528 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9529 correctly handle encrypted S/MIME data.
9532 *) Change type of various DES function arguments from des_cblock
9533 (which means, in function argument declarations, pointer to char)
9534 to des_cblock * (meaning pointer to array with 8 char elements),
9535 which allows the compiler to do more typechecking; it was like
9536 that back in SSLeay, but with lots of ugly casts.
9538 Introduce new type const_des_cblock.
9541 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9542 problems: find RecipientInfo structure that matches recipient certificate
9543 and initialise the ASN1 structures properly based on passed cipher.
9546 *) Belatedly make the BN tests actually check the results.
9549 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9550 to and from BNs: it was completely broken. New compilation option
9551 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9552 key elements as negative integers.
9555 *) Reorganize and speed up MD5.
9556 [Andy Polyakov <appro@fy.chalmers.se>]
9559 [Richard Levitte <richard@levitte.org>]
9561 *) New option -out to asn1parse to allow the parsed structure to be
9562 output to a file. This is most useful when combined with the -strparse
9563 option to examine the output of things like OCTET STRINGS.
9566 *) Make SSL library a little more fool-proof by not requiring any longer
9567 that SSL_set_{accept,connect}_state be called before
9568 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9569 in many applications because usually everything *appeared* to work as
9570 intended anyway -- now it really works as intended).
9573 *) Move openssl.cnf out of lib/.
9576 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9577 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9578 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9579 [Ralf S. Engelschall]
9581 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9582 handle PKCS#7 enveloped data properly.
9583 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9585 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9586 copying pointers. The cert_st handling is changed by this in
9587 various ways (and thus what used to be known as ctx->default_cert
9588 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9589 any longer when s->cert does not give us what we need).
9590 ssl_cert_instantiate becomes obsolete by this change.
9591 As soon as we've got the new code right (possibly it already is?),
9592 we have solved a couple of bugs of the earlier code where s->cert
9593 was used as if it could not have been shared with other SSL structures.
9595 Note that using the SSL API in certain dirty ways now will result
9596 in different behaviour than observed with earlier library versions:
9597 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9598 does not influence s as it used to.
9600 In order to clean up things more thoroughly, inside SSL_SESSION
9601 we don't use CERT any longer, but a new structure SESS_CERT
9602 that holds per-session data (if available); currently, this is
9603 the peer's certificate chain and, for clients, the server's certificate
9604 and temporary key. CERT holds only those values that can have
9605 meaningful defaults in an SSL_CTX.
9608 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9609 from the internal representation. Various PKCS#7 fixes: remove some
9610 evil casts and set the enc_dig_alg field properly based on the signing
9614 *) Allow PKCS#12 password to be set from the command line or the
9615 environment. Let 'ca' get its config file name from the environment
9616 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9620 *) Allow certificate policies extension to use an IA5STRING for the
9621 organization field. This is contrary to the PKIX definition but
9622 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9626 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9627 without disallowing inline assembler and the like for non-pedantic builds.
9630 *) Support Borland C++ builder.
9631 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9636 *) SHA-1 cleanups and performance enhancements.
9637 [Andy Polyakov <appro@fy.chalmers.se>]
9639 *) Sparc v8plus assembler for the bignum library.
9640 [Andy Polyakov <appro@fy.chalmers.se>]
9642 *) Accept any -xxx and +xxx compiler options in Configure.
9645 *) Update HPUX configuration.
9648 *) Add missing sk_<type>_unshift() function to safestack.h
9649 [Ralf S. Engelschall]
9651 *) New function SSL_CTX_use_certificate_chain_file that sets the
9652 "extra_cert"s in addition to the certificate. (This makes sense
9653 only for "PEM" format files, as chains as a whole are not
9657 *) Support verify_depth from the SSL API.
9658 x509_vfy.c had what can be considered an off-by-one-error:
9659 Its depth (which was not part of the external interface)
9660 was actually counting the number of certificates in a chain;
9661 now it really counts the depth.
9664 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9665 instead of X509err, which often resulted in confusing error
9666 messages since the error codes are not globally unique
9667 (e.g. an alleged error in ssl3_accept when a certificate
9668 didn't match the private key).
9670 *) New function SSL_CTX_set_session_id_context that allows to set a default
9671 value (so that you don't need SSL_set_session_id_context for each
9672 connection using the SSL_CTX).
9675 *) OAEP decoding bug fix.
9678 *) Support INSTALL_PREFIX for package builders, as proposed by
9682 *) New Configure options "threads" and "no-threads". For systems
9683 where the proper compiler options are known (currently Solaris
9684 and Linux), "threads" is the default.
9687 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9690 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9691 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9692 such as /usr/local/bin.
9695 *) "make linux-shared" to build shared libraries.
9696 [Niels Poppe <niels@netbox.org>]
9698 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9701 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9702 extension adding in x509 utility.
9705 *) Remove NOPROTO sections and error code comments.
9708 *) Partial rewrite of the DEF file generator to now parse the ANSI
9712 *) New Configure options --prefix=DIR and --openssldir=DIR.
9715 *) Complete rewrite of the error code script(s). It is all now handled
9716 by one script at the top level which handles error code gathering,
9717 header rewriting and C source file generation. It should be much better
9718 than the old method: it now uses a modified version of Ulf's parser to
9719 read the ANSI prototypes in all header files (thus the old K&R definitions
9720 aren't needed for error creation any more) and do a better job of
9721 translating function codes into names. The old 'ASN1 error code imbedded
9722 in a comment' is no longer necessary and it doesn't use .err files which
9723 have now been deleted. Also the error code call doesn't have to appear all
9724 on one line (which resulted in some large lines...).
9727 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9730 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9731 0 (which usually indicates a closed connection), but continue reading.
9734 *) Fix some race conditions.
9737 *) Add support for CRL distribution points extension. Add Certificate
9738 Policies and CRL distribution points documentation.
9741 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9744 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9745 8 of keying material. Merlin has also confirmed interop with this fix
9746 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9747 [Merlin Hughes <merlin@baltimore.ie>]
9749 *) Fix lots of warnings.
9750 [Richard Levitte <levitte@stacken.kth.se>]
9752 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9753 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9754 [Richard Levitte <levitte@stacken.kth.se>]
9756 *) Fix problems with sizeof(long) == 8.
9757 [Andy Polyakov <appro@fy.chalmers.se>]
9759 *) Change functions to ANSI C.
9762 *) Fix typos in error codes.
9763 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9765 *) Remove defunct assembler files from Configure.
9768 *) SPARC v8 assembler BIGNUM implementation.
9769 [Andy Polyakov <appro@fy.chalmers.se>]
9771 *) Support for Certificate Policies extension: both print and set.
9772 Various additions to support the r2i method this uses.
9775 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9776 return a const string when you are expecting an allocated buffer.
9779 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9780 types DirectoryString and DisplayText.
9783 *) Add code to allow r2i extensions to access the configuration database,
9784 add an LHASH database driver and add several ctx helper functions.
9787 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9788 fail when they extended the size of a BIGNUM.
9791 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9792 support typesafe stack.
9795 *) Fix typo in SSL_[gs]et_options().
9796 [Nils Frostberg <nils@medcom.se>]
9798 *) Delete various functions and files that belonged to the (now obsolete)
9799 old X509V3 handling code.
9802 *) New Configure option "rsaref".
9805 *) Don't auto-generate pem.h.
9808 *) Introduce type-safe ASN.1 SETs.
9811 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9812 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9814 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9815 that links with OpenSSL (well at least cause lots of warnings), but fear
9816 not: the conversion is trivial, and it eliminates loads of evil casts. A
9817 few STACKed things have been converted already. Feel free to convert more.
9818 In the fullness of time, I'll do away with the STACK type altogether.
9821 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9822 specified in <certfile> by updating the entry in the index.txt file.
9823 This way one no longer has to edit the index.txt file manually for
9824 revoking a certificate. The -revoke option does the gory details now.
9825 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9827 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9828 `-text' option at all and this way the `-noout -text' combination was
9829 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9830 [Ralf S. Engelschall]
9832 *) Make sure a corresponding plain text error message exists for the
9833 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9834 verify callback function determined that a certificate was revoked.
9835 [Ralf S. Engelschall]
9837 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9838 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9839 all available cipers including rc5, which was forgotten until now.
9840 In order to let the testing shell script know which algorithms
9841 are available, a new (up to now undocumented) command
9842 "openssl list-cipher-commands" is used.
9845 *) Bugfix: s_client occasionally would sleep in select() when
9846 it should have checked SSL_pending() first.
9849 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9850 the raw DSA values prior to ASN.1 encoding.
9853 *) Tweaks to Configure
9854 [Niels Poppe <niels@netbox.org>]
9856 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9860 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9863 *) New config option to avoid instructions that are illegal on the 80386.
9864 The default code is faster, but requires at least a 486.
9867 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9868 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9869 same as SSL2_VERSION anyway.
9872 *) New "-showcerts" option for s_client.
9875 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9876 application. Various cleanups and fixes.
9879 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9880 modify error routines to work internally. Add error codes and PBE init
9881 to library startup routines.
9884 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9885 packing functions to asn1 and evp. Changed function names and error
9886 codes along the way.
9889 *) PKCS12 integration: and so it begins... First of several patches to
9890 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9891 objects to objects.h
9894 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9895 and display support for Thawte strong extranet extension.
9898 *) Add LinuxPPC support.
9899 [Jeff Dubrule <igor@pobox.org>]
9901 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9902 bn_div_words in alpha.s.
9903 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9905 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9906 OAEP isn't supported when OpenSSL is built with RSAref.
9907 [Ulf Moeller <ulf@fitug.de>]
9909 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9910 so they no longer are missing under -DNOPROTO.
9911 [Soren S. Jorvang <soren@t.dk>]
9914 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9916 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9917 doesn't work when the session is reused. Coming soon!
9920 *) Fix a security hole, that allows sessions to be reused in the wrong
9921 context thus bypassing client cert protection! All software that uses
9922 client certs and session caches in multiple contexts NEEDS PATCHING to
9923 allow session reuse! A fuller solution is in the works.
9924 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9926 *) Some more source tree cleanups (removed obsolete files
9927 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9928 permission on "config" script to be executable) and a fix for the INSTALL
9930 [Ulf Moeller <ulf@fitug.de>]
9932 *) Remove some legacy and erroneous uses of malloc, free instead of
9934 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9936 *) Make rsa_oaep_test return non-zero on error.
9937 [Ulf Moeller <ulf@fitug.de>]
9939 *) Add support for native Solaris shared libraries. Configure
9940 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9941 if someone would make that last step automatic.
9942 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9944 *) ctx_size was not built with the right compiler during "make links". Fixed.
9947 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9948 except NULL ciphers". This means the default cipher list will no longer
9949 enable NULL ciphers. They need to be specifically enabled e.g. with
9950 the string "DEFAULT:eNULL".
9953 *) Fix to RSA private encryption routines: if p < q then it would
9954 occasionally produce an invalid result. This will only happen with
9955 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9958 *) Be less restrictive and allow also `perl util/perlpath.pl
9959 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9960 because this way one can also use an interpreter named `perl5' (which is
9961 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9962 installed as `perl').
9963 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9965 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9966 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9968 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9969 advapi32.lib to Win32 build and change the pem test comparision
9970 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9971 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9972 and crypto/des/ede_cbcm_enc.c.
9975 *) DES quad checksum was broken on big-endian architectures. Fixed.
9978 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9979 Win32 test batch file so it (might) work again. The Win32 test batch file
9980 is horrible: I feel ill....
9983 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9984 in e_os.h. Audit of header files to check ANSI and non ANSI
9985 sections: 10 functions were absent from non ANSI section and not exported
9986 from Windows DLLs. Fixed up libeay.num for new functions.
9989 *) Make `openssl version' output lines consistent.
9990 [Ralf S. Engelschall]
9992 *) Fix Win32 symbol export lists for BIO functions: Added
9993 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9994 to ms/libeay{16,32}.def.
9995 [Ralf S. Engelschall]
9997 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9998 fine under Unix and passes some trivial tests I've now added. But the
9999 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10000 added to make sure no one expects that this stuff really works in the
10001 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10002 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10004 [Ralf S. Engelschall]
10006 *) Fix the generation of two part addresses in perl.
10007 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10009 *) Add config entry for Linux on MIPS.
10010 [John Tobey <jtobey@channel1.com>]
10012 *) Make links whenever Configure is run, unless we are on Windoze.
10015 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10016 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10020 *) Add a useful kludge to allow package maintainers to specify compiler and
10021 other platforms details on the command line without having to patch the
10022 Configure script everytime: One now can use ``perl Configure
10023 <id>:<details>'', i.e. platform ids are allowed to have details appended
10024 to them (seperated by colons). This is treated as there would be a static
10025 pre-configured entry in Configure's %table under key <id> with value
10026 <details> and ``perl Configure <id>'' is called. So, when you want to
10027 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10028 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10029 now, which overrides the FreeBSD-elf entry on-the-fly.
10030 [Ralf S. Engelschall]
10032 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10035 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10036 on the `perl Configure ...' command line. This way one can compile
10037 OpenSSL libraries with Position Independent Code (PIC) which is needed
10038 for linking it into DSOs.
10039 [Ralf S. Engelschall]
10041 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10045 *) Cleaned up the LICENSE document: The official contact for any license
10046 questions now is the OpenSSL core team under openssl-core@openssl.org.
10047 And add a paragraph about the dual-license situation to make sure people
10048 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10049 to the OpenSSL toolkit.
10050 [Ralf S. Engelschall]
10052 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10053 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10054 Additonally cleaned up the `make links' target: Remove unnecessary
10055 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10056 to speed processing and no longer clutter the display with confusing
10057 stuff. Instead only the actually done links are displayed.
10058 [Ralf S. Engelschall]
10060 *) Permit null encryption ciphersuites, used for authentication only. It used
10061 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10062 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10066 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10067 signed attributes when verifying signatures (this would break them),
10068 the detached data encoding was wrong and public keys obtained using
10069 X509_get_pubkey() weren't freed.
10072 *) Add text documentation for the BUFFER functions. Also added a work around
10073 to a Win95 console bug. This was triggered by the password read stuff: the
10074 last character typed gets carried over to the next fread(). If you were
10075 generating a new cert request using 'req' for example then the last
10076 character of the passphrase would be CR which would then enter the first
10080 *) Added the new `Includes OpenSSL Cryptography Software' button as
10081 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10082 button and can be used by applications based on OpenSSL to show the
10083 relationship to the OpenSSL project.
10084 [Ralf S. Engelschall]
10086 *) Remove confusing variables in function signatures in files
10087 ssl/ssl_lib.c and ssl/ssl.h.
10088 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10090 *) Don't install bss_file.c under PREFIX/include/
10091 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10093 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10094 functions that return function pointers and has support for NT specific
10095 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10096 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10097 unsigned to signed types: this was killing the Win32 compile.
10100 *) Add new certificate file to stack functions,
10101 SSL_add_dir_cert_subjects_to_stack() and
10102 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10103 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10104 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10105 This means that Apache-SSL and similar packages don't have to mess around
10106 to add as many CAs as they want to the preferred list.
10109 *) Experiment with doxygen documentation. Currently only partially applied to
10111 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10112 openssl.doxy as the configuration file.
10115 *) Get rid of remaining C++-style comments which strict C compilers hate.
10116 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10118 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10119 compiled in by default: it has problems with large keys.
10122 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10123 DH private keys and/or callback functions which directly correspond to
10124 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10125 is needed for applications which have to configure certificates on a
10126 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10128 For the RSA certificate situation is makes no difference, but
10129 for the DSA certificate situation this fixes the "no shared cipher"
10130 problem where the OpenSSL cipher selection procedure failed because the
10131 temporary keys were not overtaken from the context and the API provided
10132 no way to reconfigure them.
10133 The new functions now let applications reconfigure the stuff and they
10134 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10135 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10136 non-public-API function ssl_cert_instantiate() is used as a helper
10137 function and also to reduce code redundancy inside ssl_rsa.c.
10138 [Ralf S. Engelschall]
10140 *) Move s_server -dcert and -dkey options out of the undocumented feature
10141 area because they are useful for the DSA situation and should be
10142 recognized by the users.
10143 [Ralf S. Engelschall]
10145 *) Fix the cipher decision scheme for export ciphers: the export bits are
10146 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10147 SSL_EXP_MASK. So, the original variable has to be used instead of the
10148 already masked variable.
10149 [Richard Levitte <levitte@stacken.kth.se>]
10151 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10152 [Richard Levitte <levitte@stacken.kth.se>]
10154 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10155 from `int' to `unsigned int' because it's a length and initialized by
10156 EVP_DigestFinal() which expects an `unsigned int *'.
10157 [Richard Levitte <levitte@stacken.kth.se>]
10159 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10160 script. Instead use the usual Shell->Perl transition trick.
10161 [Ralf S. Engelschall]
10163 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10164 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10165 -noout -modulus' as it's already the case for `openssl rsa -noout
10166 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10167 currently the public key is printed (a decision which was already done by
10168 `openssl dsa -modulus' in the past) which serves a similar purpose.
10169 Additionally the NO_RSA no longer completely removes the whole -modulus
10170 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10172 [Ralf S. Engelschall]
10174 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10175 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10176 [Arne Ansper <arne@ats.cyber.ee>]
10178 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10179 to be added. Now both 'req' and 'ca' can use new objects defined in the
10183 *) Add cool BIO that does syslog (or event log on NT).
10184 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10186 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10187 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10188 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10189 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10192 *) Add preliminary config info for new extension code.
10195 *) Make RSA_NO_PADDING really use no padding.
10196 [Ulf Moeller <ulf@fitug.de>]
10198 *) Generate errors when private/public key check is done.
10201 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10202 for some CRL extensions and new objects added.
10205 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10206 key usage extension and fuller support for authority key id.
10209 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10210 padding method for RSA, which is recommended for new applications in PKCS
10211 #1 v2.0 (RFC 2437, October 1998).
10212 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10213 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10214 against Bleichbacher's attack on RSA.
10215 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10218 *) Updates to the new SSL compression code
10219 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10221 *) Fix so that the version number in the master secret, when passed
10222 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10223 (because the server will not accept higher), that the version number
10224 is 0x03,0x01, not 0x03,0x00
10225 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10227 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10228 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10229 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10232 *) Support for RAW extensions where an arbitrary extension can be
10233 created by including its DER encoding. See apps/openssl.cnf for
10237 *) Make sure latest Perl versions don't interpret some generated C array
10238 code as Perl array code in the crypto/err/err_genc.pl script.
10239 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10241 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10242 not many people have the assembler. Various Win32 compilation fixes and
10243 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10244 build instructions.
10247 *) Modify configure script 'Configure' to automatically create crypto/date.h
10248 file under Win32 and also build pem.h from pem.org. New script
10249 util/mkfiles.pl to create the MINFO file on environments that can't do a
10250 'make files': perl util/mkfiles.pl >MINFO should work.
10253 *) Major rework of DES function declarations, in the pursuit of correctness
10254 and purity. As a result, many evil casts evaporated, and some weirdness,
10255 too. You may find this causes warnings in your code. Zapping your evil
10256 casts will probably fix them. Mostly.
10259 *) Fix for a typo in asn1.h. Bug fix to object creation script
10260 obj_dat.pl. It considered a zero in an object definition to mean
10261 "end of object": none of the objects in objects.h have any zeros
10262 so it wasn't spotted.
10263 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10265 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10266 Masking (CBCM). In the absence of test vectors, the best I have been able
10267 to do is check that the decrypt undoes the encrypt, so far. Send me test
10268 vectors if you have them.
10271 *) Correct calculation of key length for export ciphers (too much space was
10272 allocated for null ciphers). This has not been tested!
10275 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10276 message is now correct (it understands "crypto" and "ssl" on its
10277 command line). There is also now an "update" option. This will update
10278 the util/ssleay.num and util/libeay.num files with any new functions.
10280 perl util/mkdef.pl crypto ssl update
10281 it will update them.
10284 *) Overhauled the Perl interface (perl/*):
10285 - ported BN stuff to OpenSSL's different BN library
10286 - made the perl/ source tree CVS-aware
10287 - renamed the package from SSLeay to OpenSSL (the files still contain
10288 their history because I've copied them in the repository)
10289 - removed obsolete files (the test scripts will be replaced
10290 by better Test::Harness variants in the future)
10291 [Ralf S. Engelschall]
10293 *) First cut for a very conservative source tree cleanup:
10294 1. merge various obsolete readme texts into doc/ssleay.txt
10295 where we collect the old documents and readme texts.
10296 2. remove the first part of files where I'm already sure that we no
10297 longer need them because of three reasons: either they are just temporary
10298 files which were left by Eric or they are preserved original files where
10299 I've verified that the diff is also available in the CVS via "cvs diff
10300 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10301 the crypto/md/ stuff).
10302 [Ralf S. Engelschall]
10304 *) More extension code. Incomplete support for subject and issuer alt
10305 name, issuer and authority key id. Change the i2v function parameters
10306 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10307 what that's for :-) Fix to ASN1 macro which messed up
10308 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10311 *) Preliminary support for ENUMERATED type. This is largely copied from the
10315 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10316 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10318 *) Make sure `make rehash' target really finds the `openssl' program.
10319 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10321 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10322 like to hear about it if this slows down other processors.
10325 *) Add CygWin32 platform information to Configure script.
10326 [Alan Batie <batie@aahz.jf.intel.com>]
10328 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10329 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10331 *) New program nseq to manipulate netscape certificate sequences
10334 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10338 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10339 but the BN code had some problems that would cause failures when
10340 doing certificate verification and some other functions.
10341 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10343 *) Add ASN1 and PEM code to support netscape certificate sequences.
10346 *) Add ASN1 and PEM code to support netscape certificate sequences.
10349 *) Add several PKIX and private extended key usage OIDs.
10352 *) Modify the 'ca' program to handle the new extension code. Modify
10353 openssl.cnf for new extension format, add comments.
10356 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10357 and add a sample to openssl.cnf so req -x509 now adds appropriate
10361 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10362 error code, add initial support to X509_print() and x509 application.
10365 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10366 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10367 stuff is currently isolated and isn't even compiled yet.
10370 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10371 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10372 Removed the versions check from X509 routines when loading extensions:
10373 this allows certain broken certificates that don't set the version
10374 properly to be processed.
10377 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10378 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10379 can still be regenerated with "make depend".
10382 *) Spelling mistake in C version of CAST-128.
10383 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10385 *) Changes to the error generation code. The perl script err-code.pl
10386 now reads in the old error codes and retains the old numbers, only
10387 adding new ones if necessary. It also only changes the .err files if new
10388 codes are added. The makefiles have been modified to only insert errors
10389 when needed (to avoid needlessly modifying header files). This is done
10390 by only inserting errors if the .err file is newer than the auto generated
10391 C file. To rebuild all the error codes from scratch (the old behaviour)
10392 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10393 or delete all the .err files.
10396 *) CAST-128 was incorrectly implemented for short keys. The C version has
10397 been fixed, but is untested. The assembler versions are also fixed, but
10398 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10399 to regenerate it if needed.
10400 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10401 Hagino <itojun@kame.net>]
10403 *) File was opened incorrectly in randfile.c.
10404 [Ulf Möller <ulf@fitug.de>]
10406 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10407 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10408 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10409 al: it's just almost always a UTCTime. Note this patch adds new error
10410 codes so do a "make errors" if there are problems.
10413 *) Correct Linux 1 recognition in config.
10414 [Ulf Möller <ulf@fitug.de>]
10416 *) Remove pointless MD5 hash when using DSA keys in ca.
10417 [Anonymous <nobody@replay.com>]
10419 *) Generate an error if given an empty string as a cert directory. Also
10420 generate an error if handed NULL (previously returned 0 to indicate an
10421 error, but didn't set one).
10422 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10424 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10427 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10428 parameters. This was causing a warning which killed off the Win32 compile.
10431 *) Remove C++ style comments from crypto/bn/bn_local.h.
10432 [Neil Costigan <neil.costigan@celocom.com>]
10434 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10435 based on a text string, looking up short and long names and finally
10436 "dot" format. The "dot" format stuff didn't work. Added new function
10437 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10438 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10439 OID is not part of the table.
10442 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10443 X509_LOOKUP_by_alias().
10446 *) Sort openssl functions by name.
10449 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10450 encryption from sample DSA keys (in case anyone is interested the password
10454 *) Make _all_ *_free functions accept a NULL pointer.
10455 [Frans Heymans <fheymans@isaserver.be>]
10457 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10459 [Anonymous <nobody@replay.com>]
10461 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10462 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10464 *) Don't blow it for numeric -newkey arguments to apps/req.
10465 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10467 *) Temp key "for export" tests were wrong in s3_srvr.c.
10468 [Anonymous <nobody@replay.com>]
10470 *) Add prototype for temp key callback functions
10471 SSL_CTX_set_tmp_{rsa,dh}_callback().
10474 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10475 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10478 *) X509_name_add_entry() freed the wrong thing after an error.
10479 [Arne Ansper <arne@ats.cyber.ee>]
10481 *) rsa_eay.c would attempt to free a NULL context.
10482 [Arne Ansper <arne@ats.cyber.ee>]
10484 *) BIO_s_socket() had a broken should_retry() on Windoze.
10485 [Arne Ansper <arne@ats.cyber.ee>]
10487 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10488 [Arne Ansper <arne@ats.cyber.ee>]
10490 *) Make sure the already existing X509_STORE->depth variable is initialized
10491 in X509_STORE_new(), but document the fact that this variable is still
10492 unused in the certificate verification process.
10493 [Ralf S. Engelschall]
10495 *) Fix the various library and apps files to free up pkeys obtained from
10496 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10499 *) Fix reference counting in X509_PUBKEY_get(). This makes
10500 demos/maurice/example2.c work, amongst others, probably.
10501 [Steve Henson and Ben Laurie]
10503 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10504 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10505 are no longer created. This way we have a single and consistent command
10506 line interface `openssl <command>', similar to `cvs <command>'.
10507 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10509 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10510 BIT STRING wrapper always have zero unused bits.
10513 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10516 *) Make the top-level INSTALL documentation easier to understand.
10519 *) Makefiles updated to exit if an error occurs in a sub-directory
10520 make (including if user presses ^C) [Paul Sutton]
10522 *) Make Montgomery context stuff explicit in RSA data structure.
10525 *) Fix build order of pem and err to allow for generated pem.h.
10528 *) Fix renumbering bug in X509_NAME_delete_entry().
10531 *) Enhanced the err-ins.pl script so it makes the error library number
10532 global and can add a library name. This is needed for external ASN1 and
10533 other error libraries.
10536 *) Fixed sk_insert which never worked properly.
10539 *) Fix ASN1 macros so they can handle indefinite length construted
10540 EXPLICIT tags. Some non standard certificates use these: they can now
10544 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10545 into a single doc/ssleay.txt bundle. This way the information is still
10546 preserved but no longer messes up this directory. Now it's new room for
10547 the new set of documenation files.
10548 [Ralf S. Engelschall]
10550 *) SETs were incorrectly DER encoded. This was a major pain, because they
10551 shared code with SEQUENCEs, which aren't coded the same. This means that
10552 almost everything to do with SETs or SEQUENCEs has either changed name or
10553 number of arguments.
10554 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10556 *) Fix test data to work with the above.
10559 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10560 was already fixed by Eric for 0.9.1 it seems.
10561 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10563 *) Autodetect FreeBSD3.
10566 *) Fix various bugs in Configure. This affects the following platforms:
10570 unixware-2.0-pentium
10574 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10575 before they are needed.
10578 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10582 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10584 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10585 changed SSLeay to OpenSSL in version strings.
10586 [Ralf S. Engelschall]
10588 *) Some fixups to the top-level documents.
10591 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10592 because the symlink to include/ was missing.
10593 [Ralf S. Engelschall]
10595 *) Incorporated the popular no-RSA/DSA-only patches
10596 which allow to compile a RSA-free SSLeay.
10597 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10599 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10600 when "ssleay" is still not found.
10601 [Ralf S. Engelschall]
10603 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10604 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10606 *) Updated the README file.
10607 [Ralf S. Engelschall]
10609 *) Added various .cvsignore files in the CVS repository subdirs
10610 to make a "cvs update" really silent.
10611 [Ralf S. Engelschall]
10613 *) Recompiled the error-definition header files and added
10614 missing symbols to the Win32 linker tables.
10615 [Ralf S. Engelschall]
10617 *) Cleaned up the top-level documents;
10618 o new files: CHANGES and LICENSE
10619 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10620 o merged COPYRIGHT into LICENSE
10621 o removed obsolete TODO file
10622 o renamed MICROSOFT to INSTALL.W32
10623 [Ralf S. Engelschall]
10625 *) Removed dummy files from the 0.9.1b source tree:
10626 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10627 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10628 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10629 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10630 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10631 [Ralf S. Engelschall]
10633 *) Added various platform portability fixes.
10636 *) The Genesis of the OpenSSL rpject:
10637 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10638 Young and Tim J. Hudson created while they were working for C2Net until
10640 [The OpenSSL Project]
10643 Changes between 0.9.0b and 0.9.1b [not released]
10645 *) Updated a few CA certificates under certs/
10648 *) Changed some BIGNUM api stuff.
10651 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10652 DGUX x86, Linux Alpha, etc.
10655 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10656 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10660 *) Add -strparse option to asn1pars program which parses nested
10662 [Dr Stephen Henson <shenson@bigfoot.com>]
10664 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10667 *) DSA fix for "ca" program.
10670 *) Added "-genkey" option to "dsaparam" program.
10673 *) Added RIPE MD160 (rmd160) message digest.
10676 *) Added -a (all) option to "ssleay version" command.
10679 *) Added PLATFORM define which is the id given to Configure.
10682 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10685 *) Extended the ASN.1 parser routines.
10688 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10691 *) Added a BN_CTX to the BN library.
10694 *) Fixed the weak key values in DES library
10697 *) Changed API in EVP library for cipher aliases.
10700 *) Added support for RC2/64bit cipher.
10703 *) Converted the lhash library to the crypto/mem.c functions.
10706 *) Added more recognized ASN.1 object ids.
10709 *) Added more RSA padding checks for SSL/TLS.
10712 *) Added BIO proxy/filter functionality.
10715 *) Added extra_certs to SSL_CTX which can be used
10716 send extra CA certificates to the client in the CA cert chain sending
10717 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10720 *) Now Fortezza is denied in the authentication phase because
10721 this is key exchange mechanism is not supported by SSLeay at all.
10724 *) Additional PKCS1 checks.
10727 *) Support the string "TLSv1" for all TLS v1 ciphers.
10730 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10731 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10734 *) Fixed a few memory leaks.
10737 *) Fixed various code and comment typos.
10740 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10741 bytes sent in the client random.
10742 [Edward Bishop <ebishop@spyglass.com>]
projects / thirdparty / openssl.git / blob