5 This is a high-level summary of the most important changes.
6 For a full list of changes, see the git commit log; for example,
7 https://github.com/openssl/openssl/commits/ and pick the appropriate
10 Changes between 1.0.2m and 1.0.2n [xx XXX xxxx]
14 Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
16 *) bn_sqrx8x_internal carry bug on x86_64
18 There is a carry propagating bug in the x86_64 Montgomery squaring
19 procedure. No EC algorithms are affected. Analysis suggests that attacks
20 against RSA and DSA as a result of this defect would be very difficult to
21 perform and are not believed likely. Attacks against DH are considered just
22 feasible (although very difficult) because most of the work necessary to
23 deduce information about a private key may be performed offline. The amount
24 of resources required for such an attack would be very significant and
25 likely only accessible to a limited number of attackers. An attacker would
26 additionally need online access to an unpatched system using the target
27 private key in a scenario with persistent DH parameters and a private
28 key that is shared between multiple clients.
30 This only affects processors that support the BMI1, BMI2 and ADX extensions
31 like Intel Broadwell (5th generation) and later or AMD Ryzen.
33 This issue was reported to OpenSSL by the OSS-Fuzz project.
37 *) Malformed X.509 IPAddressFamily could cause OOB read
39 If an X.509 certificate has a malformed IPAddressFamily extension,
40 OpenSSL could do a one-byte buffer overread. The most likely result
41 would be an erroneous display of the certificate in text format.
43 This issue was reported to OpenSSL by the OSS-Fuzz project.
47 Changes between 1.0.2k and 1.0.2l [25 May 2017]
49 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
50 platform rather than 'mingw'.
53 Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
55 *) Truncated packet could crash via OOB read
57 If one side of an SSL/TLS path is running on a 32-bit host and a specific
58 cipher is being used, then a truncated packet can cause that host to
59 perform an out-of-bounds read, usually resulting in a crash.
61 This issue was reported to OpenSSL by Robert Święcki of Google.
65 *) BN_mod_exp may produce incorrect results on x86_64
67 There is a carry propagating bug in the x86_64 Montgomery squaring
68 procedure. No EC algorithms are affected. Analysis suggests that attacks
69 against RSA and DSA as a result of this defect would be very difficult to
70 perform and are not believed likely. Attacks against DH are considered just
71 feasible (although very difficult) because most of the work necessary to
72 deduce information about a private key may be performed offline. The amount
73 of resources required for such an attack would be very significant and
74 likely only accessible to a limited number of attackers. An attacker would
75 additionally need online access to an unpatched system using the target
76 private key in a scenario with persistent DH parameters and a private
77 key that is shared between multiple clients. For example this can occur by
78 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
79 similar to CVE-2015-3193 but must be treated as a separate problem.
81 This issue was reported to OpenSSL by the OSS-Fuzz project.
85 *) Montgomery multiplication may produce incorrect results
87 There is a carry propagating bug in the Broadwell-specific Montgomery
88 multiplication procedure that handles input lengths divisible by, but
89 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
90 and DH private keys are impossible. This is because the subroutine in
91 question is not used in operations with the private key itself and an input
92 of the attacker's direct choice. Otherwise the bug can manifest itself as
93 transient authentication and key negotiation failures or reproducible
94 erroneous outcome of public-key operations with specially crafted input.
95 Among EC algorithms only Brainpool P-512 curves are affected and one
96 presumably can attack ECDH key negotiation. Impact was not analyzed in
97 detail, because pre-requisites for attack are considered unlikely. Namely
98 multiple clients have to choose the curve in question and the server has to
99 share the private key among them, neither of which is default behaviour.
100 Even then only clients that chose the curve will be affected.
102 This issue was publicly reported as transient failures and was not
103 initially recognized as a security issue. Thanks to Richard Morgan for
104 providing reproducible case.
108 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
109 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
110 prevent issues where no progress is being made and the peer continually
111 sends unrecognised record types, using up resources processing them.
114 Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
116 *) Missing CRL sanity check
118 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
119 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
120 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
122 This issue only affects the OpenSSL 1.0.2i
126 Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
128 *) OCSP Status Request extension unbounded memory growth
130 A malicious client can send an excessively large OCSP Status Request
131 extension. If that client continually requests renegotiation, sending a
132 large OCSP Status Request extension each time, then there will be unbounded
133 memory growth on the server. This will eventually lead to a Denial Of
134 Service attack through memory exhaustion. Servers with a default
135 configuration are vulnerable even if they do not support OCSP. Builds using
136 the "no-ocsp" build time option are not affected.
138 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
142 *) In order to mitigate the SWEET32 attack, the DES ciphers were moved from
145 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
150 *) OOB write in MDC2_Update()
152 An overflow can occur in MDC2_Update() either if called directly or
153 through the EVP_DigestUpdate() function using MDC2. If an attacker
154 is able to supply very large amounts of input data after a previous
155 call to EVP_EncryptUpdate() with a partial block then a length check
156 can overflow resulting in a heap corruption.
158 The amount of data needed is comparable to SIZE_MAX which is impractical
161 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
165 *) Malformed SHA512 ticket DoS
167 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
168 DoS attack where a malformed ticket will result in an OOB read which will
171 The use of SHA512 in TLS session tickets is comparatively rare as it requires
172 a custom server callback and ticket lookup mechanism.
174 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
178 *) OOB write in BN_bn2dec()
180 The function BN_bn2dec() does not check the return value of BN_div_word().
181 This can cause an OOB write if an application uses this function with an
182 overly large BIGNUM. This could be a problem if an overly large certificate
183 or CRL is printed out from an untrusted source. TLS is not affected because
184 record limits will reject an oversized certificate before it is parsed.
186 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
190 *) OOB read in TS_OBJ_print_bio()
192 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
193 the total length the OID text representation would use and not the amount
194 of data written. This will result in OOB reads when large OIDs are
197 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
201 *) Pointer arithmetic undefined behaviour
203 Avoid some undefined pointer arithmetic
205 A common idiom in the codebase is to check limits in the following manner:
208 Where "p" points to some malloc'd data of SIZE bytes and
211 "len" here could be from some externally supplied data (e.g. from a TLS
214 The rules of C pointer arithmetic are such that "p + len" is only well
215 defined where len <= SIZE. Therefore the above idiom is actually
218 For example this could cause problems if some malloc implementation
219 provides an address for "p" such that "p + len" actually overflows for
220 values of len that are too big and therefore p + len < limit.
222 This issue was reported to OpenSSL by Guido Vranken
226 *) Constant time flag not preserved in DSA signing
228 Operations in the DSA signing algorithm should run in constant time in
229 order to avoid side channel attacks. A flaw in the OpenSSL DSA
230 implementation means that a non-constant time codepath is followed for
231 certain operations. This has been demonstrated through a cache-timing
232 attack to be sufficient for an attacker to recover the private DSA key.
234 This issue was reported by César Pereida (Aalto University), Billy Brumley
235 (Tampere University of Technology), and Yuval Yarom (The University of
240 *) DTLS buffered message DoS
242 In a DTLS connection where handshake messages are delivered out-of-order
243 those messages that OpenSSL is not yet ready to process will be buffered
244 for later use. Under certain circumstances, a flaw in the logic means that
245 those messages do not get removed from the buffer even though the handshake
246 has been completed. An attacker could force up to approx. 15 messages to
247 remain in the buffer when they are no longer required. These messages will
248 be cleared when the DTLS connection is closed. The default maximum size for
249 a message is 100k. Therefore the attacker could force an additional 1500k
250 to be consumed per connection. By opening many simulataneous connections an
251 attacker could cause a DoS attack through memory exhaustion.
253 This issue was reported to OpenSSL by Quan Luo.
257 *) DTLS replay protection DoS
259 A flaw in the DTLS replay attack protection mechanism means that records
260 that arrive for future epochs update the replay protection "window" before
261 the MAC for the record has been validated. This could be exploited by an
262 attacker by sending a record for the next epoch (which does not have to
263 decrypt or have a valid MAC), with a very large sequence number. This means
264 that all subsequent legitimate packets are dropped causing a denial of
265 service for a specific DTLS connection.
267 This issue was reported to OpenSSL by the OCAP audit team.
271 *) Certificate message OOB reads
273 In OpenSSL 1.0.2 and earlier some missing message length checks can result
274 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
275 theoretical DoS risk but this has not been observed in practice on common
278 The messages affected are client certificate, client certificate request
279 and server certificate. As a result the attack can only be performed
280 against a client or a server which enables client authentication.
282 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
286 Changes between 1.0.2g and 1.0.2h [3 May 2016]
288 *) Prevent padding oracle in AES-NI CBC MAC check
290 A MITM attacker can use a padding oracle attack to decrypt traffic
291 when the connection uses an AES CBC cipher and the server support
294 This issue was introduced as part of the fix for Lucky 13 padding
295 attack (CVE-2013-0169). The padding check was rewritten to be in
296 constant time by making sure that always the same bytes are read and
297 compared against either the MAC or padding bytes. But it no longer
298 checked that there was enough data to have both the MAC and padding
301 This issue was reported by Juraj Somorovsky using TLS-Attacker.
305 *) Fix EVP_EncodeUpdate overflow
307 An overflow can occur in the EVP_EncodeUpdate() function which is used for
308 Base64 encoding of binary data. If an attacker is able to supply very large
309 amounts of input data then a length check can overflow resulting in a heap
312 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
313 the PEM_write_bio* family of functions. These are mainly used within the
314 OpenSSL command line applications, so any application which processes data
315 from an untrusted source and outputs it as a PEM file should be considered
316 vulnerable to this issue. User applications that call these APIs directly
317 with large amounts of untrusted data may also be vulnerable.
319 This issue was reported by Guido Vranken.
323 *) Fix EVP_EncryptUpdate overflow
325 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
326 is able to supply very large amounts of input data after a previous call to
327 EVP_EncryptUpdate() with a partial block then a length check can overflow
328 resulting in a heap corruption. Following an analysis of all OpenSSL
329 internal usage of the EVP_EncryptUpdate() function all usage is one of two
330 forms. The first form is where the EVP_EncryptUpdate() call is known to be
331 the first called function after an EVP_EncryptInit(), and therefore that
332 specific call must be safe. The second form is where the length passed to
333 EVP_EncryptUpdate() can be seen from the code to be some small value and
334 therefore there is no possibility of an overflow. Since all instances are
335 one of these two forms, it is believed that there can be no overflows in
336 internal code due to this problem. It should be noted that
337 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
338 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
339 of these calls have also been analysed too and it is believed there are no
340 instances in internal usage where an overflow could occur.
342 This issue was reported by Guido Vranken.
346 *) Prevent ASN.1 BIO excessive memory allocation
348 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
349 a short invalid encoding can casuse allocation of large amounts of memory
350 potentially consuming excessive resources or exhausting memory.
352 Any application parsing untrusted data through d2i BIO functions is
353 affected. The memory based functions such as d2i_X509() are *not* affected.
354 Since the memory based functions are used by the TLS library, TLS
355 applications are not affected.
357 This issue was reported by Brian Carpenter.
363 ASN1 Strings that are over 1024 bytes can cause an overread in applications
364 using the X509_NAME_oneline() function on EBCDIC systems. This could result
365 in arbitrary stack data being returned in the buffer.
367 This issue was reported by Guido Vranken.
371 *) Modify behavior of ALPN to invoke callback after SNI/servername
372 callback, such that updates to the SSL_CTX affect ALPN.
375 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
379 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
380 methods are enabled and ssl2 is disabled the methods return NULL.
383 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
385 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
386 Builds that are not configured with "enable-weak-ssl-ciphers" will not
387 provide any "EXPORT" or "LOW" strength ciphers.
390 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
391 is by default disabled at build-time. Builds that are not configured with
392 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
393 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
394 will need to explicitly call either of:
396 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
398 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
400 as appropriate. Even if either of those is used, or the application
401 explicitly uses the version-specific SSLv2_method() or its client and
402 server variants, SSLv2 ciphers vulnerable to exhaustive search key
403 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
404 ciphers, and SSLv2 56-bit DES are no longer available.
408 *) Fix a double-free in DSA code
410 A double free bug was discovered when OpenSSL parses malformed DSA private
411 keys and could lead to a DoS attack or memory corruption for applications
412 that receive DSA private keys from untrusted sources. This scenario is
415 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
420 *) Disable SRP fake user seed to address a server memory leak.
422 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
424 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
425 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
426 was changed to ignore the "fake user" SRP seed, even if the seed
429 Users should use SRP_VBASE_get1_by_user instead. Note that in
430 SRP_VBASE_get1_by_user, caller must free the returned value. Note
431 also that even though configuring the SRP seed attempts to hide
432 invalid usernames by continuing the handshake with fake
433 credentials, this behaviour is not constant time and no strong
434 guarantees are made that the handshake is indistinguishable from
435 that of a valid user.
439 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
441 In the BN_hex2bn function the number of hex digits is calculated using an
442 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
443 large values of |i| this can result in |bn_expand| not allocating any
444 memory because |i * 4| is negative. This can leave the internal BIGNUM data
445 field as NULL leading to a subsequent NULL ptr deref. For very large values
446 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
447 In this case memory is allocated to the internal BIGNUM data field, but it
448 is insufficiently sized leading to heap corruption. A similar issue exists
449 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
450 is ever called by user applications with very large untrusted hex/dec data.
451 This is anticipated to be a rare occurrence.
453 All OpenSSL internal usage of these functions use data that is not expected
454 to be untrusted, e.g. config file data or application command line
455 arguments. If user developed applications generate config file data based
456 on untrusted data then it is possible that this could also lead to security
457 consequences. This is also anticipated to be rare.
459 This issue was reported to OpenSSL by Guido Vranken.
463 *) Fix memory issues in BIO_*printf functions
465 The internal |fmtstr| function used in processing a "%s" format string in
466 the BIO_*printf functions could overflow while calculating the length of a
467 string and cause an OOB read when printing very long strings.
469 Additionally the internal |doapr_outch| function can attempt to write to an
470 OOB memory location (at an offset from the NULL pointer) in the event of a
471 memory allocation failure. In 1.0.2 and below this could be caused where
472 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
473 could be in processing a very long "%s" format string. Memory leaks can
476 The first issue may mask the second issue dependent on compiler behaviour.
477 These problems could enable attacks where large amounts of untrusted data
478 is passed to the BIO_*printf functions. If applications use these functions
479 in this way then they could be vulnerable. OpenSSL itself uses these
480 functions when printing out human-readable dumps of ASN.1 data. Therefore
481 applications that print this data could be vulnerable if the data is from
482 untrusted sources. OpenSSL command line applications could also be
483 vulnerable where they print out ASN.1 data, or if untrusted data is passed
484 as command line arguments.
486 Libssl is not considered directly vulnerable. Additionally certificates etc
487 received via remote connections via libssl are also unlikely to be able to
488 trigger these issues because of message size limits enforced within libssl.
490 This issue was reported to OpenSSL Guido Vranken.
494 *) Side channel attack on modular exponentiation
496 A side-channel attack was found which makes use of cache-bank conflicts on
497 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
498 of RSA keys. The ability to exploit this issue is limited as it relies on
499 an attacker who has control of code in a thread running on the same
500 hyper-threaded core as the victim thread which is performing decryptions.
502 This issue was reported to OpenSSL by Yuval Yarom, The University of
503 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
504 Nadia Heninger, University of Pennsylvania with more information at
505 http://cachebleed.info.
509 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
510 if no keysize is specified with default_bits. This fixes an
511 omission in an earlier change that changed all RSA/DSA key generation
512 apps to use 2048 bits by default.
515 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
517 *) DH small subgroups
519 Historically OpenSSL only ever generated DH parameters based on "safe"
520 primes. More recently (in version 1.0.2) support was provided for
521 generating X9.42 style parameter files such as those required for RFC 5114
522 support. The primes used in such files may not be "safe". Where an
523 application is using DH configured with parameters based on primes that are
524 not "safe" then an attacker could use this fact to find a peer's private
525 DH exponent. This attack requires that the attacker complete multiple
526 handshakes in which the peer uses the same private DH exponent. For example
527 this could be used to discover a TLS server's private DH exponent if it's
528 reusing the private DH exponent or it's using a static DH ciphersuite.
530 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
531 TLS. It is not on by default. If the option is not set then the server
532 reuses the same private DH exponent for the life of the server process and
533 would be vulnerable to this attack. It is believed that many popular
534 applications do set this option and would therefore not be at risk.
536 The fix for this issue adds an additional check where a "q" parameter is
537 available (as is the case in X9.42 based parameters). This detects the
538 only known attack, and is the only possible defense for static DH
539 ciphersuites. This could have some performance impact.
541 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
542 default and cannot be disabled. This could have some performance impact.
544 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
548 *) SSLv2 doesn't block disabled ciphers
550 A malicious client can negotiate SSLv2 ciphers that have been disabled on
551 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
552 been disabled, provided that the SSLv2 protocol was not also disabled via
555 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
556 and Sebastian Schinzel.
560 *) Reject DH handshakes with parameters shorter than 1024 bits.
563 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
565 *) BN_mod_exp may produce incorrect results on x86_64
567 There is a carry propagating bug in the x86_64 Montgomery squaring
568 procedure. No EC algorithms are affected. Analysis suggests that attacks
569 against RSA and DSA as a result of this defect would be very difficult to
570 perform and are not believed likely. Attacks against DH are considered just
571 feasible (although very difficult) because most of the work necessary to
572 deduce information about a private key may be performed offline. The amount
573 of resources required for such an attack would be very significant and
574 likely only accessible to a limited number of attackers. An attacker would
575 additionally need online access to an unpatched system using the target
576 private key in a scenario with persistent DH parameters and a private
577 key that is shared between multiple clients. For example this can occur by
578 default in OpenSSL DHE based SSL/TLS ciphersuites.
580 This issue was reported to OpenSSL by Hanno Böck.
584 *) Certificate verify crash with missing PSS parameter
586 The signature verification routines will crash with a NULL pointer
587 dereference if presented with an ASN.1 signature using the RSA PSS
588 algorithm and absent mask generation function parameter. Since these
589 routines are used to verify certificate signature algorithms this can be
590 used to crash any certificate verification operation and exploited in a
591 DoS attack. Any application which performs certificate verification is
592 vulnerable including OpenSSL clients and servers which enable client
595 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
599 *) X509_ATTRIBUTE memory leak
601 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
602 memory. This structure is used by the PKCS#7 and CMS routines so any
603 application which reads PKCS#7 or CMS data from untrusted sources is
604 affected. SSL/TLS is not affected.
606 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
611 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
612 This changes the decoding behaviour for some invalid messages,
613 though the change is mostly in the more lenient direction, and
614 legacy behaviour is preserved as much as possible.
617 *) In DSA_generate_parameters_ex, if the provided seed is too short,
618 use a random seed, as already documented.
619 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
621 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
623 *) Alternate chains certificate forgery
625 During certificate verfification, OpenSSL will attempt to find an
626 alternative certificate chain if the first attempt to build such a chain
627 fails. An error in the implementation of this logic can mean that an
628 attacker could cause certain checks on untrusted certificates to be
629 bypassed, such as the CA flag, enabling them to use a valid leaf
630 certificate to act as a CA and "issue" an invalid certificate.
632 This issue was reported to OpenSSL by Adam Langley/David Benjamin
637 *) Race condition handling PSK identify hint
639 If PSK identity hints are received by a multi-threaded client then
640 the values are wrongly updated in the parent SSL_CTX structure. This can
641 result in a race condition potentially leading to a double free of the
646 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
648 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
649 incompatibility in the handling of HMAC. The previous ABI has now been
652 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
654 *) Malformed ECParameters causes infinite loop
656 When processing an ECParameters structure OpenSSL enters an infinite loop
657 if the curve specified is over a specially malformed binary polynomial
660 This can be used to perform denial of service against any
661 system which processes public keys, certificate requests or
662 certificates. This includes TLS clients and TLS servers with
663 client authentication enabled.
665 This issue was reported to OpenSSL by Joseph Barr-Pixton.
669 *) Exploitable out-of-bounds read in X509_cmp_time
671 X509_cmp_time does not properly check the length of the ASN1_TIME
672 string and can read a few bytes out of bounds. In addition,
673 X509_cmp_time accepts an arbitrary number of fractional seconds in the
676 An attacker can use this to craft malformed certificates and CRLs of
677 various sizes and potentially cause a segmentation fault, resulting in
678 a DoS on applications that verify certificates or CRLs. TLS clients
679 that verify CRLs are affected. TLS clients and servers with client
680 authentication enabled may be affected if they use custom verification
683 This issue was reported to OpenSSL by Robert Swiecki (Google), and
684 independently by Hanno Böck.
688 *) PKCS7 crash with missing EnvelopedContent
690 The PKCS#7 parsing code does not handle missing inner EncryptedContent
691 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
692 with missing content and trigger a NULL pointer dereference on parsing.
694 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
695 structures from untrusted sources are affected. OpenSSL clients and
696 servers are not affected.
698 This issue was reported to OpenSSL by Michal Zalewski (Google).
702 *) CMS verify infinite loop with unknown hash function
704 When verifying a signedData message the CMS code can enter an infinite loop
705 if presented with an unknown hash function OID. This can be used to perform
706 denial of service against any system which verifies signedData messages using
708 This issue was reported to OpenSSL by Johannes Bauer.
712 *) Race condition handling NewSessionTicket
714 If a NewSessionTicket is received by a multi-threaded client when attempting to
715 reuse a previous ticket then a race condition can occur potentially leading to
716 a double free of the ticket data.
720 *) Removed support for the two export grade static DH ciphersuites
721 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
722 were newly added (along with a number of other static DH ciphersuites) to
723 1.0.2. However the two export ones have *never* worked since they were
724 introduced. It seems strange in any case to be adding new export
725 ciphersuites, and given "logjam" it also does not seem correct to fix them.
728 *) Only support 256-bit or stronger elliptic curves with the
729 'ecdh_auto' setting (server) or by default (client). Of supported
730 curves, prefer P-256 (both).
733 *) Reject DH handshakes with parameters shorter than 768 bits.
734 [Kurt Roeckx and Emilia Kasper]
736 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
738 *) ClientHello sigalgs DoS fix
740 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
741 invalid signature algorithms extension a NULL pointer dereference will
742 occur. This can be exploited in a DoS attack against the server.
744 This issue was was reported to OpenSSL by David Ramos of Stanford
747 [Stephen Henson and Matt Caswell]
749 *) Multiblock corrupted pointer fix
751 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
752 feature only applies on 64 bit x86 architecture platforms that support AES
753 NI instructions. A defect in the implementation of "multiblock" can cause
754 OpenSSL's internal write buffer to become incorrectly set to NULL when
755 using non-blocking IO. Typically, when the user application is using a
756 socket BIO for writing, this will only result in a failed connection.
757 However if some other BIO is used then it is likely that a segmentation
758 fault will be triggered, thus enabling a potential DoS attack.
760 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
764 *) Segmentation fault in DTLSv1_listen fix
766 The DTLSv1_listen function is intended to be stateless and processes the
767 initial ClientHello from many peers. It is common for user code to loop
768 over the call to DTLSv1_listen until a valid ClientHello is received with
769 an associated cookie. A defect in the implementation of DTLSv1_listen means
770 that state is preserved in the SSL object from one invocation to the next
771 that can lead to a segmentation fault. Errors processing the initial
772 ClientHello can trigger this scenario. An example of such an error could be
773 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
776 This issue was reported to OpenSSL by Per Allansson.
780 *) Segmentation fault in ASN1_TYPE_cmp fix
782 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
783 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
784 certificate signature algorithm consistency this can be used to crash any
785 certificate verification operation and exploited in a DoS attack. Any
786 application which performs certificate verification is vulnerable including
787 OpenSSL clients and servers which enable client authentication.
791 *) Segmentation fault for invalid PSS parameters fix
793 The signature verification routines will crash with a NULL pointer
794 dereference if presented with an ASN.1 signature using the RSA PSS
795 algorithm and invalid parameters. Since these routines are used to verify
796 certificate signature algorithms this can be used to crash any
797 certificate verification operation and exploited in a DoS attack. Any
798 application which performs certificate verification is vulnerable including
799 OpenSSL clients and servers which enable client authentication.
801 This issue was was reported to OpenSSL by Brian Carpenter.
805 *) ASN.1 structure reuse memory corruption fix
807 Reusing a structure in ASN.1 parsing may allow an attacker to cause
808 memory corruption via an invalid write. Such reuse is and has been
809 strongly discouraged and is believed to be rare.
811 Applications that parse structures containing CHOICE or ANY DEFINED BY
812 components may be affected. Certificate parsing (d2i_X509 and related
813 functions) are however not affected. OpenSSL clients and servers are
818 *) PKCS7 NULL pointer dereferences fix
820 The PKCS#7 parsing code does not handle missing outer ContentInfo
821 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
822 missing content and trigger a NULL pointer dereference on parsing.
824 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
825 otherwise parse PKCS#7 structures from untrusted sources are
826 affected. OpenSSL clients and servers are not affected.
828 This issue was reported to OpenSSL by Michal Zalewski (Google).
832 *) DoS via reachable assert in SSLv2 servers fix
834 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
835 servers that both support SSLv2 and enable export cipher suites by sending
836 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
838 This issue was discovered by Sean Burford (Google) and Emilia Käsper
839 (OpenSSL development team).
843 *) Empty CKE with client auth and DHE fix
845 If client auth is used then a server can seg fault in the event of a DHE
846 ciphersuite being selected and a zero length ClientKeyExchange message
847 being sent by the client. This could be exploited in a DoS attack.
851 *) Handshake with unseeded PRNG fix
853 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
854 with an unseeded PRNG. The conditions are:
855 - The client is on a platform where the PRNG has not been seeded
856 automatically, and the user has not seeded manually
857 - A protocol specific client method version has been used (i.e. not
858 SSL_client_methodv23)
859 - A ciphersuite is used that does not require additional random data from
860 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
862 If the handshake succeeds then the client random that has been used will
863 have been generated from a PRNG with insufficient entropy and therefore the
864 output may be predictable.
866 For example using the following command with an unseeded openssl will
867 succeed on an unpatched platform:
869 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
873 *) Use After Free following d2i_ECPrivatekey error fix
875 A malformed EC private key file consumed via the d2i_ECPrivateKey function
876 could cause a use after free condition. This, in turn, could cause a double
877 free in several private key parsing functions (such as d2i_PrivateKey
878 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
879 for applications that receive EC private keys from untrusted
880 sources. This scenario is considered rare.
882 This issue was discovered by the BoringSSL project and fixed in their
887 *) X509_to_X509_REQ NULL pointer deref fix
889 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
890 the certificate key is invalid. This function is rarely used in practice.
892 This issue was discovered by Brian Carpenter.
896 *) Removed the export ciphers from the DEFAULT ciphers
899 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
901 *) Change RSA and DH/DSA key generation apps to generate 2048-bit
905 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
906 ARMv5 through ARMv8, as opposite to "locking" it to single one.
907 So far those who have to target multiple plaforms would compromise
908 and argue that binary targeting say ARMv5 would still execute on
909 ARMv8. "Universal" build resolves this compromise by providing
910 near-optimal performance even on newer platforms.
913 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
914 (other platforms pending).
915 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
917 *) Add support for the SignedCertificateTimestampList certificate and
918 OCSP response extensions from RFC6962.
921 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
922 for corner cases. (Certain input points at infinity could lead to
923 bogus results, with non-infinity inputs mapped to infinity too.)
926 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
927 This covers AES, SHA256/512 and GHASH. "Initial" means that most
928 common cases are optimized and there still is room for further
929 improvements. Vector Permutation AES for Altivec is also added.
932 *) Add support for little-endian ppc64 Linux target.
933 [Marcelo Cerri (IBM)]
935 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
936 SHA1, SHA256 and GHASH. "Initial" means that most common cases
937 are optimized and there still is room for further improvements.
938 Both 32- and 64-bit modes are supported.
939 [Andy Polyakov, Ard Biesheuvel (Linaro)]
941 *) Improved ARMv7 NEON support.
944 *) Support for SPARC Architecture 2011 crypto extensions, first
945 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
946 SHA256/512, MD5, GHASH and modular exponentiation.
947 [Andy Polyakov, David Miller]
949 *) Accelerated modular exponentiation for Intel processors, a.k.a.
951 [Shay Gueron & Vlad Krasnov (Intel Corp)]
953 *) Support for new and upcoming Intel processors, including AVX2,
954 BMI and SHA ISA extensions. This includes additional "stitched"
955 implementations, AESNI-SHA256 and GCM, and multi-buffer support
958 This work was sponsored by Intel Corp.
961 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
962 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
963 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
966 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
967 this fixes a limiation in previous versions of OpenSSL.
970 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
971 MGF1 digest and OAEP label.
974 *) Add EVP support for key wrapping algorithms, to avoid problems with
975 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
976 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
977 algorithms and include tests cases.
980 *) Add functions to allocate and set the fields of an ECDSA_METHOD
982 [Douglas E. Engert, Steve Henson]
984 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
985 difference in days and seconds between two tm or ASN1_TIME structures.
988 *) Add -rev test option to s_server to just reverse order of characters
989 received by client and send back to server. Also prints an abbreviated
990 summary of the connection parameters.
993 *) New option -brief for s_client and s_server to print out a brief summary
994 of connection parameters.
997 *) Add callbacks for arbitrary TLS extensions.
998 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1000 *) New option -crl_download in several openssl utilities to download CRLs
1001 from CRLDP extension in certificates.
1004 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1007 *) New function X509_CRL_diff to generate a delta CRL from the difference
1008 of two full CRLs. Add support to "crl" utility.
1011 *) New functions to set lookup_crls function and to retrieve
1012 X509_STORE from X509_STORE_CTX.
1015 *) Print out deprecated issuer and subject unique ID fields in
1019 *) Extend OCSP I/O functions so they can be used for simple general purpose
1020 HTTP as well as OCSP. New wrapper function which can be used to download
1021 CRLs using the OCSP API.
1024 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1027 *) SSL_CONF* functions. These provide a common framework for application
1028 configuration using configuration files or command lines.
1031 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1032 message callback and prints the results. Needs compile time option
1033 "enable-ssl-trace". New options to s_client and s_server to enable
1037 *) New ctrl and macro to retrieve supported points extensions.
1038 Print out extension in s_server and s_client.
1041 *) New functions to retrieve certificate signature and signature
1045 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1049 *) New Suite B modes for TLS code. These use and enforce the requirements
1050 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1051 only use Suite B curves. The Suite B modes can be set by using the
1052 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1055 *) New chain verification flags for Suite B levels of security. Check
1056 algorithms are acceptable when flags are set in X509_verify_cert.
1059 *) Make tls1_check_chain return a set of flags indicating checks passed
1060 by a certificate chain. Add additional tests to handle client
1061 certificates: checks for matching certificate type and issuer name
1065 *) If an attempt is made to use a signature algorithm not in the peer
1066 preference list abort the handshake. If client has no suitable
1067 signature algorithms in response to a certificate request do not
1068 use the certificate.
1071 *) If server EC tmp key is not in client preference list abort handshake.
1074 *) Add support for certificate stores in CERT structure. This makes it
1075 possible to have different stores per SSL structure or one store in
1076 the parent SSL_CTX. Include distint stores for certificate chain
1077 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1078 to build and store a certificate chain in CERT structure: returing
1079 an error if the chain cannot be built: this will allow applications
1080 to test if a chain is correctly configured.
1082 Note: if the CERT based stores are not set then the parent SSL_CTX
1083 store is used to retain compatibility with existing behaviour.
1087 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1088 mask based on the current session, check mask when sending client
1089 hello and checking the requested ciphersuite.
1092 *) New ctrls to retrieve and set certificate types in a certificate
1093 request message. Print out received values in s_client. If certificate
1094 types is not set with custom values set sensible values based on
1095 supported signature algorithms.
1098 *) Support for distinct client and server supported signature algorithms.
1101 *) Add certificate callback. If set this is called whenever a certificate
1102 is required by client or server. An application can decide which
1103 certificate chain to present based on arbitrary criteria: for example
1104 supported signature algorithms. Add very simple example to s_server.
1105 This fixes many of the problems and restrictions of the existing client
1106 certificate callback: for example you can now clear an existing
1107 certificate and specify the whole chain.
1110 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1111 the certificate can be used for (if anything). Set valid_flags field
1112 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1113 to have similar checks in it.
1115 Add new "cert_flags" field to CERT structure and include a "strict mode".
1116 This enforces some TLS certificate requirements (such as only permitting
1117 certificate signature algorithms contained in the supported algorithms
1118 extension) which some implementations ignore: this option should be used
1119 with caution as it could cause interoperability issues.
1122 *) Update and tidy signature algorithm extension processing. Work out
1123 shared signature algorithms based on preferences and peer algorithms
1124 and print them out in s_client and s_server. Abort handshake if no
1125 shared signature algorithms.
1128 *) Add new functions to allow customised supported signature algorithms
1129 for SSL and SSL_CTX structures. Add options to s_client and s_server
1133 *) New function SSL_certs_clear() to delete all references to certificates
1134 from an SSL structure. Before this once a certificate had been added
1135 it couldn't be removed.
1138 *) Integrate hostname, email address and IP address checking with certificate
1139 verification. New verify options supporting checking in opensl utility.
1142 *) Fixes and wildcard matching support to hostname and email checking
1143 functions. Add manual page.
1144 [Florian Weimer (Red Hat Product Security Team)]
1146 *) New functions to check a hostname email or IP address against a
1147 certificate. Add options x509 utility to print results of checks against
1151 *) Fix OCSP checking.
1152 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1154 *) Initial experimental support for explicitly trusted non-root CAs.
1155 OpenSSL still tries to build a complete chain to a root but if an
1156 intermediate CA has a trust setting included that is used. The first
1157 setting is used: whether to trust (e.g., -addtrust option to the x509
1161 *) Add -trusted_first option which attempts to find certificates in the
1162 trusted store even if an untrusted chain is also supplied.
1165 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1166 platform support for Linux and Android.
1169 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1172 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1173 When in FIPS mode the approved implementations are used as normal,
1174 when not in FIPS mode the internal unapproved versions are used instead.
1175 This means that the FIPS capable OpenSSL isn't forced to use the
1176 (often lower perfomance) FIPS implementations outside FIPS mode.
1179 *) Transparently support X9.42 DH parameters when calling
1180 PEM_read_bio_DHparameters. This means existing applications can handle
1181 the new parameter format automatically.
1184 *) Initial experimental support for X9.42 DH parameter format: mainly
1185 to support use of 'q' parameter for RFC5114 parameters.
1188 *) Add DH parameters from RFC5114 including test data to dhtest.
1191 *) Support for automatic EC temporary key parameter selection. If enabled
1192 the most preferred EC parameters are automatically used instead of
1193 hardcoded fixed parameters. Now a server just has to call:
1194 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1195 support ECDH and use the most appropriate parameters.
1198 *) Enhance and tidy EC curve and point format TLS extension code. Use
1199 static structures instead of allocation if default values are used.
1200 New ctrls to set curves we wish to support and to retrieve shared curves.
1201 Print out shared curves in s_server. New options to s_server and s_client
1202 to set list of supported curves.
1205 *) New ctrls to retrieve supported signature algorithms and
1206 supported curve values as an array of NIDs. Extend openssl utility
1207 to print out received values.
1210 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1211 between NIDs and the more common NIST names such as "P-256". Enhance
1212 ecparam utility and ECC method to recognise the NIST names for curves.
1215 *) Enhance SSL/TLS certificate chain handling to support different
1216 chains for each certificate instead of one chain in the parent SSL_CTX.
1219 *) Support for fixed DH ciphersuite client authentication: where both
1220 server and client use DH certificates with common parameters.
1223 *) Support for fixed DH ciphersuites: those requiring DH server
1227 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1229 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1230 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1231 X509_CINF_get_signature were reverted post internal team review.
1233 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1235 *) Build fixes for the Windows and OpenVMS platforms
1236 [Matt Caswell and Richard Levitte]
1238 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1240 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1241 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1242 dereference. This could lead to a Denial Of Service attack. Thanks to
1243 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1247 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1248 dtls1_buffer_record function under certain conditions. In particular this
1249 could occur if an attacker sent repeated DTLS records with the same
1250 sequence number but for the next epoch. The memory leak could be exploited
1251 by an attacker in a Denial of Service attack through memory exhaustion.
1252 Thanks to Chris Mueller for reporting this issue.
1256 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1257 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1258 method would be set to NULL which could later result in a NULL pointer
1259 dereference. Thanks to Frank Schmirler for reporting this issue.
1263 *) Abort handshake if server key exchange message is omitted for ephemeral
1266 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1267 reporting this issue.
1271 *) Remove non-export ephemeral RSA code on client and server. This code
1272 violated the TLS standard by allowing the use of temporary RSA keys in
1273 non-export ciphersuites and could be used by a server to effectively
1274 downgrade the RSA key length used to a value smaller than the server
1275 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1276 INRIA or reporting this issue.
1280 *) Fixed issue where DH client certificates are accepted without verification.
1281 An OpenSSL server will accept a DH certificate for client authentication
1282 without the certificate verify message. This effectively allows a client to
1283 authenticate without the use of a private key. This only affects servers
1284 which trust a client certificate authority which issues certificates
1285 containing DH keys: these are extremely rare and hardly ever encountered.
1286 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1291 *) Ensure that the session ID context of an SSL is updated when its
1292 SSL_CTX is updated via SSL_set_SSL_CTX.
1294 The session ID context is typically set from the parent SSL_CTX,
1295 and can vary with the CTX.
1298 *) Fix various certificate fingerprint issues.
1300 By using non-DER or invalid encodings outside the signed portion of a
1301 certificate the fingerprint can be changed without breaking the signature.
1302 Although no details of the signed portion of the certificate can be changed
1303 this can cause problems with some applications: e.g. those using the
1304 certificate fingerprint for blacklists.
1306 1. Reject signatures with non zero unused bits.
1308 If the BIT STRING containing the signature has non zero unused bits reject
1309 the signature. All current signature algorithms require zero unused bits.
1311 2. Check certificate algorithm consistency.
1313 Check the AlgorithmIdentifier inside TBS matches the one in the
1314 certificate signature. NB: this will result in signature failure
1315 errors for some broken certificates.
1317 Thanks to Konrad Kraszewski from Google for reporting this issue.
1319 3. Check DSA/ECDSA signatures use DER.
1321 Reencode DSA/ECDSA signatures and compare with the original received
1322 signature. Return an error if there is a mismatch.
1324 This will reject various cases including garbage after signature
1325 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1326 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1327 (negative or with leading zeroes).
1329 Further analysis was conducted and fixes were developed by Stephen Henson
1330 of the OpenSSL core team.
1335 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1336 results on some platforms, including x86_64. This bug occurs at random
1337 with a very low probability, and is not known to be exploitable in any
1338 way, though its exact impact is difficult to determine. Thanks to Pieter
1339 Wuille (Blockstream) who reported this issue and also suggested an initial
1340 fix. Further analysis was conducted by the OpenSSL development team and
1341 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1342 the OpenSSL core team.
1346 *) Do not resume sessions on the server if the negotiated protocol
1347 version does not match the session's version. Resuming with a different
1348 version, while not strictly forbidden by the RFC, is of questionable
1349 sanity and breaks all known clients.
1350 [David Benjamin, Emilia Käsper]
1352 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1353 early CCS messages during renegotiation. (Note that because
1354 renegotiation is encrypted, this early CCS was not exploitable.)
1357 *) Tighten client-side session ticket handling during renegotiation:
1358 ensure that the client only accepts a session ticket if the server sends
1359 the extension anew in the ServerHello. Previously, a TLS client would
1360 reuse the old extension state and thus accept a session ticket if one was
1361 announced in the initial ServerHello.
1363 Similarly, ensure that the client requires a session ticket if one
1364 was advertised in the ServerHello. Previously, a TLS client would
1365 ignore a missing NewSessionTicket message.
1368 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1370 *) SRTP Memory Leak.
1372 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1373 sends a carefully crafted handshake message, to cause OpenSSL to fail
1374 to free up to 64k of memory causing a memory leak. This could be
1375 exploited in a Denial Of Service attack. This issue affects OpenSSL
1376 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1377 whether SRTP is used or configured. Implementations of OpenSSL that
1378 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1380 The fix was developed by the OpenSSL team.
1384 *) Session Ticket Memory Leak.
1386 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1387 integrity of that ticket is first verified. In the event of a session
1388 ticket integrity check failing, OpenSSL will fail to free memory
1389 causing a memory leak. By sending a large number of invalid session
1390 tickets an attacker could exploit this issue in a Denial Of Service
1395 *) Build option no-ssl3 is incomplete.
1397 When OpenSSL is configured with "no-ssl3" as a build option, servers
1398 could accept and complete a SSL 3.0 handshake, and clients could be
1399 configured to send them.
1401 [Akamai and the OpenSSL team]
1403 *) Add support for TLS_FALLBACK_SCSV.
1404 Client applications doing fallback retries should call
1405 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1407 [Adam Langley, Bodo Moeller]
1409 *) Add additional DigestInfo checks.
1411 Reencode DigestInto in DER and check against the original when
1412 verifying RSA signature: this will reject any improperly encoded
1413 DigestInfo structures.
1415 Note: this is a precautionary measure and no attacks are currently known.
1419 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1421 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1422 SRP code can be overrun an internal buffer. Add sanity check that
1423 g, A, B < N to SRP code.
1425 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1426 Group for discovering this issue.
1430 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1431 TLS 1.0 instead of higher protocol versions when the ClientHello message
1432 is badly fragmented. This allows a man-in-the-middle attacker to force a
1433 downgrade to TLS 1.0 even if both the server and the client support a
1434 higher protocol version, by modifying the client's TLS records.
1436 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1437 researching this issue.
1441 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1442 to a denial of service attack. A malicious server can crash the client
1443 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1444 ciphersuite and sending carefully crafted handshake messages.
1446 Thanks to Felix Gröbert (Google) for discovering and researching this
1451 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1452 to leak memory. This can be exploited through a Denial of Service attack.
1453 Thanks to Adam Langley for discovering and researching this issue.
1457 *) An attacker can force openssl to consume large amounts of memory whilst
1458 processing DTLS handshake messages. This can be exploited through a
1459 Denial of Service attack.
1460 Thanks to Adam Langley for discovering and researching this issue.
1464 *) An attacker can force an error condition which causes openssl to crash
1465 whilst processing DTLS packets due to memory being freed twice. This
1466 can be exploited through a Denial of Service attack.
1467 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1472 *) If a multithreaded client connects to a malicious server using a resumed
1473 session and the server sends an ec point format extension it could write
1474 up to 255 bytes to freed memory.
1476 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1481 *) A malicious server can crash an OpenSSL client with a null pointer
1482 dereference (read) by specifying an SRP ciphersuite even though it was not
1483 properly negotiated with the client. This can be exploited through a
1484 Denial of Service attack.
1486 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1487 discovering and researching this issue.
1491 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1492 X509_name_oneline, X509_name_print_ex et al. to leak some information
1493 from the stack. Applications may be affected if they echo pretty printing
1494 output to the attacker.
1496 Thanks to Ivan Fratric (Google) for discovering this issue.
1498 [Emilia Käsper, and Steve Henson]
1500 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1501 for corner cases. (Certain input points at infinity could lead to
1502 bogus results, with non-infinity inputs mapped to infinity too.)
1505 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1507 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1508 handshake can force the use of weak keying material in OpenSSL
1509 SSL/TLS clients and servers.
1511 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1512 researching this issue. (CVE-2014-0224)
1513 [KIKUCHI Masashi, Steve Henson]
1515 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1516 OpenSSL DTLS client the code can be made to recurse eventually crashing
1519 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1521 [Imre Rad, Steve Henson]
1523 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1524 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1525 client or server. This is potentially exploitable to run arbitrary
1526 code on a vulnerable client or server.
1528 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1529 [Jüri Aedla, Steve Henson]
1531 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1532 are subject to a denial of service attack.
1534 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1535 this issue. (CVE-2014-3470)
1536 [Felix Gröbert, Ivan Fratric, Steve Henson]
1538 *) Harmonize version and its documentation. -f flag is used to display
1540 [mancha <mancha1@zoho.com>]
1542 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1543 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1544 [mancha <mancha1@zoho.com>]
1546 *) Fix some double frees. These are not thought to be exploitable.
1547 [mancha <mancha1@zoho.com>]
1549 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1551 *) A missing bounds check in the handling of the TLS heartbeat extension
1552 can be used to reveal up to 64k of memory to a connected client or
1555 Thanks for Neel Mehta of Google Security for discovering this bug and to
1556 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1557 preparing the fix (CVE-2014-0160)
1558 [Adam Langley, Bodo Moeller]
1560 *) Fix for the attack described in the paper "Recovering OpenSSL
1561 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1562 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1563 http://eprint.iacr.org/2014/140
1565 Thanks to Yuval Yarom and Naomi Benger for discovering this
1566 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1567 [Yuval Yarom and Naomi Benger]
1569 *) TLS pad extension: draft-agl-tls-padding-03
1571 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1572 TLS client Hello record length value would otherwise be > 255 and
1573 less that 512 pad with a dummy extension containing zeroes so it
1574 is at least 512 bytes long.
1576 [Adam Langley, Steve Henson]
1578 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1580 *) Fix for TLS record tampering bug. A carefully crafted invalid
1581 handshake could crash OpenSSL with a NULL pointer exception.
1582 Thanks to Anton Johansson for reporting this issues.
1585 *) Keep original DTLS digest and encryption contexts in retransmission
1586 structures so we can use the previous session parameters if they need
1587 to be resent. (CVE-2013-6450)
1590 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1591 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1592 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1593 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1594 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1595 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1596 [Rob Stradling, Adam Langley]
1598 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1600 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1601 supporting platforms or when small records were transferred.
1602 [Andy Polyakov, Steve Henson]
1604 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1606 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1608 This addresses the flaw in CBC record processing discovered by
1609 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1610 at: http://www.isg.rhul.ac.uk/tls/
1612 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1613 Security Group at Royal Holloway, University of London
1614 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1615 Emilia Käsper for the initial patch.
1617 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1619 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1620 ciphersuites which can be exploited in a denial of service attack.
1621 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1622 and detecting this bug and to Wolfgang Ettlinger
1623 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1627 *) Return an error when checking OCSP signatures when key is NULL.
1628 This fixes a DoS attack. (CVE-2013-0166)
1631 *) Make openssl verify return errors.
1632 [Chris Palmer <palmer@google.com> and Ben Laurie]
1634 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1635 the right response is stapled. Also change SSL_get_certificate()
1636 so it returns the certificate actually sent.
1637 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1638 [Rob Stradling <rob.stradling@comodo.com>]
1640 *) Fix possible deadlock when decoding public keys.
1643 *) Don't use TLS 1.0 record version number in initial client hello
1647 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1649 *) Sanity check record length before skipping explicit IV in TLS
1650 1.2, 1.1 and DTLS to fix DoS attack.
1652 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1653 fuzzing as a service testing platform.
1657 *) Initialise tkeylen properly when encrypting CMS messages.
1658 Thanks to Solar Designer of Openwall for reporting this issue.
1661 *) In FIPS mode don't try to use composite ciphers as they are not
1665 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1667 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1668 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1669 mean any application compiled against OpenSSL 1.0.0 headers setting
1670 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1671 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1672 0x10000000L Any application which was previously compiled against
1673 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1674 will need to be recompiled as a result. Letting be results in
1675 inability to disable specifically TLS 1.1 and in client context,
1676 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1679 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1680 disable just protocol X, but all protocols above X *if* there are
1681 protocols *below* X still enabled. In more practical terms it means
1682 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1683 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1684 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1688 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1690 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1691 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1692 in CRYPTO_realloc_clean.
1694 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1695 issue and to Adam Langley <agl@chromium.org> for fixing it.
1697 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1699 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1702 *) Workarounds for some broken servers that "hang" if a client hello
1703 record length exceeds 255 bytes.
1705 1. Do not use record version number > TLS 1.0 in initial client
1706 hello: some (but not all) hanging servers will now work.
1707 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1708 the number of ciphers sent in the client hello. This should be
1709 set to an even number, such as 50, for example by passing:
1710 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1711 Most broken servers should now work.
1712 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1713 TLS 1.2 client support entirely.
1716 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1719 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1721 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1722 STRING form instead of a DigestInfo.
1725 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1726 and the RSA_sign/RSA_verify functions. This was made more apparent when
1727 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1728 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1729 the correct format in RSA_verify so both forms transparently work.
1732 *) Some servers which support TLS 1.0 can choke if we initially indicate
1733 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1734 encrypted premaster secret. As a workaround use the maximum pemitted
1735 client version in client hello, this should keep such servers happy
1736 and still work with previous versions of OpenSSL.
1739 *) Add support for TLS/DTLS heartbeats.
1740 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1742 *) Add support for SCTP.
1743 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1745 *) Improved PRNG seeding for VOS.
1746 [Paul Green <Paul.Green@stratus.com>]
1748 *) Extensive assembler packs updates, most notably:
1750 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1751 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1752 - x86_64: bit-sliced AES implementation;
1753 - ARM: NEON support, contemporary platforms optimizations;
1754 - s390x: z196 support;
1755 - *: GHASH and GF(2^m) multiplication implementations;
1759 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1760 (removal of unnecessary code)
1761 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1763 *) Add TLS key material exporter from RFC 5705.
1766 *) Add DTLS-SRTP negotiation from RFC 5764.
1769 *) Add Next Protocol Negotiation,
1770 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1771 disabled with a no-npn flag to config or Configure. Code donated
1773 [Adam Langley <agl@google.com> and Ben Laurie]
1775 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1776 NIST-P256, NIST-P521, with constant-time single point multiplication on
1777 typical inputs. Compiler support for the nonstandard type __uint128_t is
1778 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1779 Code made available under Apache License version 2.0.
1781 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1782 line to include this in your build of OpenSSL, and run "make depend" (or
1783 "make update"). This enables the following EC_METHODs:
1785 EC_GFp_nistp224_method()
1786 EC_GFp_nistp256_method()
1787 EC_GFp_nistp521_method()
1789 EC_GROUP_new_by_curve_name() will automatically use these (while
1790 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1792 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1794 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1795 all platforms. Move ssize_t definition from e_os.h to the public
1796 header file e_os2.h as it now appears in public header file cms.h
1799 *) New -sigopt option to the ca, req and x509 utilities. Additional
1800 signature parameters can be passed using this option and in
1804 *) Add RSA PSS signing function. This will generate and set the
1805 appropriate AlgorithmIdentifiers for PSS based on those in the
1806 corresponding EVP_MD_CTX structure. No application support yet.
1809 *) Support for companion algorithm specific ASN1 signing routines.
1810 New function ASN1_item_sign_ctx() signs a pre-initialised
1811 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1812 the appropriate parameters.
1815 *) Add new algorithm specific ASN1 verification initialisation function
1816 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1817 handling will be the same no matter what EVP_PKEY_METHOD is used.
1818 Add a PSS handler to support verification of PSS signatures: checked
1819 against a number of sample certificates.
1822 *) Add signature printing for PSS. Add PSS OIDs.
1823 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1825 *) Add algorithm specific signature printing. An individual ASN1 method
1826 can now print out signatures instead of the standard hex dump.
1828 More complex signatures (e.g. PSS) can print out more meaningful
1829 information. Include DSA version that prints out the signature
1833 *) Password based recipient info support for CMS library: implementing
1837 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1838 neatly separates the code into cipher and PBE sections and is required
1839 for some algorithms that split PBES2 into separate pieces (such as
1840 password based CMS).
1843 *) Session-handling fixes:
1844 - Fix handling of connections that are resuming with a session ID,
1845 but also support Session Tickets.
1846 - Fix a bug that suppressed issuing of a new ticket if the client
1847 presented a ticket with an expired session.
1848 - Try to set the ticket lifetime hint to something reasonable.
1849 - Make tickets shorter by excluding irrelevant information.
1850 - On the client side, don't ignore renewed tickets.
1851 [Adam Langley, Bodo Moeller (Google)]
1853 *) Fix PSK session representation.
1856 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1858 This work was sponsored by Intel.
1861 *) Add GCM support to TLS library. Some custom code is needed to split
1862 the IV between the fixed (from PRF) and explicit (from TLS record)
1863 portions. This adds all GCM ciphersuites supported by RFC5288 and
1864 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
1865 add a special AESGCM string for GCM only.
1868 *) Expand range of ctrls for AES GCM. Permit setting invocation
1869 field on decrypt and retrieval of invocation field only on encrypt.
1872 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1873 As required by RFC5289 these ciphersuites cannot be used if for
1874 versions of TLS earlier than 1.2.
1877 *) For FIPS capable OpenSSL interpret a NULL default public key method
1878 as unset and return the appopriate default but do *not* set the default.
1879 This means we can return the appopriate method in applications that
1880 swicth between FIPS and non-FIPS modes.
1883 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1884 ENGINE is used then we cannot handle that in the FIPS module so we
1885 keep original code iff non-FIPS operations are allowed.
1888 *) Add -attime option to openssl utilities.
1889 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1891 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1894 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1895 FIPS EC methods unconditionally for now.
1898 *) New build option no-ec2m to disable characteristic 2 code.
1901 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1902 all cases can be covered as some introduce binary incompatibilities.
1905 *) Redirect RSA operations to FIPS module including keygen,
1906 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1909 *) Add similar low level API blocking to ciphers.
1912 *) Low level digest APIs are not approved in FIPS mode: any attempt
1913 to use these will cause a fatal error. Applications that *really* want
1914 to use them can use the private_* version instead.
1917 *) Redirect cipher operations to FIPS module for FIPS builds.
1920 *) Redirect digest operations to FIPS module for FIPS builds.
1923 *) Update build system to add "fips" flag which will link in fipscanister.o
1924 for static and shared library builds embedding a signature if needed.
1927 *) Output TLS supported curves in preference order instead of numerical
1928 order. This is currently hardcoded for the highest order curves first.
1929 This should be configurable so applications can judge speed vs strength.
1932 *) Add TLS v1.2 server support for client authentication.
1935 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1939 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1940 FIPS modules versions.
1943 *) Add TLS v1.2 client side support for client authentication. Keep cache
1944 of handshake records longer as we don't know the hash algorithm to use
1945 until after the certificate request message is received.
1948 *) Initial TLS v1.2 client support. Add a default signature algorithms
1949 extension including all the algorithms we support. Parse new signature
1950 format in client key exchange. Relax some ECC signing restrictions for
1951 TLS v1.2 as indicated in RFC5246.
1954 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1955 to new signature format when needed using client digest preference.
1956 All server ciphersuites should now work correctly in TLS v1.2. No client
1957 support yet and no support for client certificates.
1960 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1961 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1962 ciphersuites. At present only RSA key exchange ciphersuites work with
1963 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1964 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1965 and version checking.
1968 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1969 with this defined it will not be affected by any changes to ssl internal
1970 structures. Add several utility functions to allow openssl application
1971 to work with OPENSSL_NO_SSL_INTERN defined.
1975 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1977 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1980 *) Permit abbreviated handshakes when renegotiating using the function
1981 SSL_renegotiate_abbreviated().
1982 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1984 *) Add call to ENGINE_register_all_complete() to
1985 ENGINE_load_builtin_engines(), so some implementations get used
1986 automatically instead of needing explicit application support.
1989 *) Add support for TLS key exporter as described in RFC5705.
1990 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1992 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1993 a few changes are required:
1995 Add SSL_OP_NO_TLSv1_1 flag.
1996 Add TLSv1_1 methods.
1997 Update version checking logic to handle version 1.1.
1998 Add explicit IV handling (ported from DTLS code).
1999 Add command line options to s_client/s_server.
2002 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2004 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2005 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2006 content decryption and always return the same error. Note: this attack
2007 needs on average 2^20 messages so it only affects automated senders. The
2008 old behaviour can be reenabled in the CMS code by setting the
2009 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2010 an MMA defence is not necessary.
2011 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2012 this issue. (CVE-2012-0884)
2015 *) Fix CVE-2011-4619: make sure we really are receiving a
2016 client hello before rejecting multiple SGC restarts. Thanks to
2017 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2020 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2022 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2023 Thanks to Antonio Martin, Enterprise Secure Access Research and
2024 Development, Cisco Systems, Inc. for discovering this bug and
2025 preparing a fix. (CVE-2012-0050)
2028 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2030 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2031 of the Vaudenay padding oracle attack on CBC mode encryption
2032 which enables an efficient plaintext recovery attack against
2033 the OpenSSL implementation of DTLS. Their attack exploits timing
2034 differences arising during decryption processing. A research
2035 paper describing this attack can be found at:
2036 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2037 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2038 Security Group at Royal Holloway, University of London
2039 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2040 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2041 for preparing the fix. (CVE-2011-4108)
2042 [Robin Seggelmann, Michael Tuexen]
2044 *) Clear bytes used for block padding of SSL 3.0 records.
2046 [Adam Langley (Google)]
2048 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2049 Kadianakis <desnacked@gmail.com> for discovering this issue and
2050 Adam Langley for preparing the fix. (CVE-2011-4619)
2051 [Adam Langley (Google)]
2053 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2054 [Andrey Kulikov <amdeich@gmail.com>]
2056 *) Prevent malformed RFC3779 data triggering an assertion failure.
2057 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2058 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2059 [Rob Austein <sra@hactrn.net>]
2061 *) Improved PRNG seeding for VOS.
2062 [Paul Green <Paul.Green@stratus.com>]
2064 *) Fix ssl_ciph.c set-up race.
2065 [Adam Langley (Google)]
2067 *) Fix spurious failures in ecdsatest.c.
2068 [Emilia Käsper (Google)]
2070 *) Fix the BIO_f_buffer() implementation (which was mixing different
2071 interpretations of the '..._len' fields).
2072 [Adam Langley (Google)]
2074 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2075 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2076 threads won't reuse the same blinding coefficients.
2078 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2079 lock to call BN_BLINDING_invert_ex, and avoids one use of
2080 BN_BLINDING_update for each BN_BLINDING structure (previously,
2081 the last update always remained unused).
2082 [Emilia Käsper (Google)]
2084 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2085 [Bob Buckholz (Google)]
2087 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2089 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2090 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2091 [Kaspar Brand <ossl@velox.ch>]
2093 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2094 for multi-threaded use of ECDH. (CVE-2011-3210)
2095 [Adam Langley (Google)]
2097 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2100 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2101 signature public key algorithm by using OID xref utilities instead.
2102 Before this you could only use some ECC ciphersuites with SHA1 only.
2105 *) Add protection against ECDSA timing attacks as mentioned in the paper
2106 by Billy Bob Brumley and Nicola Tuveri, see:
2108 http://eprint.iacr.org/2011/232.pdf
2110 [Billy Bob Brumley and Nicola Tuveri]
2112 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2114 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2115 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2117 *) Fix bug in string printing code: if *any* escaping is enabled we must
2118 escape the escape character (backslash) or the resulting string is
2122 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2124 *) Disable code workaround for ancient and obsolete Netscape browsers
2125 and servers: an attacker can use it in a ciphersuite downgrade attack.
2126 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2129 *) Fixed J-PAKE implementation error, originally discovered by
2130 Sebastien Martini, further info and confirmation from Stefan
2131 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2134 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2136 *) Fix extension code to avoid race conditions which can result in a buffer
2137 overrun vulnerability: resumed sessions must not be modified as they can
2138 be shared by multiple threads. CVE-2010-3864
2141 *) Fix WIN32 build system to correctly link an ENGINE directory into
2145 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2147 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2149 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2151 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2153 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2154 context. The operation can be customised via the ctrl mechanism in
2155 case ENGINEs want to include additional functionality.
2158 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2161 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2162 output hashes compatible with older versions of OpenSSL.
2163 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2165 *) Fix compression algorithm handling: if resuming a session use the
2166 compression algorithm of the resumed session instead of determining
2167 it from client hello again. Don't allow server to change algorithm.
2170 *) Add load_crls() function to apps tidying load_certs() too. Add option
2171 to verify utility to allow additional CRLs to be included.
2174 *) Update OCSP request code to permit adding custom headers to the request:
2175 some responders need this.
2178 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2180 [Julia Lawall <julia@diku.dk>]
2182 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2183 needlessly dereferenced structures, used obsolete functions and
2184 didn't handle all updated verify codes correctly.
2187 *) Disable MD2 in the default configuration.
2190 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2191 indicate the initial BIO being pushed or popped. This makes it possible
2192 to determine whether the BIO is the one explicitly called or as a result
2193 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2194 it handles reference counts correctly and doesn't zero out the I/O bio
2195 when it is not being explicitly popped. WARNING: applications which
2196 included workarounds for the old buggy behaviour will need to be modified
2197 or they could free up already freed BIOs.
2200 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2201 renaming to all platforms (within the 0.9.8 branch, this was
2202 done conditionally on Netware platforms to avoid a name clash).
2203 [Guenter <lists@gknw.net>]
2205 *) Add ECDHE and PSK support to DTLS.
2206 [Michael Tuexen <tuexen@fh-muenster.de>]
2208 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2212 *) Add "missing" function EVP_MD_flags() (without this the only way to
2213 retrieve a digest flags is by accessing the structure directly. Update
2214 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2215 or cipher is registered as in the "from" argument. Print out all
2216 registered digests in the dgst usage message instead of manually
2217 attempting to work them out.
2220 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2221 this allows the use of compression and extensions. Change default cipher
2222 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2223 by default unless an application cipher string requests it.
2226 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2227 key ids to find matching certificates and keys but some PKCS#12 files
2228 don't follow the (somewhat unwritten) rules and this strategy fails.
2229 Now just gather all certificates together and the first private key
2230 then look for the first certificate that matches the key.
2233 *) Support use of registered digest and cipher names for dgst and cipher
2234 commands instead of having to add each one as a special case. So now
2241 openssl dgst -sha256 foo
2243 and this works for ENGINE based algorithms too.
2247 *) Update Gost ENGINE to support parameter files.
2248 [Victor B. Wagner <vitus@cryptocom.ru>]
2250 *) Support GeneralizedTime in ca utility.
2251 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2253 *) Enhance the hash format used for certificate directory links. The new
2254 form uses the canonical encoding (meaning equivalent names will work
2255 even if they aren't identical) and uses SHA1 instead of MD5. This form
2256 is incompatible with the older format and as a result c_rehash should
2257 be used to rebuild symbolic links.
2260 *) Make PKCS#8 the default write format for private keys, replacing the
2261 traditional format. This form is standardised, more secure and doesn't
2262 include an implicit MD5 dependency.
2265 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2266 committed to OpenSSL should pass this lot as a minimum.
2269 *) Add session ticket override functionality for use by EAP-FAST.
2270 [Jouni Malinen <j@w1.fi>]
2272 *) Modify HMAC functions to return a value. Since these can be implemented
2273 in an ENGINE errors can occur.
2276 *) Type-checked OBJ_bsearch_ex.
2279 *) Type-checked OBJ_bsearch. Also some constification necessitated
2280 by type-checking. Still to come: TXT_DB, bsearch(?),
2281 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2285 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2286 seconds to a tm structure directly, instead of going through OS
2287 specific date routines. This avoids any issues with OS routines such
2288 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2289 and X509_time_adj_ex() to cover the extended range. The existing
2290 X509_time_adj() is still usable and will no longer have any date issues.
2293 *) Delta CRL support. New use deltas option which will attempt to locate
2294 and search any appropriate delta CRLs available.
2296 This work was sponsored by Google.
2299 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2300 code and add additional score elements. Validate alternate CRL paths
2301 as part of the CRL checking and indicate a new error "CRL path validation
2302 error" in this case. Applications wanting additional details can use
2303 the verify callback and check the new "parent" field. If this is not
2304 NULL CRL path validation is taking place. Existing applications wont
2305 see this because it requires extended CRL support which is off by
2308 This work was sponsored by Google.
2311 *) Support for freshest CRL extension.
2313 This work was sponsored by Google.
2316 *) Initial indirect CRL support. Currently only supported in the CRLs
2317 passed directly and not via lookup. Process certificate issuer
2318 CRL entry extension and lookup CRL entries by bother issuer name
2319 and serial number. Check and process CRL issuer entry in IDP extension.
2321 This work was sponsored by Google.
2324 *) Add support for distinct certificate and CRL paths. The CRL issuer
2325 certificate is validated separately in this case. Only enabled if
2326 an extended CRL support flag is set: this flag will enable additional
2327 CRL functionality in future.
2329 This work was sponsored by Google.
2332 *) Add support for policy mappings extension.
2334 This work was sponsored by Google.
2337 *) Fixes to pathlength constraint, self issued certificate handling,
2338 policy processing to align with RFC3280 and PKITS tests.
2340 This work was sponsored by Google.
2343 *) Support for name constraints certificate extension. DN, email, DNS
2344 and URI types are currently supported.
2346 This work was sponsored by Google.
2349 *) To cater for systems that provide a pointer-based thread ID rather
2350 than numeric, deprecate the current numeric thread ID mechanism and
2351 replace it with a structure and associated callback type. This
2352 mechanism allows a numeric "hash" to be extracted from a thread ID in
2353 either case, and on platforms where pointers are larger than 'long',
2354 mixing is done to help ensure the numeric 'hash' is usable even if it
2355 can't be guaranteed unique. The default mechanism is to use "&errno"
2356 as a pointer-based thread ID to distinguish between threads.
2358 Applications that want to provide their own thread IDs should now use
2359 CRYPTO_THREADID_set_callback() to register a callback that will call
2360 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2362 Note that ERR_remove_state() is now deprecated, because it is tied
2363 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2364 to free the current thread's error state should be replaced by
2365 ERR_remove_thread_state(NULL).
2367 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2368 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2369 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2370 application was previously providing a numeric thread callback that
2371 was inappropriate for distinguishing threads, then uniqueness might
2372 have been obtained with &errno that happened immediately in the
2373 intermediate development versions of OpenSSL; this is no longer the
2374 case, the numeric thread callback will now override the automatic use
2376 [Geoff Thorpe, with help from Bodo Moeller]
2378 *) Initial support for different CRL issuing certificates. This covers a
2379 simple case where the self issued certificates in the chain exist and
2380 the real CRL issuer is higher in the existing chain.
2382 This work was sponsored by Google.
2385 *) Removed effectively defunct crypto/store from the build.
2388 *) Revamp of STACK to provide stronger type-checking. Still to come:
2389 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2390 ASN1_STRING, CONF_VALUE.
2393 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2394 RAM on SSL connections. This option can save about 34k per idle SSL.
2397 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2398 STACK, TXT_DB, bsearch, qsort.
2401 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2402 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2403 support for data, signedData, compressedData, digestedData and
2404 encryptedData, envelopedData types included. Scripts to check against
2405 RFC4134 examples draft and interop and consistency checks of many
2406 content types and variants.
2409 *) Add options to enc utility to support use of zlib compression BIO.
2412 *) Extend mk1mf to support importing of options and assembly language
2413 files from Configure script, currently only included in VC-WIN32.
2414 The assembly language rules can now optionally generate the source
2415 files from the associated perl scripts.
2418 *) Implement remaining functionality needed to support GOST ciphersuites.
2419 Interop testing has been performed using CryptoPro implementations.
2420 [Victor B. Wagner <vitus@cryptocom.ru>]
2422 *) s390x assembler pack.
2425 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2429 *) Implement Opaque PRF Input TLS extension as specified in
2430 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2431 official specification yet and no extension type assignment by
2432 IANA exists, this extension (for now) will have to be explicitly
2433 enabled when building OpenSSL by providing the extension number
2434 to use. For example, specify an option
2436 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2438 to the "config" or "Configure" script to enable the extension,
2439 assuming extension number 0x9527 (which is a completely arbitrary
2440 and unofficial assignment based on the MD5 hash of the Internet
2441 Draft). Note that by doing so, you potentially lose
2442 interoperability with other TLS implementations since these might
2443 be using the same extension number for other purposes.
2445 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2446 opaque PRF input value to use in the handshake. This will create
2447 an interal copy of the length-'len' string at 'src', and will
2448 return non-zero for success.
2450 To get more control and flexibility, provide a callback function
2453 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2454 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2458 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2461 Callback function 'cb' will be called in handshakes, and is
2462 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2463 Argument 'arg' is for application purposes (the value as given to
2464 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2465 be provided to the callback function). The callback function
2466 has to return non-zero to report success: usually 1 to use opaque
2467 PRF input just if possible, or 2 to enforce use of the opaque PRF
2468 input. In the latter case, the library will abort the handshake
2469 if opaque PRF input is not successfully negotiated.
2471 Arguments 'peerinput' and 'len' given to the callback function
2472 will always be NULL and 0 in the case of a client. A server will
2473 see the client's opaque PRF input through these variables if
2474 available (NULL and 0 otherwise). Note that if the server
2475 provides an opaque PRF input, the length must be the same as the
2476 length of the client's opaque PRF input.
2478 Note that the callback function will only be called when creating
2479 a new session (session resumption can resume whatever was
2480 previously negotiated), and will not be called in SSL 2.0
2481 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2482 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2483 for applications that need to enforce opaque PRF input.
2487 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2490 [Victor B. Wagner <vitus@cryptocom.ru>]
2492 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2493 RFC4507bis. The encrypted ticket format is an encrypted encoded
2494 SSL_SESSION structure, that way new session features are automatically
2497 If a client application caches session in an SSL_SESSION structure
2498 support is transparent because tickets are now stored in the encoded
2501 The SSL_CTX structure automatically generates keys for ticket
2502 protection in servers so again support should be possible
2503 with no application modification.
2505 If a client or server wishes to disable RFC4507 support then the option
2506 SSL_OP_NO_TICKET can be set.
2508 Add a TLS extension debugging callback to allow the contents of any client
2509 or server extensions to be examined.
2511 This work was sponsored by Google.
2514 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2515 OpenSSL should now compile cleanly on gcc 4.2
2516 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2518 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2519 support including streaming MAC support: this is required for GOST
2520 ciphersuite support.
2521 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2523 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2524 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2525 to output in BER and PEM format.
2528 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2529 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2530 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2531 ENGINE support for HMAC keys which are unextractable. New -mac and
2532 -macopt options to dgst utility.
2535 *) New option -sigopt to dgst utility. Update dgst to use
2536 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2537 alternative signing paramaters such as X9.31 or PSS in the dgst
2541 *) Change ssl_cipher_apply_rule(), the internal function that does
2542 the work each time a ciphersuite string requests enabling
2543 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2544 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2545 the order of disabled ciphersuites such that those ciphersuites
2546 that most recently went from enabled to disabled not only stay
2547 in order with respect to each other, but also have higher priority
2548 than other disabled ciphersuites the next time ciphersuites are
2551 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2552 the same ciphersuites as with "HIGH" alone, but in a specific
2553 order where the PSK ciphersuites come first (since they are the
2554 most recently disabled ciphersuites when "HIGH" is parsed).
2556 Also, change ssl_create_cipher_list() (using this new
2557 funcionality) such that between otherwise identical
2558 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2562 *) Change ssl_create_cipher_list() so that it automatically
2563 arranges the ciphersuites in reasonable order before starting
2564 to process the rule string. Thus, the definition for "DEFAULT"
2565 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2566 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2567 This makes it much easier to arrive at a reasonable default order
2568 in applications for which anonymous ciphers are OK (meaning
2569 that you can't actually use DEFAULT).
2570 [Bodo Moeller; suggested by Victor Duchovni]
2572 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2573 processing) into multiple integers instead of setting
2574 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2575 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2576 (These masks as well as the individual bit definitions are hidden
2577 away into the non-exported interface ssl/ssl_locl.h, so this
2578 change to the definition of the SSL_CIPHER structure shouldn't
2579 affect applications.) This give us more bits for each of these
2580 categories, so there is no longer a need to coagulate AES128 and
2581 AES256 into a single algorithm bit, and to coagulate Camellia128
2582 and Camellia256 into a single algorithm bit, which has led to all
2585 Thus, among other things, the kludge introduced in 0.9.7m and
2586 0.9.8e for masking out AES256 independently of AES128 or masking
2587 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2589 With the change, we also introduce new ciphersuite aliases that
2590 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2594 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2595 Use the leftmost N bytes of the signature input if the input is
2596 larger than the prime q (with N being the size in bytes of q).
2599 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2600 it yet and it is largely untested.
2603 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2606 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2607 some compilers (gcc 4.2 and later) reject their use. Safestack is
2608 reimplemented. Update ASN1 to avoid use of legacy functions.
2611 *) Win32/64 targets are linked with Winsock2.
2614 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2615 to external functions. This can be used to increase CRL handling
2616 efficiency especially when CRLs are very large by (for example) storing
2617 the CRL revoked certificates in a database.
2620 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2621 new CRLs added to a directory can be used. New command line option
2622 -verify_return_error to s_client and s_server. This causes real errors
2623 to be returned by the verify callback instead of carrying on no matter
2624 what. This reflects the way a "real world" verify callback would behave.
2627 *) GOST engine, supporting several GOST algorithms and public key formats.
2628 Kindly donated by Cryptocom.
2631 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2632 partitioned by DP are handled but no indirect CRL or reason partitioning
2633 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2634 selected via a scoring technique which handles IDP and AKID in CRLs.
2637 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2638 will ultimately be used for all verify operations: this will remove the
2639 X509_STORE dependency on certificate verification and allow alternative
2640 lookup methods. X509_STORE based implementations of these two callbacks.
2643 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2644 Modify get_crl() to find a valid (unexpired) CRL if possible.
2647 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2648 this would be called X509_CRL_cmp() but that name is already used by
2649 a function that just compares CRL issuer names. Cache several CRL
2650 extensions in X509_CRL structure and cache CRLDP in X509.
2653 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2654 this maps equivalent X509_NAME structures into a consistent structure.
2655 Name comparison can then be performed rapidly using memcmp().
2658 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2662 *) Allow digests to supply their own micalg string for S/MIME type using
2663 the ctrl EVP_MD_CTRL_MICALG.
2666 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2667 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2668 ctrl. It can then customise the structure before and/or after signing
2672 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2673 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2674 to free up any added signature OIDs.
2677 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2678 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2679 digest and cipher tables. New options added to openssl utility:
2680 list-message-digest-algorithms and list-cipher-algorithms.
2683 *) Change the array representation of binary polynomials: the list
2684 of degrees of non-zero coefficients is now terminated with -1.
2685 Previously it was terminated with 0, which was also part of the
2686 value; thus, the array representation was not applicable to
2687 polynomials where t^0 has coefficient zero. This change makes
2688 the array representation useful in a more general context.
2691 *) Various modifications and fixes to SSL/TLS cipher string
2692 handling. For ECC, the code now distinguishes between fixed ECDH
2693 with RSA certificates on the one hand and with ECDSA certificates
2694 on the other hand, since these are separate ciphersuites. The
2695 unused code for Fortezza ciphersuites has been removed.
2697 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2698 (not "ECDHE"). For consistency with the code for DH
2699 certificates, use of ECDH certificates is now considered ECDH
2700 authentication, not RSA or ECDSA authentication (the latter is
2701 merely the CA's signing algorithm and not actively used in the
2704 The temporary ciphersuite alias "ECCdraft" is no longer
2705 available, and ECC ciphersuites are no longer excluded from "ALL"
2706 and "DEFAULT". The following aliases now exist for RFC 4492
2707 ciphersuites, most of these by analogy with the DH case:
2709 kECDHr - ECDH cert, signed with RSA
2710 kECDHe - ECDH cert, signed with ECDSA
2711 kECDH - ECDH cert (signed with either RSA or ECDSA)
2712 kEECDH - ephemeral ECDH
2713 ECDH - ECDH cert or ephemeral ECDH
2719 AECDH - anonymous ECDH
2720 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2724 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2725 Use correct micalg parameters depending on digest(s) in signed message.
2728 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2729 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2732 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2733 an engine to register a method. Add ENGINE lookups for methods and
2734 functional reference processing.
2737 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2738 EVP_{Sign,Verify}* which allow an application to customise the signature
2742 *) New -resign option to smime utility. This adds one or more signers
2743 to an existing PKCS#7 signedData structure. Also -md option to use an
2744 alternative message digest algorithm for signing.
2747 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2748 create PKCS7 structures containing multiple signers. Update smime
2749 application to support multiple signers.
2752 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2756 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2757 Reorganize PBE internals to lookup from a static table using NIDs,
2758 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2759 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2760 PRF which will be automatically used with PBES2.
2763 *) Replace the algorithm specific calls to generate keys in "req" with the
2767 *) Update PKCS#7 enveloped data routines to use new API. This is now
2768 supported by any public key method supporting the encrypt operation. A
2769 ctrl is added to allow the public key algorithm to examine or modify
2770 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2774 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2775 a default digest type to use. In most cases this will be SHA1 but some
2776 algorithms (such as GOST) need to specify an alternative digest. The
2777 return value indicates how strong the prefernce is 1 means optional and
2778 2 is mandatory (that is it is the only supported type). Modify
2779 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2780 use the default md. Update openssl utilities to use the default digest
2781 type for signing if it is not explicitly indicated.
2784 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2785 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2786 signing method from the key type. This effectively removes the link
2787 between digests and public key types.
2790 *) Add an OID cross reference table and utility functions. Its purpose is to
2791 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2792 rsaEncryption. This will allow some of the algorithm specific hackery
2793 needed to use the correct OID to be removed.
2796 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2797 structures for PKCS7_sign(). They are now set up by the relevant public
2801 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2804 *) Add support for key derivation (agreement) in the API, DH method and
2808 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2809 public and private key formats. As a side effect these add additional
2810 command line functionality not previously available: DSA signatures can be
2811 generated and verified using pkeyutl and DH key support and generation in
2816 [Oliver Tappe <zooey@hirschkaefer.de>]
2818 *) New make target "install_html_docs" installs HTML renditions of the
2820 [Oliver Tappe <zooey@hirschkaefer.de>]
2822 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
2823 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2824 support key and parameter generation and add initial key generation
2825 functionality for RSA.
2828 *) Add functions for main EVP_PKEY_method operations. The undocumented
2829 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2830 EVP_PKEY_{encrypt,decrypt}_old.
2833 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2834 key API, doesn't do much yet.
2837 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2838 public key algorithms. New option to openssl utility:
2839 "list-public-key-algorithms" to print out info.
2842 *) Implement the Supported Elliptic Curves Extension for
2843 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2846 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2847 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2850 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2851 utilities such as rsa, dsa, dsaparam etc except they process any key
2855 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2856 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2857 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2861 *) Initial support for pluggable public key ASN1.
2862 De-spaghettify the public key ASN1 handling. Move public and private
2863 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2864 algorithm specific handling to a single module within the relevant
2865 algorithm directory. Add functions to allow (near) opaque processing
2866 of public and private key structures.
2869 *) Implement the Supported Point Formats Extension for
2870 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2873 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2874 for the psk identity [hint] and the psk callback functions to the
2875 SSL_SESSION, SSL and SSL_CTX structure.
2878 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2882 SSL_CTX_use_psk_identity_hint
2883 SSL_get_psk_identity_hint
2884 SSL_get_psk_identity
2885 SSL_use_psk_identity_hint
2887 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2889 *) Add RFC 3161 compliant time stamp request creation, response generation
2890 and response verification functionality.
2891 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2893 *) Add initial support for TLS extensions, specifically for the server_name
2894 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2895 have new members for a host name. The SSL data structure has an
2896 additional member SSL_CTX *initial_ctx so that new sessions can be
2897 stored in that context to allow for session resumption, even after the
2898 SSL has been switched to a new SSL_CTX in reaction to a client's
2899 server_name extension.
2901 New functions (subject to change):
2903 SSL_get_servername()
2904 SSL_get_servername_type()
2907 New CTRL codes and macros (subject to change):
2909 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2910 - SSL_CTX_set_tlsext_servername_callback()
2911 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2912 - SSL_CTX_set_tlsext_servername_arg()
2913 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2915 openssl s_client has a new '-servername ...' option.
2917 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2918 '-key2 ...', '-servername_fatal' (subject to change). This allows
2919 testing the HostName extension for a specific single host name ('-cert'
2920 and '-key' remain fallbacks for handshakes without HostName
2921 negotiation). If the unrecogninzed_name alert has to be sent, this by
2922 default is a warning; it becomes fatal with the '-servername_fatal'
2925 [Peter Sylvester, Remy Allais, Christophe Renou]
2927 *) Whirlpool hash implementation is added.
2930 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2931 bn(64,32). Because of instruction set limitations it doesn't have
2932 any negative impact on performance. This was done mostly in order
2933 to make it possible to share assembler modules, such as bn_mul_mont
2934 implementations, between 32- and 64-bit builds without hassle.
2937 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2938 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2942 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2943 dedicated Montgomery multiplication procedure, is introduced.
2944 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2945 "64-bit" performance on certain 32-bit targets.
2948 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2949 in SSL structures. New SSL ctrl to set maximum send fragment size.
2950 Save memory by seeting the I/O buffer sizes dynamically instead of
2951 using the maximum available value.
2954 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2955 in addition to the text details.
2958 *) Very, very preliminary EXPERIMENTAL support for printing of general
2959 ASN1 structures. This currently produces rather ugly output and doesn't
2960 handle several customised structures at all.
2963 *) Integrated support for PVK file format and some related formats such
2964 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2965 these in the 'rsa' and 'dsa' utilities.
2968 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2971 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2972 place for the (very old) "NETSCAPE" format certificates which are now
2973 handled using new ASN1 code equivalents.
2976 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2977 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2978 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2981 *) Modify CRL distribution points extension code to print out previously
2982 unsupported fields. Enhance extension setting code to allow setting of
2986 *) Add print and set support for Issuing Distribution Point CRL extension.
2989 *) Change 'Configure' script to enable Camellia by default.
2992 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2994 *) When rejecting SSL/TLS records due to an incorrect version number, never
2995 update s->server with a new major version number. As of
2996 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2997 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2998 the previous behavior could result in a read attempt at NULL when
2999 receiving specific incorrect SSL/TLS records once record payload
3000 protection is active. (CVE-2010-0740)
3001 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3003 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3004 could be crashed if the relevant tables were not present (e.g. chrooted).
3005 [Tomas Hoger <thoger@redhat.com>]
3007 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3009 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3010 [Martin Olsson, Neel Mehta]
3012 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3013 accommodate for stack sorting, always a write lock!).
3016 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3017 excessive delays in the RAND_poll(): over a minute. As a workaround
3018 include a time check in the inner Heap32Next loop too.
3021 *) The code that handled flushing of data in SSL/TLS originally used the
3022 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3023 the problem outlined in PR#1949. The fix suggested there however can
3024 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3025 of Apache). So instead simplify the code to flush unconditionally.
3026 This should be fine since flushing with no data to flush is a no op.
3029 *) Handle TLS versions 2.0 and later properly and correctly use the
3030 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3031 off ancient servers have a habit of sticking around for a while...
3034 *) Modify compression code so it frees up structures without using the
3035 ex_data callbacks. This works around a problem where some applications
3036 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3037 restarting) then use compression (e.g. SSL with compression) later.
3038 This results in significant per-connection memory leaks and
3039 has caused some security issues including CVE-2008-1678 and
3043 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3044 change when encrypting or decrypting.
3047 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3048 connect and renegotiate with servers which do not support RI.
3049 Until RI is more widely deployed this option is enabled by default.
3052 *) Add "missing" ssl ctrls to clear options and mode.
3055 *) If client attempts to renegotiate and doesn't support RI respond with
3056 a no_renegotiation alert as required by RFC5746. Some renegotiating
3057 TLS clients will continue a connection gracefully when they receive
3058 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3059 waiting for a server hello which it will never receive. Now we treat a
3060 received no_renegotiation alert as a fatal error. This is because
3061 applications requesting a renegotiation might well expect it to succeed
3062 and would have no code in place to handle the server denying it so the
3063 only safe thing to do is to terminate the connection.
3066 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3067 peer supports secure renegotiation and 0 otherwise. Print out peer
3068 renegotiation support in s_client/s_server.
3071 *) Replace the highly broken and deprecated SPKAC certification method with
3072 the updated NID creation version. This should correctly handle UTF8.
3075 *) Implement RFC5746. Re-enable renegotiation but require the extension
3076 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3077 turns out to be a bad idea. It has been replaced by
3078 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3079 SSL_CTX_set_options(). This is really not recommended unless you
3080 know what you are doing.
3081 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3083 *) Fixes to stateless session resumption handling. Use initial_ctx when
3084 issuing and attempting to decrypt tickets in case it has changed during
3085 servername handling. Use a non-zero length session ID when attempting
3086 stateless session resumption: this makes it possible to determine if
3087 a resumption has occurred immediately after receiving server hello
3088 (several places in OpenSSL subtly assume this) instead of later in
3092 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3093 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3094 fixes for a few places where the return code is not checked
3096 [Julia Lawall <julia@diku.dk>]
3098 *) Add --strict-warnings option to Configure script to include devteam
3099 warnings in other configurations.
3102 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3103 makes it possible to install openssl libraries in locations which
3104 have names other than "lib", for example "/usr/lib64" which some
3106 [Steve Henson, based on patch from Jeremy Utley]
3108 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3109 X690 8.9.12 and can produce some misleading textual output of OIDs.
3110 [Steve Henson, reported by Dan Kaminsky]
3112 *) Delete MD2 from algorithm tables. This follows the recommendation in
3113 several standards that it is not used in new applications due to
3114 several cryptographic weaknesses. For binary compatibility reasons
3115 the MD2 API is still compiled in by default.
3118 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3122 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3123 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3125 [Guenter <lists@gknw.net>]
3127 *) Fix the server certificate chain building code to use X509_verify_cert(),
3128 it used to have an ad-hoc builder which was unable to cope with anything
3129 other than a simple chain.
3130 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3132 *) Don't check self signed certificate signatures in X509_verify_cert()
3133 by default (a flag can override this): it just wastes time without
3134 adding any security. As a useful side effect self signed root CAs
3135 with non-FIPS digests are now usable in FIPS mode.
3138 *) In dtls1_process_out_of_seq_message() the check if the current message
3139 is already buffered was missing. For every new message was memory
3140 allocated, allowing an attacker to perform an denial of service attack
3141 with sending out of seq handshake messages until there is no memory
3142 left. Additionally every future messege was buffered, even if the
3143 sequence number made no sense and would be part of another handshake.
3144 So only messages with sequence numbers less than 10 in advance will be
3145 buffered. (CVE-2009-1378)
3146 [Robin Seggelmann, discovered by Daniel Mentz]
3148 *) Records are buffered if they arrive with a future epoch to be
3149 processed after finishing the corresponding handshake. There is
3150 currently no limitation to this buffer allowing an attacker to perform
3151 a DOS attack with sending records with future epochs until there is no
3152 memory left. This patch adds the pqueue_size() function to detemine
3153 the size of a buffer and limits the record buffer to 100 entries.
3155 [Robin Seggelmann, discovered by Daniel Mentz]
3157 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3158 parent structure is freed. (CVE-2009-1379)
3161 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3162 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3165 [Ilya O. <vrghost@gmail.com>]
3167 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3169 *) Disable renegotiation completely - this fixes a severe security
3170 problem (CVE-2009-3555) at the cost of breaking all
3171 renegotiation. Renegotiation can be re-enabled by setting
3172 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3173 run-time. This is really not recommended unless you know what
3177 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3179 *) Don't set val to NULL when freeing up structures, it is freed up by
3180 underlying code. If sizeof(void *) > sizeof(long) this can result in
3181 zeroing past the valid field. (CVE-2009-0789)
3182 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3184 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3185 checked correctly. This would allow some invalid signed attributes to
3186 appear to verify correctly. (CVE-2009-0591)
3187 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3189 *) Reject UniversalString and BMPString types with invalid lengths. This
3190 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3191 a legal length. (CVE-2009-0590)
3194 *) Set S/MIME signing as the default purpose rather than setting it
3195 unconditionally. This allows applications to override it at the store
3199 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3200 to handle some structures.
3203 *) Improve efficiency of mem_gets: don't search whole buffer each time
3205 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3207 *) New -hex option for openssl rand.
3210 *) Print out UTF8String and NumericString when parsing ASN1.
3213 *) Support NumericString type for name components.
3216 *) Allow CC in the environment to override the automatically chosen
3217 compiler. Note that nothing is done to ensure flags work with the
3221 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3223 *) Properly check EVP_VerifyFinal() and similar return values
3225 [Ben Laurie, Bodo Moeller, Google Security Team]
3227 *) Enable TLS extensions by default.
3230 *) Allow the CHIL engine to be loaded, whether the application is
3231 multithreaded or not. (This does not release the developer from the
3232 obligation to set up the dynamic locking callbacks.)
3233 [Sander Temme <sander@temme.net>]
3235 *) Use correct exit code if there is an error in dgst command.
3236 [Steve Henson; problem pointed out by Roland Dirlewanger]
3238 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3239 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3242 *) Add experimental JPAKE support, including demo authentication in
3243 s_client and s_server.
3246 *) Set the comparison function in v3_addr_canonize().
3247 [Rob Austein <sra@hactrn.net>]
3249 *) Add support for XMPP STARTTLS in s_client.
3250 [Philip Paeps <philip@freebsd.org>]
3252 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3253 to ensure that even with this option, only ciphersuites in the
3254 server's preference list will be accepted. (Note that the option
3255 applies only when resuming a session, so the earlier behavior was
3256 just about the algorithm choice for symmetric cryptography.)
3259 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3261 *) Fix NULL pointer dereference if a DTLS server received
3262 ChangeCipherSpec as first record (CVE-2009-1386).
3265 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
3266 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3269 *) The fix in 0.9.8c that supposedly got rid of unsafe
3270 double-checked locking was incomplete for RSA blinding,
3271 addressing just one layer of what turns out to have been
3272 doubly unsafe triple-checked locking.
3274 So now fix this for real by retiring the MONT_HELPER macro
3275 in crypto/rsa/rsa_eay.c.
3277 [Bodo Moeller; problem pointed out by Marius Schilder]
3279 *) Various precautionary measures:
3281 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3283 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3284 (NB: This would require knowledge of the secret session ticket key
3285 to exploit, in which case you'd be SOL either way.)
3287 - Change bn_nist.c so that it will properly handle input BIGNUMs
3288 outside the expected range.
3290 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3293 [Neel Mehta, Bodo Moeller]
3295 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3296 the load fails. Useful for distros.
3297 [Ben Laurie and the FreeBSD team]
3299 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3302 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3305 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3307 This work was sponsored by Logica.
3310 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3311 keystores. Support for SSL/TLS client authentication too.
3312 Not compiled unless enable-capieng specified to Configure.
3314 This work was sponsored by Logica.
3317 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3318 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3319 attribute creation routines such as certifcate requests and PKCS#12
3323 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3325 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3326 handshake which could lead to a cilent crash as found using the
3327 Codenomicon TLS test suite (CVE-2008-1672)
3328 [Steve Henson, Mark Cox]
3330 *) Fix double free in TLS server name extensions which could lead to
3331 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3334 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3336 Clear the error queue to ensure that error entries left from
3337 older function calls do not interfere with the correct operation.
3338 [Lutz Jaenicke, Erik de Castro Lopo]
3340 *) Remove root CA certificates of commercial CAs:
3342 The OpenSSL project does not recommend any specific CA and does not
3343 have any policy with respect to including or excluding any CA.
3344 Therefore it does not make any sense to ship an arbitrary selection
3345 of root CA certificates with the OpenSSL software.
3348 *) RSA OAEP patches to fix two separate invalid memory reads.
3349 The first one involves inputs when 'lzero' is greater than
3350 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3351 before the beginning of from). The second one involves inputs where
3352 the 'db' section contains nothing but zeroes (there is a one-byte
3353 invalid read after the end of 'db').
3354 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3356 *) Partial backport from 0.9.9-dev:
3358 Introduce bn_mul_mont (dedicated Montgomery multiplication
3359 procedure) as a candidate for BIGNUM assembler implementation.
3360 While 0.9.9-dev uses assembler for various architectures, only
3361 x86_64 is available by default here in the 0.9.8 branch, and
3362 32-bit x86 is available through a compile-time setting.
3364 To try the 32-bit x86 assembler implementation, use Configure
3365 option "enable-montasm" (which exists only for this backport).
3367 As "enable-montasm" for 32-bit x86 disclaims code stability
3368 anyway, in this constellation we activate additional code
3369 backported from 0.9.9-dev for further performance improvements,
3370 namely BN_from_montgomery_word. (To enable this otherwise,
3371 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3373 [Andy Polyakov (backport partially by Bodo Moeller)]
3375 *) Add TLS session ticket callback. This allows an application to set
3376 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3377 values. This is useful for key rollover for example where several key
3378 sets may exist with different names.
3381 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3382 This was broken until now in 0.9.8 releases, such that the only way
3383 a registered ENGINE could be used (assuming it initialises
3384 successfully on the host) was to explicitly set it as the default
3385 for the relevant algorithms. This is in contradiction with 0.9.7
3386 behaviour and the documentation. With this fix, when an ENGINE is
3387 registered into a given algorithm's table of implementations, the
3388 'uptodate' flag is reset so that auto-discovery will be used next
3389 time a new context for that algorithm attempts to select an
3391 [Ian Lister (tweaked by Geoff Thorpe)]
3393 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3394 implemention in the following ways:
3396 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3399 Lack of BER streaming support means one pass streaming processing is
3400 only supported if data is detached: setting the streaming flag is
3401 ignored for embedded content.
3403 CMS support is disabled by default and must be explicitly enabled
3404 with the enable-cms configuration option.
3407 *) Update the GMP engine glue to do direct copies between BIGNUM and
3408 mpz_t when openssl and GMP use the same limb size. Otherwise the
3409 existing "conversion via a text string export" trick is still used.
3410 [Paul Sheer <paulsheer@gmail.com>]
3412 *) Zlib compression BIO. This is a filter BIO which compressed and
3413 uncompresses any data passed through it.
3416 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3417 RFC3394 compatible AES key wrapping.
3420 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3421 sets string data without copying. X509_ALGOR_set0() and
3422 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3423 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3424 from an X509_ATTRIBUTE structure optionally checking it occurs only
3425 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3429 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3430 to get the expected BN_FLG_CONSTTIME behavior.
3431 [Bodo Moeller (Google)]
3435 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3436 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3437 - added some more tests to do_tests.pl
3438 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3439 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3440 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3441 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3442 - various changes to netware.pl to enable gcc-cross builds on Win32
3444 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3445 - various changes to fix missing prototype warnings
3446 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3447 - added AES, WHIRLPOOL and CPUID assembler code to build files
3448 - added missing AES assembler make rules to mk1mf.pl
3449 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3450 [Guenter Knauf <eflash@gmx.net>]
3452 *) Implement certificate status request TLS extension defined in RFC3546.
3453 A client can set the appropriate parameters and receive the encoded
3454 OCSP response via a callback. A server can query the supplied parameters
3455 and set the encoded OCSP response in the callback. Add simplified examples
3456 to s_client and s_server.
3459 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3461 *) Fix various bugs:
3462 + Binary incompatibility of ssl_ctx_st structure
3463 + DTLS interoperation with non-compliant servers
3464 + Don't call get_session_cb() without proposed session
3465 + Fix ia64 assembler code
3466 [Andy Polyakov, Steve Henson]
3468 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3470 *) DTLS Handshake overhaul. There were longstanding issues with
3471 OpenSSL DTLS implementation, which were making it impossible for
3472 RFC 4347 compliant client to communicate with OpenSSL server.
3473 Unfortunately just fixing these incompatibilities would "cut off"
3474 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3475 server keeps tolerating non RFC compliant syntax. The opposite is
3476 not true, 0.9.8f client can not communicate with earlier server.
3477 This update even addresses CVE-2007-4995.
3480 *) Changes to avoid need for function casts in OpenSSL: some compilers
3481 (gcc 4.2 and later) reject their use.
3482 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3485 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3486 RFC4507bis. The encrypted ticket format is an encrypted encoded
3487 SSL_SESSION structure, that way new session features are automatically
3490 If a client application caches session in an SSL_SESSION structure
3491 support is transparent because tickets are now stored in the encoded
3494 The SSL_CTX structure automatically generates keys for ticket
3495 protection in servers so again support should be possible
3496 with no application modification.
3498 If a client or server wishes to disable RFC4507 support then the option
3499 SSL_OP_NO_TICKET can be set.
3501 Add a TLS extension debugging callback to allow the contents of any client
3502 or server extensions to be examined.
3504 This work was sponsored by Google.
3507 *) Add initial support for TLS extensions, specifically for the server_name
3508 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3509 have new members for a host name. The SSL data structure has an
3510 additional member SSL_CTX *initial_ctx so that new sessions can be
3511 stored in that context to allow for session resumption, even after the
3512 SSL has been switched to a new SSL_CTX in reaction to a client's
3513 server_name extension.
3515 New functions (subject to change):
3517 SSL_get_servername()
3518 SSL_get_servername_type()
3521 New CTRL codes and macros (subject to change):
3523 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3524 - SSL_CTX_set_tlsext_servername_callback()
3525 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3526 - SSL_CTX_set_tlsext_servername_arg()
3527 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3529 openssl s_client has a new '-servername ...' option.
3531 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3532 '-key2 ...', '-servername_fatal' (subject to change). This allows
3533 testing the HostName extension for a specific single host name ('-cert'
3534 and '-key' remain fallbacks for handshakes without HostName
3535 negotiation). If the unrecogninzed_name alert has to be sent, this by
3536 default is a warning; it becomes fatal with the '-servername_fatal'
3539 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3541 *) Add AES and SSE2 assembly language support to VC++ build.
3544 *) Mitigate attack on final subtraction in Montgomery reduction.
3547 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3548 (which previously caused an internal error).
3551 *) Squeeze another 10% out of IGE mode when in != out.
3554 *) AES IGE mode speedup.
3555 [Dean Gaudet (Google)]
3557 *) Add the Korean symmetric 128-bit cipher SEED (see
3558 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3559 add SEED ciphersuites from RFC 4162:
3561 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3562 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3563 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3564 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3566 To minimize changes between patchlevels in the OpenSSL 0.9.8
3567 series, SEED remains excluded from compilation unless OpenSSL
3568 is configured with 'enable-seed'.
3569 [KISA, Bodo Moeller]
3571 *) Mitigate branch prediction attacks, which can be practical if a
3572 single processor is shared, allowing a spy process to extract
3573 information. For detailed background information, see
3574 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3575 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3576 and Necessary Software Countermeasures"). The core of the change
3577 are new versions BN_div_no_branch() and
3578 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3579 respectively, which are slower, but avoid the security-relevant
3580 conditional branches. These are automatically called by BN_div()
3581 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3582 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3583 remove a conditional branch.
3585 BN_FLG_CONSTTIME is the new name for the previous
3586 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3587 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3588 in the exponent causes BN_mod_exp_mont() to use the alternative
3589 implementation in BN_mod_exp_mont_consttime().) The old name
3590 remains as a deprecated alias.
3592 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3593 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3594 constant-time implementations for more than just exponentiation.
3595 Here too the old name is kept as a deprecated alias.
3597 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3598 the BN_BLINDING structure gets an independent copy of the
3599 modulus. This means that the previous "BIGNUM *m" argument to
3600 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3601 essentially becomes "const BIGNUM *m", although we can't actually
3602 change this in the header file before 0.9.9. It allows
3603 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3604 enable BN_FLG_CONSTTIME.
3606 [Matthew D Wood (Intel Corp)]
3608 *) In the SSL/TLS server implementation, be strict about session ID
3609 context matching (which matters if an application uses a single
3610 external cache for different purposes). Previously,
3611 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3612 set. This did ensure strict client verification, but meant that,
3613 with applications using a single external cache for quite
3614 different requirements, clients could circumvent ciphersuite
3615 restrictions for a given session ID context by starting a session
3616 in a different context.
3619 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3620 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3621 authentication-only ciphersuites.
3624 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3625 not complete and could lead to a possible single byte overflow
3626 (CVE-2007-5135) [Ben Laurie]
3628 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3630 *) Since AES128 and AES256 (and similarly Camellia128 and
3631 Camellia256) share a single mask bit in the logic of
3632 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3633 kludge to work properly if AES128 is available and AES256 isn't
3634 (or if Camellia128 is available and Camellia256 isn't).
3637 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3638 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3639 When a point or a seed is encoded in a BIT STRING, we need to
3640 prevent the removal of trailing zero bits to get the proper DER
3641 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3642 of a NamedBitList, for which trailing 0 bits need to be removed.)
3645 *) Have SSL/TLS server implementation tolerate "mismatched" record
3646 protocol version while receiving ClientHello even if the
3647 ClientHello is fragmented. (The server can't insist on the
3648 particular protocol version it has chosen before the ServerHello
3649 message has informed the client about his choice.)
3652 *) Add RFC 3779 support.
3653 [Rob Austein for ARIN, Ben Laurie]
3655 *) Load error codes if they are not already present instead of using a
3656 static variable. This allows them to be cleanly unloaded and reloaded.
3657 Improve header file function name parsing.
3660 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3661 or CAPABILITY handshake as required by RFCs.
3664 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3666 *) Introduce limits to prevent malicious keys being able to
3667 cause a denial of service. (CVE-2006-2940)
3668 [Steve Henson, Bodo Moeller]
3670 *) Fix ASN.1 parsing of certain invalid structures that can result
3671 in a denial of service. (CVE-2006-2937) [Steve Henson]
3673 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3674 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3676 *) Fix SSL client code which could crash if connecting to a
3677 malicious SSLv2 server. (CVE-2006-4343)
3678 [Tavis Ormandy and Will Drewry, Google Security Team]
3680 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3681 match only those. Before that, "AES256-SHA" would be interpreted
3682 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3683 the same strength classification in 0.9.7h) as we currently only
3684 have a single AES bit in the ciphersuite description bitmap.
3685 That change, however, also applied to ciphersuite strings such as
3686 "RC4-MD5" that intentionally matched multiple ciphersuites --
3687 namely, SSL 2.0 ciphersuites in addition to the more common ones
3688 from SSL 3.0/TLS 1.0.
3690 So we change the selection algorithm again: Naming an explicit
3691 ciphersuite selects this one ciphersuite, and any other similar
3692 ciphersuite (same bitmap) from *other* protocol versions.
3693 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3694 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3696 Since SSL 2.0 does not have any ciphersuites for which the
3697 128/256 bit distinction would be relevant, this works for now.
3698 The proper fix will be to use different bits for AES128 and
3699 AES256, which would have avoided the problems from the beginning;
3700 however, bits are scarce, so we can only do this in a new release
3701 (not just a patchlevel) when we can change the SSL_CIPHER
3702 definition to split the single 'unsigned long mask' bitmap into
3703 multiple values to extend the available space.
3707 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3709 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3710 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3712 *) Add AES IGE and biIGE modes.
3715 *) Change the Unix randomness entropy gathering to use poll() when
3716 possible instead of select(), since the latter has some
3717 undesirable limitations.
3718 [Darryl Miles via Richard Levitte and Bodo Moeller]
3720 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3721 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3722 cannot be implicitly activated as part of, e.g., the "AES" alias.
3723 However, please upgrade to OpenSSL 0.9.9[-dev] for
3724 non-experimental use of the ECC ciphersuites to get TLS extension
3725 support, which is required for curve and point format negotiation
3726 to avoid potential handshake problems.
3729 *) Disable rogue ciphersuites:
3731 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3732 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3733 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3735 The latter two were purportedly from
3736 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3739 Also deactivate the remaining ciphersuites from
3740 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3741 unofficial, and the ID has long expired.
3744 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3745 dual-core machines) and other potential thread-safety issues.
3748 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3749 versions), which is now available for royalty-free use
3750 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3751 Also, add Camellia TLS ciphersuites from RFC 4132.
3753 To minimize changes between patchlevels in the OpenSSL 0.9.8
3754 series, Camellia remains excluded from compilation unless OpenSSL
3755 is configured with 'enable-camellia'.
3758 *) Disable the padding bug check when compression is in use. The padding
3759 bug check assumes the first packet is of even length, this is not
3760 necessarily true if compresssion is enabled and can result in false
3761 positives causing handshake failure. The actual bug test is ancient
3762 code so it is hoped that implementations will either have fixed it by
3763 now or any which still have the bug do not support compression.
3766 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3768 *) When applying a cipher rule check to see if string match is an explicit
3769 cipher suite and only match that one cipher suite if it is.
3772 *) Link in manifests for VC++ if needed.
3773 [Austin Ziegler <halostatue@gmail.com>]
3775 *) Update support for ECC-based TLS ciphersuites according to
3776 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3777 TLS extensions, which are supported starting with the 0.9.9
3778 branch, not in the OpenSSL 0.9.8 branch).
3781 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3782 opaque EVP_CIPHER_CTX handling.
3785 *) Fixes and enhancements to zlib compression code. We now only use
3786 "zlib1.dll" and use the default __cdecl calling convention on Win32
3787 to conform with the standards mentioned here:
3788 http://www.zlib.net/DLL_FAQ.txt
3789 Static zlib linking now works on Windows and the new --with-zlib-include
3790 --with-zlib-lib options to Configure can be used to supply the location
3791 of the headers and library. Gracefully handle case where zlib library
3795 *) Several fixes and enhancements to the OID generation code. The old code
3796 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3797 handle numbers larger than ULONG_MAX, truncated printing and had a
3798 non standard OBJ_obj2txt() behaviour.
3801 *) Add support for building of engines under engine/ as shared libraries
3802 under VC++ build system.
3805 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3806 Hopefully, we will not see any false combination of paths any more.
3809 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3811 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3812 (part of SSL_OP_ALL). This option used to disable the
3813 countermeasure against man-in-the-middle protocol-version
3814 rollback in the SSL 2.0 server implementation, which is a bad
3815 idea. (CVE-2005-2969)
3817 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3818 for Information Security, National Institute of Advanced Industrial
3819 Science and Technology [AIST], Japan)]
3821 *) Add two function to clear and return the verify parameter flags.
3824 *) Keep cipherlists sorted in the source instead of sorting them at
3825 runtime, thus removing the need for a lock.
3828 *) Avoid some small subgroup attacks in Diffie-Hellman.
3829 [Nick Mathewson and Ben Laurie]
3831 *) Add functions for well-known primes.
3834 *) Extended Windows CE support.
3835 [Satoshi Nakamura and Andy Polyakov]
3837 *) Initialize SSL_METHOD structures at compile time instead of during
3838 runtime, thus removing the need for a lock.
3841 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3842 attempting to decrypt each encrypted key in turn. Add support to
3846 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3848 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3851 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3854 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3855 key into the same file any more.
3858 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3861 *) Add -utf8 command line and config file option to 'ca'.
3862 [Stefan <stf@udoma.org]
3864 *) Removed the macro des_crypt(), as it seems to conflict with some
3865 libraries. Use DES_crypt().
3868 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3869 involves renaming the source and generated shared-libs for
3870 both. The engines will accept the corrected or legacy ids
3871 ('ncipher' and '4758_cca' respectively) when binding. NB,
3872 this only applies when building 'shared'.
3873 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3875 *) Add attribute functions to EVP_PKEY structure. Modify
3876 PKCS12_create() to recognize a CSP name attribute and
3877 use it. Make -CSP option work again in pkcs12 utility.
3880 *) Add new functionality to the bn blinding code:
3881 - automatic re-creation of the BN_BLINDING parameters after
3882 a fixed number of uses (currently 32)
3883 - add new function for parameter creation
3884 - introduce flags to control the update behaviour of the
3885 BN_BLINDING parameters
3886 - hide BN_BLINDING structure
3887 Add a second BN_BLINDING slot to the RSA structure to improve
3888 performance when a single RSA object is shared among several
3892 *) Add support for DTLS.
3893 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3895 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3896 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3899 *) Remove buggy and incompletet DH cert support from
3900 ssl/ssl_rsa.c and ssl/s3_both.c
3903 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3904 the apps/openssl applications.
3907 *) Compile clean with "-Wall -Wmissing-prototypes
3908 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3909 DEBUG_SAFESTACK must also be set.
3912 *) Change ./Configure so that certain algorithms can be disabled by default.
3913 The new counterpiece to "no-xxx" is "enable-xxx".
3915 The patented RC5 and MDC2 algorithms will now be disabled unless
3916 "enable-rc5" and "enable-mdc2", respectively, are specified.
3918 (IDEA remains enabled despite being patented. This is because IDEA
3919 is frequently required for interoperability, and there is no license
3920 fee for non-commercial use. As before, "no-idea" can be used to
3921 avoid this algorithm.)
3925 *) Add processing of proxy certificates (see RFC 3820). This work was
3926 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3927 EGEE (Enabling Grids for E-science in Europe).
3930 *) RC4 performance overhaul on modern architectures/implementations, such
3931 as Intel P4, IA-64 and AMD64.
3934 *) New utility extract-section.pl. This can be used specify an alternative
3935 section number in a pod file instead of having to treat each file as
3936 a separate case in Makefile. This can be done by adding two lines to the
3939 =for comment openssl_section:XXX
3941 The blank line is mandatory.
3945 *) New arguments -certform, -keyform and -pass for s_client and s_server
3946 to allow alternative format key and certificate files and passphrase
3950 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3951 update associated structures and add various utility functions.
3953 Add new policy related verify parameters, include policy checking in
3954 standard verify code. Enhance 'smime' application with extra parameters
3955 to support policy checking and print out.
3958 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3959 Nehemiah processors. These extensions support AES encryption in hardware
3960 as well as RNG (though RNG support is currently disabled).
3961 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3963 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3966 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3967 [Andy Polyakov and a number of other people]
3969 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3970 implementation contributed by IBM.
3971 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3973 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3974 exponent rather than 'unsigned long'. There is a corresponding change to
3975 the new 'rsa_keygen' element of the RSA_METHOD structure.
3976 [Jelte Jansen, Geoff Thorpe]
3978 *) Functionality for creating the initial serial number file is now
3979 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3981 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3982 number file to 1, which is bound to cause problems. To avoid
3983 the problems while respecting compatibility between different 0.9.7
3984 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3985 CA.pl for serial number initialization. With the new release 0.9.8,
3986 we can fix the problem directly in the 'ca' utility.)
3989 *) Reduced header interdepencies by declaring more opaque objects in
3990 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3991 give fewer recursive includes, which could break lazy source code - so
3992 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3993 developers should define this symbol when building and using openssl to
3994 ensure they track the recommended behaviour, interfaces, [etc], but
3995 backwards-compatible behaviour prevails when this isn't defined.
3998 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4001 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4002 This will generate a random key of the appropriate length based on the
4003 cipher context. The EVP_CIPHER can provide its own random key generation
4004 routine to support keys of a specific form. This is used in the des and
4005 3des routines to generate a key of the correct parity. Update S/MIME
4006 code to use new functions and hence generate correct parity DES keys.
4007 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4008 valid (weak or incorrect parity).
4011 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4012 as looking them up. This is useful when the verified structure may contain
4013 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4014 present unless the new PKCS7_NO_CRL flag is asserted.
4017 *) Extend ASN1 oid configuration module. It now additionally accepts the
4020 shortName = some long name, 1.2.3.4
4023 *) Reimplemented the BN_CTX implementation. There is now no more static
4024 limitation on the number of variables it can handle nor the depth of the
4025 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4026 information can now expand as required, and rather than having a single
4027 static array of bignums, BN_CTX now uses a linked-list of such arrays
4028 allowing it to expand on demand whilst maintaining the usefulness of
4029 BN_CTX's "bundling".
4032 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4033 to allow all RSA operations to function using a single BN_CTX.
4036 *) Preliminary support for certificate policy evaluation and checking. This
4037 is initially intended to pass the tests outlined in "Conformance Testing
4038 of Relying Party Client Certificate Path Processing Logic" v1.07.
4041 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4042 remained unused and not that useful. A variety of other little bignum
4043 tweaks and fixes have also been made continuing on from the audit (see
4047 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4048 associated ASN1, EVP and SSL functions and old ASN1 macros.
4051 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4052 and this should never fail. So the return value from the use of
4053 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4054 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4057 *) BN_CTX_get() should return zero-valued bignums, providing the same
4058 initialised value as BN_new().
4059 [Geoff Thorpe, suggested by Ulf Möller]
4061 *) Support for inhibitAnyPolicy certificate extension.
4064 *) An audit of the BIGNUM code is underway, for which debugging code is
4065 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4066 is considered valid when processing BIGNUMs, and causes execution to
4067 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4068 further steps are taken to deliberately pollute unused data in BIGNUM
4069 structures to try and expose faulty code further on. For now, openssl will
4070 (in its default mode of operation) continue to tolerate the inconsistent
4071 forms that it has tolerated in the past, but authors and packagers should
4072 consider trying openssl and their own applications when compiled with
4073 these debugging symbols defined. It will help highlight potential bugs in
4074 their own code, and will improve the test coverage for OpenSSL itself. At
4075 some point, these tighter rules will become openssl's default to improve
4076 maintainability, though the assert()s and other overheads will remain only
4077 in debugging configurations. See bn.h for more details.
4078 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4080 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4081 that can only be obtained through BN_CTX_new() (which implicitly
4082 initialises it). The presence of this function only made it possible
4083 to overwrite an existing structure (and cause memory leaks).
4086 *) Because of the callback-based approach for implementing LHASH as a
4087 template type, lh_insert() adds opaque objects to hash-tables and
4088 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4089 to clean up those corresponding objects before destroying the hash table
4090 (and losing the object pointers). So some over-zealous constifications in
4091 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4092 objects as "const" and the lh_doall[_arg] callback wrappers are not
4093 prototyped to have "const" restrictions on the object pointers they are
4094 given (and so aren't required to cast them away any more).
4097 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4098 (speed) prefers to use its own implementation. The two implementations
4099 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4100 its object type properly exposed (MS_TM) instead of casting to/from "char
4101 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4102 aren't necessarily the greatest nomenclatures - but this is what was used
4103 internally to the implementation so I've used that for now.
4106 *) Ensure that deprecated functions do not get compiled when
4107 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4108 the self-tests were still using deprecated key-generation functions so
4109 these have been updated also.
4112 *) Reorganise PKCS#7 code to separate the digest location functionality
4113 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
4114 New function PKCS7_set_digest() to set the digest type for PKCS#7
4115 digestedData type. Add additional code to correctly generate the
4116 digestedData type and add support for this type in PKCS7 initialization
4120 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4121 structure of type "other".
4124 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4125 sure the loop does correctly stop and breaking ("division by zero")
4126 modulus operations are not performed. The (pre-generated) prime
4127 table crypto/bn/bn_prime.h was already correct, but it could not be
4128 re-generated on some platforms because of the "division by zero"
4129 situation in the script.
4130 [Ralf S. Engelschall]
4132 *) Update support for ECC-based TLS ciphersuites according to
4133 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4134 SHA-1 now is only used for "small" curves (where the
4135 representation of a field element takes up to 24 bytes); for
4136 larger curves, the field element resulting from ECDH is directly
4137 used as premaster secret.
4138 [Douglas Stebila (Sun Microsystems Laboratories)]
4140 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4141 curve secp160r1 to the tests.
4142 [Douglas Stebila (Sun Microsystems Laboratories)]
4144 *) Add the possibility to load symbols globally with DSO.
4145 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4147 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4148 control of the error stack.
4151 *) Add support for STORE in ENGINE.
4154 *) Add the STORE type. The intention is to provide a common interface
4155 to certificate and key stores, be they simple file-based stores, or
4156 HSM-type store, or LDAP stores, or...
4157 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4160 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4161 pass a list of arguments to any function as well as provide a way
4162 for a function to pass data back to the caller.
4165 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4166 works like BUF_strdup() but can be used to duplicate a portion of
4167 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4171 *) Add the function sk_find_ex() which works like sk_find(), but will
4172 return an index to an element even if an exact match couldn't be
4173 found. The index is guaranteed to point at the element where the
4174 searched-for key would be inserted to preserve sorting order.
4177 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4178 takes an extra flags argument for optional functionality. Currently,
4179 the following flags are defined:
4181 OBJ_BSEARCH_VALUE_ON_NOMATCH
4182 This one gets OBJ_bsearch_ex() to return a pointer to the first
4183 element where the comparing function returns a negative or zero
4186 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4187 This one gets OBJ_bsearch_ex() to return a pointer to the first
4188 element where the comparing function returns zero. This is useful
4189 if there are more than one element where the comparing function
4193 *) Make it possible to create self-signed certificates with 'openssl ca'
4194 in such a way that the self-signed certificate becomes part of the
4195 CA database and uses the same mechanisms for serial number generation
4196 as all other certificate signing. The new flag '-selfsign' enables
4197 this functionality. Adapt CA.sh and CA.pl.in.
4200 *) Add functionality to check the public key of a certificate request
4201 against a given private. This is useful to check that a certificate
4202 request can be signed by that key (self-signing).
4205 *) Make it possible to have multiple active certificates with the same
4206 subject in the CA index file. This is done only if the keyword
4207 'unique_subject' is set to 'no' in the main CA section (default
4208 if 'CA_default') of the configuration file. The value is saved
4209 with the database itself in a separate index attribute file,
4210 named like the index file with '.attr' appended to the name.
4213 *) Generate muti valued AVAs using '+' notation in config files for
4217 *) Support for nameConstraints certificate extension.
4220 *) Support for policyConstraints certificate extension.
4223 *) Support for policyMappings certificate extension.
4226 *) Make sure the default DSA_METHOD implementation only uses its
4227 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4228 and change its own handlers to be NULL so as to remove unnecessary
4229 indirection. This lets alternative implementations fallback to the
4230 default implementation more easily.
4233 *) Support for directoryName in GeneralName related extensions
4237 *) Make it possible to link applications using Makefile.shared.
4238 Make that possible even when linking against static libraries!
4241 *) Support for single pass processing for S/MIME signing. This now
4242 means that S/MIME signing can be done from a pipe, in addition
4243 cleartext signing (multipart/signed type) is effectively streaming
4244 and the signed data does not need to be all held in memory.
4246 This is done with a new flag PKCS7_STREAM. When this flag is set
4247 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4248 is done after the data is output (and digests calculated) in
4249 SMIME_write_PKCS7().
4252 *) Add full support for -rpath/-R, both in shared libraries and
4253 applications, at least on the platforms where it's known how
4257 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4258 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4259 will now compute a table of multiples of the generator that
4260 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4261 faster (notably in the case of a single point multiplication,
4262 scalar * generator).
4263 [Nils Larsch, Bodo Moeller]
4265 *) IPv6 support for certificate extensions. The various extensions
4266 which use the IP:a.b.c.d can now take IPv6 addresses using the
4267 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4271 *) Added an ENGINE that implements RSA by performing private key
4272 exponentiations with the GMP library. The conversions to and from
4273 GMP's mpz_t format aren't optimised nor are any montgomery forms
4274 cached, and on x86 it appears OpenSSL's own performance has caught up.
4275 However there are likely to be other architectures where GMP could
4276 provide a boost. This ENGINE is not built in by default, but it can be
4277 specified at Configure time and should be accompanied by the necessary
4278 linker additions, eg;
4279 ./config -DOPENSSL_USE_GMP -lgmp
4282 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4283 testing availability of engines with "-t" - the old behaviour is
4284 produced by increasing the feature's verbosity with "-tt".
4287 *) ECDSA routines: under certain error conditions uninitialized BN objects
4288 could be freed. Solution: make sure initialization is performed early
4289 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4293 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4294 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4295 software implementations. For DSA and DH, parameter generation can
4296 also be overriden by providing the appropriate method callbacks.
4299 *) Change the "progress" mechanism used in key-generation and
4300 primality testing to functions that take a new BN_GENCB pointer in
4301 place of callback/argument pairs. The new API functions have "_ex"
4302 postfixes and the older functions are reimplemented as wrappers for
4303 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4304 declarations of the old functions to help (graceful) attempts to
4305 migrate to the new functions. Also, the new key-generation API
4306 functions operate on a caller-supplied key-structure and return
4307 success/failure rather than returning a key or NULL - this is to
4308 help make "keygen" another member function of RSA_METHOD etc.
4310 Example for using the new callback interface:
4312 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4316 BN_GENCB_set(&my_cb, my_callback, my_arg);
4318 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4319 /* For the meaning of a, b in calls to my_callback(), see the
4320 * documentation of the function that calls the callback.
4321 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4322 * my_callback should return 1 if it wants BN_is_prime_ex()
4323 * to continue, or 0 to stop.
4328 *) Change the ZLIB compression method to be stateful, and make it
4329 available to TLS with the number defined in
4330 draft-ietf-tls-compression-04.txt.
4333 *) Add the ASN.1 structures and functions for CertificatePair, which
4334 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4336 CertificatePair ::= SEQUENCE {
4337 forward [0] Certificate OPTIONAL,
4338 reverse [1] Certificate OPTIONAL,
4339 -- at least one of the pair shall be present -- }
4341 Also implement the PEM functions to read and write certificate
4342 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4344 This needed to be defined, mostly for the sake of the LDAP
4345 attribute crossCertificatePair, but may prove useful elsewhere as
4349 *) Make it possible to inhibit symlinking of shared libraries in
4350 Makefile.shared, for Cygwin's sake.
4353 *) Extend the BIGNUM API by creating a function
4354 void BN_set_negative(BIGNUM *a, int neg);
4355 and a macro that behave like
4356 int BN_is_negative(const BIGNUM *a);
4358 to avoid the need to access 'a->neg' directly in applications.
4361 *) Implement fast modular reduction for pseudo-Mersenne primes
4362 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4363 EC_GROUP_new_curve_GFp() will now automatically use this
4365 [Nils Larsch <nla@trustcenter.de>]
4367 *) Add new lock type (CRYPTO_LOCK_BN).
4370 *) Change the ENGINE framework to automatically load engines
4371 dynamically from specific directories unless they could be
4372 found to already be built in or loaded. Move all the
4373 current engines except for the cryptodev one to a new
4375 The engines in engines/ are built as shared libraries if
4376 the "shared" options was given to ./Configure or ./config.
4377 Otherwise, they are inserted in libcrypto.a.
4378 /usr/local/ssl/engines is the default directory for dynamic
4379 engines, but that can be overriden at configure time through
4380 the usual use of --prefix and/or --openssldir, and at run
4381 time with the environment variable OPENSSL_ENGINES.
4382 [Geoff Thorpe and Richard Levitte]
4384 *) Add Makefile.shared, a helper makefile to build shared
4385 libraries. Addapt Makefile.org.
4388 *) Add version info to Win32 DLLs.
4389 [Peter 'Luna' Runestig" <peter@runestig.com>]
4391 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4392 can be added using this API to created arbitrary PKCS#12
4393 files while avoiding the low level API.
4395 New options to PKCS12_create(), key or cert can be NULL and
4396 will then be omitted from the output file. The encryption
4397 algorithm NIDs can be set to -1 for no encryption, the mac
4398 iteration count can be set to 0 to omit the mac.
4400 Enhance pkcs12 utility by making the -nokeys and -nocerts
4401 options work when creating a PKCS#12 file. New option -nomac
4402 to omit the mac, NONE can be set for an encryption algorithm.
4403 New code is modified to use the enhanced PKCS12_create()
4404 instead of the low level API.
4407 *) Extend ASN1 encoder to support indefinite length constructed
4408 encoding. This can output sequences tags and octet strings in
4409 this form. Modify pk7_asn1.c to support indefinite length
4410 encoding. This is experimental and needs additional code to
4411 be useful, such as an ASN1 bio and some enhanced streaming
4414 Extend template encode functionality so that tagging is passed
4415 down to the template encoder.
4418 *) Let 'openssl req' fail if an argument to '-newkey' is not
4419 recognized instead of using RSA as a default.
4422 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4423 As these are not official, they are not included in "ALL";
4424 the "ECCdraft" ciphersuite group alias can be used to select them.
4425 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4427 *) Add ECDH engine support.
4428 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4430 *) Add ECDH in new directory crypto/ecdh/.
4431 [Douglas Stebila (Sun Microsystems Laboratories)]
4433 *) Let BN_rand_range() abort with an error after 100 iterations
4434 without success (which indicates a broken PRNG).
4437 *) Change BN_mod_sqrt() so that it verifies that the input value
4438 is really the square of the return value. (Previously,
4439 BN_mod_sqrt would show GIGO behaviour.)
4442 *) Add named elliptic curves over binary fields from X9.62, SECG,
4443 and WAP/WTLS; add OIDs that were still missing.
4445 [Sheueling Chang Shantz and Douglas Stebila
4446 (Sun Microsystems Laboratories)]
4448 *) Extend the EC library for elliptic curves over binary fields
4449 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4452 EC_GF2m_simple_method
4456 EC_GROUP_new_curve_GF2m
4457 EC_GROUP_set_curve_GF2m
4458 EC_GROUP_get_curve_GF2m
4459 EC_POINT_set_affine_coordinates_GF2m
4460 EC_POINT_get_affine_coordinates_GF2m
4461 EC_POINT_set_compressed_coordinates_GF2m
4463 Point compression for binary fields is disabled by default for
4464 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4467 As binary polynomials are represented as BIGNUMs, various members
4468 of the EC_GROUP and EC_POINT data structures can be shared
4469 between the implementations for prime fields and binary fields;
4470 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4471 are essentially identical to their ..._GFp counterparts.
4472 (For simplicity, the '..._GFp' prefix has been dropped from
4473 various internal method names.)
4475 An internal 'field_div' method (similar to 'field_mul' and
4476 'field_sqr') has been added; this is used only for binary fields.
4478 [Sheueling Chang Shantz and Douglas Stebila
4479 (Sun Microsystems Laboratories)]
4481 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4482 through methods ('mul', 'precompute_mult').
4484 The generic implementations (now internally called 'ec_wNAF_mul'
4485 and 'ec_wNAF_precomputed_mult') remain the default if these
4486 methods are undefined.
4488 [Sheueling Chang Shantz and Douglas Stebila
4489 (Sun Microsystems Laboratories)]
4491 *) New function EC_GROUP_get_degree, which is defined through
4492 EC_METHOD. For curves over prime fields, this returns the bit
4493 length of the modulus.
4495 [Sheueling Chang Shantz and Douglas Stebila
4496 (Sun Microsystems Laboratories)]
4498 *) New functions EC_GROUP_dup, EC_POINT_dup.
4499 (These simply call ..._new and ..._copy).
4501 [Sheueling Chang Shantz and Douglas Stebila
4502 (Sun Microsystems Laboratories)]
4504 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4505 Polynomials are represented as BIGNUMs (where the sign bit is not
4506 used) in the following functions [macros]:
4509 BN_GF2m_sub [= BN_GF2m_add]
4510 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4511 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4512 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4514 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4515 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4516 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4517 BN_GF2m_cmp [= BN_ucmp]
4519 (Note that only the 'mod' functions are actually for fields GF(2^m).
4520 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4522 For some functions, an the irreducible polynomial defining a
4523 field can be given as an 'unsigned int[]' with strictly
4524 decreasing elements giving the indices of those bits that are set;
4525 i.e., p[] represents the polynomial
4526 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4528 p[0] > p[1] > ... > p[k] = 0.
4529 This applies to the following functions:
4534 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4535 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4537 BN_GF2m_mod_sqrt_arr
4538 BN_GF2m_mod_solve_quad_arr
4542 Conversion can be performed by the following functions:
4547 bntest.c has additional tests for binary polynomial arithmetic.
4549 Two implementations for BN_GF2m_mod_div() are available.
4550 The default algorithm simply uses BN_GF2m_mod_inv() and
4551 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4552 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4553 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4555 [Sheueling Chang Shantz and Douglas Stebila
4556 (Sun Microsystems Laboratories)]
4558 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4559 functionality is disabled at compile-time.
4560 [Douglas Stebila <douglas.stebila@sun.com>]
4562 *) Change default behaviour of 'openssl asn1parse' so that more
4563 information is visible when viewing, e.g., a certificate:
4565 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4566 mode the content of non-printable OCTET STRINGs is output in a
4567 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4568 avoid the appearance of a printable string.
4569 [Nils Larsch <nla@trustcenter.de>]
4571 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4573 EC_GROUP_set_asn1_flag()
4574 EC_GROUP_get_asn1_flag()
4575 EC_GROUP_set_point_conversion_form()
4576 EC_GROUP_get_point_conversion_form()
4577 These control ASN1 encoding details:
4578 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4579 has been set to OPENSSL_EC_NAMED_CURVE.
4580 - Points are encoded in uncompressed form by default; options for
4581 asn1_for are as for point2oct, namely
4582 POINT_CONVERSION_COMPRESSED
4583 POINT_CONVERSION_UNCOMPRESSED
4584 POINT_CONVERSION_HYBRID
4586 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4589 EC_GROUP_get0_seed()
4590 EC_GROUP_get_seed_len()
4591 This is used only for ASN1 purposes (so far).
4592 [Nils Larsch <nla@trustcenter.de>]
4594 *) Add 'field_type' member to EC_METHOD, which holds the NID
4595 of the appropriate field type OID. The new function
4596 EC_METHOD_get_field_type() returns this value.
4597 [Nils Larsch <nla@trustcenter.de>]
4602 EC_POINT_point2hex()
4603 EC_POINT_hex2point()
4604 providing useful interfaces to EC_POINT_point2oct() and
4605 EC_POINT_oct2point().
4606 [Nils Larsch <nla@trustcenter.de>]
4608 *) Change internals of the EC library so that the functions
4609 EC_GROUP_set_generator()
4610 EC_GROUP_get_generator()
4611 EC_GROUP_get_order()
4612 EC_GROUP_get_cofactor()
4613 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4614 to methods, which would lead to unnecessary code duplication when
4615 adding different types of curves.
4616 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4618 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4619 arithmetic, and such that modified wNAFs are generated
4620 (which avoid length expansion in many cases).
4623 *) Add a function EC_GROUP_check_discriminant() (defined via
4624 EC_METHOD) that verifies that the curve discriminant is non-zero.
4626 Add a function EC_GROUP_check() that makes some sanity tests
4627 on a EC_GROUP, its generator and order. This includes
4628 EC_GROUP_check_discriminant().
4629 [Nils Larsch <nla@trustcenter.de>]
4631 *) Add ECDSA in new directory crypto/ecdsa/.
4633 Add applications 'openssl ecparam' and 'openssl ecdsa'
4634 (these are based on 'openssl dsaparam' and 'openssl dsa').
4636 ECDSA support is also included in various other files across the
4637 library. Most notably,
4638 - 'openssl req' now has a '-newkey ecdsa:file' option;
4639 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4640 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4641 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4642 them suitable for ECDSA where domain parameters must be
4643 extracted before the specific public key;
4644 - ECDSA engine support has been added.
4645 [Nils Larsch <nla@trustcenter.de>]
4647 *) Include some named elliptic curves, and add OIDs from X9.62,
4648 SECG, and WAP/WTLS. Each curve can be obtained from the new
4650 EC_GROUP_new_by_curve_name(),
4651 and the list of available named curves can be obtained with
4652 EC_get_builtin_curves().
4653 Also add a 'curve_name' member to EC_GROUP objects, which can be
4655 EC_GROUP_set_curve_name()
4656 EC_GROUP_get_curve_name()
4657 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4659 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4660 was actually never needed) and in BN_mul(). The removal in BN_mul()
4661 required a small change in bn_mul_part_recursive() and the addition
4662 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4663 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4664 bn_sub_words() and bn_add_words() except they take arrays with
4668 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4670 *) Cleanse PEM buffers before freeing them since they may contain
4672 [Benjamin Bennett <ben@psc.edu>]
4674 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4675 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4676 authentication-only ciphersuites.
4679 *) Since AES128 and AES256 share a single mask bit in the logic of
4680 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4681 kludge to work properly if AES128 is available and AES256 isn't.
4684 *) Expand security boundary to match 1.1.1 module.
4687 *) Remove redundant features: hash file source, editing of test vectors
4688 modify fipsld to use external fips_premain.c signature.
4691 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4692 run algorithm test programs.
4695 *) Make algorithm test programs more tolerant of whitespace.
4698 *) Have SSL/TLS server implementation tolerate "mismatched" record
4699 protocol version while receiving ClientHello even if the
4700 ClientHello is fragmented. (The server can't insist on the
4701 particular protocol version it has chosen before the ServerHello
4702 message has informed the client about his choice.)
4705 *) Load error codes if they are not already present instead of using a
4706 static variable. This allows them to be cleanly unloaded and reloaded.
4709 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4711 *) Introduce limits to prevent malicious keys being able to
4712 cause a denial of service. (CVE-2006-2940)
4713 [Steve Henson, Bodo Moeller]
4715 *) Fix ASN.1 parsing of certain invalid structures that can result
4716 in a denial of service. (CVE-2006-2937) [Steve Henson]
4718 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4719 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4721 *) Fix SSL client code which could crash if connecting to a
4722 malicious SSLv2 server. (CVE-2006-4343)
4723 [Tavis Ormandy and Will Drewry, Google Security Team]
4725 *) Change ciphersuite string processing so that an explicit
4726 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4727 will no longer include "AES128-SHA"), and any other similar
4728 ciphersuite (same bitmap) from *other* protocol versions (so that
4729 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4730 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4731 changes from 0.9.8b and 0.9.8d.
4734 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4736 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4737 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4739 *) Change the Unix randomness entropy gathering to use poll() when
4740 possible instead of select(), since the latter has some
4741 undesirable limitations.
4742 [Darryl Miles via Richard Levitte and Bodo Moeller]
4744 *) Disable rogue ciphersuites:
4746 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4747 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4748 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4750 The latter two were purportedly from
4751 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4754 Also deactive the remaining ciphersuites from
4755 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4756 unofficial, and the ID has long expired.
4759 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4760 dual-core machines) and other potential thread-safety issues.
4763 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4765 *) Adapt fipsld and the build system to link against the validated FIPS
4766 module in FIPS mode.
4769 *) Fixes for VC++ 2005 build under Windows.
4772 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4773 from a Windows bash shell such as MSYS. It is autodetected from the
4774 "config" script when run from a VC++ environment. Modify standard VC++
4775 build to use fipscanister.o from the GNU make build.
4778 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4780 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4781 The value now differs depending on if you build for FIPS or not.
4782 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4783 safely run with a non-FIPSed libcrypto, as it may crash because of
4784 the difference induced by this change.
4787 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4789 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4790 (part of SSL_OP_ALL). This option used to disable the
4791 countermeasure against man-in-the-middle protocol-version
4792 rollback in the SSL 2.0 server implementation, which is a bad
4793 idea. (CVE-2005-2969)
4795 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4796 for Information Security, National Institute of Advanced Industrial
4797 Science and Technology [AIST], Japan)]
4799 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4800 mainly for FIPS compliance and not fully integrated at this stage.
4803 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4804 the exponentiation using a fixed-length exponent. (Otherwise,
4805 the information leaked through timing could expose the secret key
4806 after many signatures; cf. Bleichenbacher's attack on DSA with
4810 *) Make a new fixed-window mod_exp implementation the default for
4811 RSA, DSA, and DH private-key operations so that the sequence of
4812 squares and multiplies and the memory access pattern are
4813 independent of the particular secret key. This will mitigate
4814 cache-timing and potential related attacks.
4816 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4817 and this is automatically used by BN_mod_exp_mont() if the new flag
4818 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4819 will use this BN flag for private exponents unless the flag
4820 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4821 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4823 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4825 *) Change the client implementation for SSLv23_method() and
4826 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4827 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4828 (Previously, the SSL 2.0 backwards compatible Client Hello
4829 message format would be used even with SSL_OP_NO_SSLv2.)
4832 *) Add support for smime-type MIME parameter in S/MIME messages which some
4836 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4837 a threadsafe manner. Modify rsa code to use new function and add calls
4838 to dsa and dh code (which had race conditions before).
4841 *) Include the fixed error library code in the C error file definitions
4842 instead of fixing them up at runtime. This keeps the error code
4843 structures constant.
4846 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4848 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4851 *) Fixes for newer kerberos headers. NB: the casts are needed because
4852 the 'length' field is signed on one version and unsigned on another
4853 with no (?) obvious way to tell the difference, without these VC++
4854 complains. Also the "definition" of FAR (blank) is no longer included
4855 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4856 some needed definitions.
4859 *) Undo Cygwin change.
4862 *) Added support for proxy certificates according to RFC 3820.
4863 Because they may be a security thread to unaware applications,
4864 they must be explicitely allowed in run-time. See
4865 docs/HOWTO/proxy_certificates.txt for further information.
4868 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4870 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4871 server and client random values. Previously
4872 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4873 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4875 This change has negligible security impact because:
4877 1. Server and client random values still have 24 bytes of pseudo random
4880 2. Server and client random values are sent in the clear in the initial
4883 3. The master secret is derived using the premaster secret (48 bytes in
4884 size for static RSA ciphersuites) as well as client server and random
4887 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4890 [Stephen Henson, reported by UK NISCC]
4892 *) Use Windows randomness collection on Cygwin.
4895 *) Fix hang in EGD/PRNGD query when communication socket is closed
4896 prematurely by EGD/PRNGD.
4897 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4899 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4902 *) Back-port of selected performance improvements from development
4903 branch, as well as improved support for PowerPC platforms.
4906 *) Add lots of checks for memory allocation failure, error codes to indicate
4907 failure and freeing up memory if a failure occurs.
4908 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4910 *) Add new -passin argument to dgst.
4913 *) Perform some character comparisons of different types in X509_NAME_cmp:
4914 this is needed for some certificates that reencode DNs into UTF8Strings
4915 (in violation of RFC3280) and can't or wont issue name rollover
4919 *) Make an explicit check during certificate validation to see that
4920 the CA setting in each certificate on the chain is correct. As a
4921 side effect always do the following basic checks on extensions,
4922 not just when there's an associated purpose to the check:
4924 - if there is an unhandled critical extension (unless the user
4925 has chosen to ignore this fault)
4926 - if the path length has been exceeded (if one is set at all)
4927 - that certain extensions fit the associated purpose (if one has
4931 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4933 *) Avoid a race condition when CRLs are checked in a multi threaded
4934 environment. This would happen due to the reordering of the revoked
4935 entries during signature checking and serial number lookup. Now the
4936 encoding is cached and the serial number sort performed under a lock.
4937 Add new STACK function sk_is_sorted().
4940 *) Add Delta CRL to the extension code.
4943 *) Various fixes to s3_pkt.c so alerts are sent properly.
4944 [David Holmes <d.holmes@f5.com>]
4946 *) Reduce the chances of duplicate issuer name and serial numbers (in
4947 violation of RFC3280) using the OpenSSL certificate creation utilities.
4948 This is done by creating a random 64 bit value for the initial serial
4949 number when a serial number file is created or when a self signed
4950 certificate is created using 'openssl req -x509'. The initial serial
4951 number file is created using 'openssl x509 -next_serial' in CA.pl
4952 rather than being initialized to 1.
4955 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4957 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4958 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4959 [Joe Orton, Steve Henson]
4961 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4963 [Joe Orton, Steve Henson]
4965 *) Make it possible to have multiple active certificates with the same
4966 subject in the CA index file. This is done only if the keyword
4967 'unique_subject' is set to 'no' in the main CA section (default
4968 if 'CA_default') of the configuration file. The value is saved
4969 with the database itself in a separate index attribute file,
4970 named like the index file with '.attr' appended to the name.
4973 *) X509 verify fixes. Disable broken certificate workarounds when
4974 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4975 keyUsage extension present. Don't accept CRLs with unhandled critical
4976 extensions: since verify currently doesn't process CRL extensions this
4977 rejects a CRL with *any* critical extensions. Add new verify error codes
4981 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4982 A clarification of RFC2560 will require the use of OCTET STRINGs and
4983 some implementations cannot handle the current raw format. Since OpenSSL
4984 copies and compares OCSP nonces as opaque blobs without any attempt at
4985 parsing them this should not create any compatibility issues.
4988 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4989 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4990 this HMAC (and other) operations are several times slower than OpenSSL
4994 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4995 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4997 *) Use the correct content when signing type "other".
5000 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5002 *) Fix various bugs revealed by running the NISCC test suite:
5004 Stop out of bounds reads in the ASN1 code when presented with
5005 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5007 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5009 If verify callback ignores invalid public key errors don't try to check
5010 certificate signature with the NULL public key.
5014 *) New -ignore_err option in ocsp application to stop the server
5015 exiting on the first error in a request.
5018 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5019 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5023 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5024 extra data after the compression methods not only for TLS 1.0
5025 but also for SSL 3.0 (as required by the specification).
5026 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5028 *) Change X509_certificate_type() to mark the key as exported/exportable
5029 when it's 512 *bits* long, not 512 bytes.
5032 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5033 blocks during encryption.
5036 *) Various fixes to base64 BIO and non blocking I/O. On write
5037 flushes were not handled properly if the BIO retried. On read
5038 data was not being buffered properly and had various logic bugs.
5039 This also affects blocking I/O when the data being decoded is a
5043 *) Various S/MIME bugfixes and compatibility changes:
5044 output correct application/pkcs7 MIME type if
5045 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5046 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5047 of files as .eml work). Correctly handle very long lines in MIME
5051 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5053 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5054 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5055 a protocol version number mismatch like a decryption error
5056 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5059 *) Turn on RSA blinding by default in the default implementation
5060 to avoid a timing attack. Applications that don't want it can call
5061 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5062 They would be ill-advised to do so in most cases.
5063 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5065 *) Change RSA blinding code so that it works when the PRNG is not
5066 seeded (in this case, the secret RSA exponent is abused as
5067 an unpredictable seed -- if it is not unpredictable, there
5068 is no point in blinding anyway). Make RSA blinding thread-safe
5069 by remembering the creator's thread ID in rsa->blinding and
5070 having all other threads use local one-time blinding factors
5071 (this requires more computation than sharing rsa->blinding, but
5072 avoids excessive locking; and if an RSA object is not shared
5073 between threads, blinding will still be very fast).
5076 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5077 ENGINE as defaults for all supported algorithms irrespective of
5078 the 'flags' parameter. 'flags' is now honoured, so applications
5079 should make sure they are passing it correctly.
5082 *) Target "mingw" now allows native Windows code to be generated in
5083 the Cygwin environment as well as with the MinGW compiler.
5086 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5088 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5089 via timing by performing a MAC computation even if incorrrect
5090 block cipher padding has been found. This is a countermeasure
5091 against active attacks where the attacker has to distinguish
5092 between bad padding and a MAC verification error. (CVE-2003-0078)
5094 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5095 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5096 Martin Vuagnoux (EPFL, Ilion)]
5098 *) Make the no-err option work as intended. The intention with no-err
5099 is not to have the whole error stack handling routines removed from
5100 libcrypto, it's only intended to remove all the function name and
5101 reason texts, thereby removing some of the footprint that may not
5102 be interesting if those errors aren't displayed anyway.
5104 NOTE: it's still possible for any application or module to have it's
5105 own set of error texts inserted. The routines are there, just not
5106 used by default when no-err is given.
5109 *) Add support for FreeBSD on IA64.
5110 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5112 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5113 Kerberos function mit_des_cbc_cksum(). Before this change,
5114 the value returned by DES_cbc_cksum() was like the one from
5115 mit_des_cbc_cksum(), except the bytes were swapped.
5116 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5118 *) Allow an application to disable the automatic SSL chain building.
5119 Before this a rather primitive chain build was always performed in
5120 ssl3_output_cert_chain(): an application had no way to send the
5121 correct chain if the automatic operation produced an incorrect result.
5123 Now the chain builder is disabled if either:
5125 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5127 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5129 The reasoning behind this is that an application would not want the
5130 auto chain building to take place if extra chain certificates are
5131 present and it might also want a means of sending no additional
5132 certificates (for example the chain has two certificates and the
5136 *) Add the possibility to build without the ENGINE framework.
5137 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5139 *) Under Win32 gmtime() can return NULL: check return value in
5140 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5143 *) DSA routines: under certain error conditions uninitialized BN objects
5144 could be freed. Solution: make sure initialization is performed early
5145 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5146 Nils Larsch <nla@trustcenter.de> via PR#459)
5149 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5150 checked on reconnect on the client side, therefore session resumption
5151 could still fail with a "ssl session id is different" error. This
5152 behaviour is masked when SSL_OP_ALL is used due to
5153 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5154 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5155 followup to PR #377.
5158 *) IA-32 assembler support enhancements: unified ELF targets, support
5159 for SCO/Caldera platforms, fix for Cygwin shared build.
5162 *) Add support for FreeBSD on sparc64. As a consequence, support for
5163 FreeBSD on non-x86 processors is separate from x86 processors on
5164 the config script, much like the NetBSD support.
5165 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5167 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5169 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5172 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5173 code (06) was taken as the first octet of the session ID and the last
5174 octet was ignored consequently. As a result SSLv2 client side session
5175 caching could not have worked due to the session ID mismatch between
5177 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5181 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5182 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5186 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5187 seems that in spite of existing for more than a year, many application
5188 author have done nothing to provide the necessary callbacks, which
5189 means that this particular engine will not work properly anywhere.
5190 This is a very unfortunate situation which forces us, in the name
5191 of usability, to give the hw_ncipher.c a static lock, which is part
5193 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5194 appear in 0.9.8 or later. We EXPECT application authors to have
5195 dealt properly with this when 0.9.8 is released (unless we actually
5196 make such changes in the libcrypto locking code that changes will
5197 have to be made anyway).
5200 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5201 octets have been read, EOF or an error occurs. Without this change
5202 some truncated ASN1 structures will not produce an error.
5205 *) Disable Heimdal support, since it hasn't been fully implemented.
5206 Still give the possibility to force the use of Heimdal, but with
5207 warnings and a request that patches get sent to openssl-dev.
5210 *) Add the VC-CE target, introduce the WINCE sysname, and add
5211 INSTALL.WCE and appropriate conditionals to make it build.
5212 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5214 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5215 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5216 edit numbers of the version.
5217 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5219 *) Introduce safe string copy and catenation functions
5220 (BUF_strlcpy() and BUF_strlcat()).
5221 [Ben Laurie (CHATS) and Richard Levitte]
5223 *) Avoid using fixed-size buffers for one-line DNs.
5224 [Ben Laurie (CHATS)]
5226 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5227 resizing buffers containing secrets, and use where appropriate.
5228 [Ben Laurie (CHATS)]
5230 *) Avoid using fixed size buffers for configuration file location.
5231 [Ben Laurie (CHATS)]
5233 *) Avoid filename truncation for various CA files.
5234 [Ben Laurie (CHATS)]
5236 *) Use sizeof in preference to magic numbers.
5237 [Ben Laurie (CHATS)]
5239 *) Avoid filename truncation in cert requests.
5240 [Ben Laurie (CHATS)]
5242 *) Add assertions to check for (supposedly impossible) buffer
5244 [Ben Laurie (CHATS)]
5246 *) Don't cache truncated DNS entries in the local cache (this could
5247 potentially lead to a spoofing attack).
5248 [Ben Laurie (CHATS)]
5250 *) Fix various buffers to be large enough for hex/decimal
5251 representations in a platform independent manner.
5252 [Ben Laurie (CHATS)]
5254 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5255 resizing buffers containing secrets, and use where appropriate.
5256 [Ben Laurie (CHATS)]
5258 *) Add BIO_indent() to avoid much slightly worrying code to do
5260 [Ben Laurie (CHATS)]
5262 *) Convert sprintf()/BIO_puts() to BIO_printf().
5263 [Ben Laurie (CHATS)]
5265 *) buffer_gets() could terminate with the buffer only half
5267 [Ben Laurie (CHATS)]
5269 *) Add assertions to prevent user-supplied crypto functions from
5270 overflowing internal buffers by having large block sizes, etc.
5271 [Ben Laurie (CHATS)]
5273 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5275 [Ben Laurie (CHATS)]
5277 *) Eliminate unused copy of key in RC4.
5278 [Ben Laurie (CHATS)]
5280 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5281 [Ben Laurie (CHATS)]
5283 *) Fix off-by-one error in EGD path.
5284 [Ben Laurie (CHATS)]
5286 *) If RANDFILE path is too long, ignore instead of truncating.
5287 [Ben Laurie (CHATS)]
5289 *) Eliminate unused and incorrectly sized X.509 structure
5291 [Ben Laurie (CHATS)]
5293 *) Eliminate unused and dangerous function knumber().
5294 [Ben Laurie (CHATS)]
5296 *) Eliminate unused and dangerous structure, KSSL_ERR.
5297 [Ben Laurie (CHATS)]
5299 *) Protect against overlong session ID context length in an encoded
5300 session object. Since these are local, this does not appear to be
5302 [Ben Laurie (CHATS)]
5304 *) Change from security patch (see 0.9.6e below) that did not affect
5305 the 0.9.6 release series:
5307 Remote buffer overflow in SSL3 protocol - an attacker could
5308 supply an oversized master key in Kerberos-enabled versions.
5310 [Ben Laurie (CHATS)]
5312 *) Change the SSL kerb5 codes to match RFC 2712.
5315 *) Make -nameopt work fully for req and add -reqopt switch.
5316 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5318 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5319 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5321 *) Make sure tests can be performed even if the corresponding algorithms
5322 have been removed entirely. This was also the last step to make
5323 OpenSSL compilable with DJGPP under all reasonable conditions.
5324 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5326 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5327 to allow version independent disabling of normally unselected ciphers,
5328 which may be activated as a side-effect of selecting a single cipher.
5330 (E.g., cipher list string "RSA" enables ciphersuites that are left
5331 out of "ALL" because they do not provide symmetric encryption.
5332 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5333 [Lutz Jaenicke, Bodo Moeller]
5335 *) Add appropriate support for separate platform-dependent build
5336 directories. The recommended way to make a platform-dependent
5337 build directory is the following (tested on Linux), maybe with
5340 # Place yourself outside of the OpenSSL source tree. In
5341 # this example, the environment variable OPENSSL_SOURCE
5342 # is assumed to contain the absolute OpenSSL source directory.
5343 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5344 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5345 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5346 mkdir -p `dirname $F`
5347 ln -s $OPENSSL_SOURCE/$F $F
5350 To be absolutely sure not to disturb the source tree, a "make clean"
5351 is a good thing. If it isn't successfull, don't worry about it,
5352 it probably means the source directory is very clean.
5355 *) Make sure any ENGINE control commands make local copies of string
5356 pointers passed to them whenever necessary. Otherwise it is possible
5357 the caller may have overwritten (or deallocated) the original string
5358 data when a later ENGINE operation tries to use the stored values.
5359 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5361 *) Improve diagnostics in file reading and command-line digests.
5362 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5364 *) Add AES modes CFB and OFB to the object database. Correct an
5365 error in AES-CFB decryption.
5368 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5369 allows existing EVP_CIPHER_CTX structures to be reused after
5370 calling EVP_*Final(). This behaviour is used by encryption
5371 BIOs and some applications. This has the side effect that
5372 applications must explicitly clean up cipher contexts with
5373 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5376 *) Check the values of dna and dnb in bn_mul_recursive before calling
5377 bn_mul_comba (a non zero value means the a or b arrays do not contain
5378 n2 elements) and fallback to bn_mul_normal if either is not zero.
5381 *) Fix escaping of non-ASCII characters when using the -subj option
5382 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5385 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5386 form for "surname", serialNumber has no short form.
5387 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5388 therefore remove "mail" short name for "internet 7".
5389 The OID for unique identifiers in X509 certificates is
5390 x500UniqueIdentifier, not uniqueIdentifier.
5391 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5394 *) Add an "init" command to the ENGINE config module and auto initialize
5395 ENGINEs. Without any "init" command the ENGINE will be initialized
5396 after all ctrl commands have been executed on it. If init=1 the
5397 ENGINE is initailized at that point (ctrls before that point are run
5398 on the uninitialized ENGINE and after on the initialized one). If
5399 init=0 then the ENGINE will not be iniatialized at all.
5402 *) Fix the 'app_verify_callback' interface so that the user-defined
5403 argument is actually passed to the callback: In the
5404 SSL_CTX_set_cert_verify_callback() prototype, the callback
5405 declaration has been changed from
5408 int (*cb)(X509_STORE_CTX *,void *);
5409 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5410 i=s->ctx->app_verify_callback(&ctx)
5411 has been changed into
5412 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5414 To update applications using SSL_CTX_set_cert_verify_callback(),
5415 a dummy argument can be added to their callback functions.
5416 [D. K. Smetters <smetters@parc.xerox.com>]
5418 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5419 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5421 *) Add and OPENSSL_LOAD_CONF define which will cause
5422 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5423 This allows older applications to transparently support certain
5424 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5425 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5426 load the config file and OPENSSL_add_all_algorithms_conf() which will
5427 always load it have also been added.
5430 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5431 Adjust NIDs and EVP layer.
5432 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5434 *) Config modules support in openssl utility.
5436 Most commands now load modules from the config file,
5437 though in a few (such as version) this isn't done
5438 because it couldn't be used for anything.
5440 In the case of ca and req the config file used is
5441 the same as the utility itself: that is the -config
5442 command line option can be used to specify an
5446 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5447 use "openssl_conf" if filename is NULL use default openssl config file.
5450 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5451 config section name. Add a new flag to tolerate a missing config file
5452 and move code to CONF_modules_load_file().
5455 *) Support for crypto accelerator cards from Accelerated Encryption
5456 Processing, www.aep.ie. (Use engine 'aep')
5457 The support was copied from 0.9.6c [engine] and adapted/corrected
5458 to work with the new engine framework.
5459 [AEP Inc. and Richard Levitte]
5461 *) Support for SureWare crypto accelerator cards from Baltimore
5462 Technologies. (Use engine 'sureware')
5463 The support was copied from 0.9.6c [engine] and adapted
5464 to work with the new engine framework.
5467 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5468 make the newer ENGINE framework commands for the CHIL engine work.
5469 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5471 *) Make it possible to produce shared libraries on ReliantUNIX.
5472 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5474 *) Add the configuration target debug-linux-ppro.
5475 Make 'openssl rsa' use the general key loading routines
5476 implemented in apps.c, and make those routines able to
5477 handle the key format FORMAT_NETSCAPE and the variant
5479 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5481 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5482 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5484 *) Add -keyform to rsautl, and document -engine.
5485 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5487 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5488 BIO_R_NO_SUCH_FILE error code rather than the generic
5489 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5492 *) Add new functions
5494 ERR_peek_last_error_line
5495 ERR_peek_last_error_line_data.
5496 These are similar to
5499 ERR_peek_error_line_data,
5500 but report on the latest error recorded rather than the first one
5501 still in the error queue.
5502 [Ben Laurie, Bodo Moeller]
5504 *) default_algorithms option in ENGINE config module. This allows things
5506 default_algorithms = ALL
5507 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5510 *) Prelminary ENGINE config module.
5513 *) New experimental application configuration code.
5516 *) Change the AES code to follow the same name structure as all other
5517 symmetric ciphers, and behave the same way. Move everything to
5518 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5519 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5521 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5522 [Ben Laurie and Theo de Raadt]
5524 *) Add option to output public keys in req command.
5525 [Massimiliano Pala madwolf@openca.org]
5527 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5528 (up to about 10% better than before for P-192 and P-224).
5531 *) New functions/macros
5533 SSL_CTX_set_msg_callback(ctx, cb)
5534 SSL_CTX_set_msg_callback_arg(ctx, arg)
5535 SSL_set_msg_callback(ssl, cb)
5536 SSL_set_msg_callback_arg(ssl, arg)
5538 to request calling a callback function
5540 void cb(int write_p, int version, int content_type,
5541 const void *buf, size_t len, SSL *ssl, void *arg)
5543 whenever a protocol message has been completely received
5544 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5545 protocol version according to which the SSL library interprets
5546 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5547 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5548 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5549 specification (change_cipher_spec(20), alert(21), handshake(22)).
5550 'buf' and 'len' point to the actual message, 'ssl' to the
5551 SSL object, and 'arg' is the application-defined value set by
5552 SSL[_CTX]_set_msg_callback_arg().
5554 'openssl s_client' and 'openssl s_server' have new '-msg' options
5555 to enable a callback that displays all protocol messages.
5558 *) Change the shared library support so shared libraries are built as
5559 soon as the corresponding static library is finished, and thereby get
5560 openssl and the test programs linked against the shared library.
5561 This still only happens when the keyword "shard" has been given to
5562 the configuration scripts.
5564 NOTE: shared library support is still an experimental thing, and
5565 backward binary compatibility is still not guaranteed.
5566 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5568 *) Add support for Subject Information Access extension.
5569 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5571 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5572 additional bytes when new memory had to be allocated, not just
5573 when reusing an existing buffer.
5576 *) New command line and configuration option 'utf8' for the req command.
5577 This allows field values to be specified as UTF8 strings.
5580 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5581 runs for the former and machine-readable output for the latter.
5584 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5585 of the e-mail address in the DN (i.e., it will go into a certificate
5586 extension only). The new configuration file option 'email_in_dn = no'
5587 has the same effect.
5588 [Massimiliano Pala madwolf@openca.org]
5590 *) Change all functions with names starting with des_ to be starting
5591 with DES_ instead. Add wrappers that are compatible with libdes,
5592 but are named _ossl_old_des_*. Finally, add macros that map the
5593 des_* symbols to the corresponding _ossl_old_des_* if libdes
5594 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5595 desired, the des_* symbols will be mapped to DES_*, with one
5598 Since we provide two compatibility mappings, the user needs to
5599 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5600 compatibility is desired. The default (i.e., when that macro
5601 isn't defined) is OpenSSL 0.9.6c compatibility.
5603 There are also macros that enable and disable the support of old
5604 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5605 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5606 are defined, the default will apply: to support the old des routines.
5608 In either case, one must include openssl/des.h to get the correct
5609 definitions. Do not try to just include openssl/des_old.h, that
5612 NOTE: This is a major break of an old API into a new one. Software
5613 authors are encouraged to switch to the DES_ style functions. Some
5614 time in the future, des_old.h and the libdes compatibility functions
5615 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5616 default), and then completely removed.
5619 *) Test for certificates which contain unsupported critical extensions.
5620 If such a certificate is found during a verify operation it is
5621 rejected by default: this behaviour can be overridden by either
5622 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5623 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5624 X509_supported_extension() has also been added which returns 1 if a
5625 particular extension is supported.
5628 *) Modify the behaviour of EVP cipher functions in similar way to digests
5629 to retain compatibility with existing code.
5632 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5633 compatibility with existing code. In particular the 'ctx' parameter does
5634 not have to be to be initialized before the call to EVP_DigestInit() and
5635 it is tidied up after a call to EVP_DigestFinal(). New function
5636 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5637 EVP_MD_CTX_copy() changed to not require the destination to be
5638 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5639 requires the destination to be valid.
5641 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5642 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5645 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5646 so that complete 'Handshake' protocol structures are kept in memory
5647 instead of overwriting 'msg_type' and 'length' with 'body' data.
5650 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5651 [Massimo Santin via Richard Levitte]
5653 *) Major restructuring to the underlying ENGINE code. This includes
5654 reduction of linker bloat, separation of pure "ENGINE" manipulation
5655 (initialisation, etc) from functionality dealing with implementations
5656 of specific crypto iterfaces. This change also introduces integrated
5657 support for symmetric ciphers and digest implementations - so ENGINEs
5658 can now accelerate these by providing EVP_CIPHER and EVP_MD
5659 implementations of their own. This is detailed in crypto/engine/README
5660 as it couldn't be adequately described here. However, there are a few
5661 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5662 were changed in the original introduction of ENGINE code have now
5663 reverted back - the hooking from this code to ENGINE is now a good
5664 deal more passive and at run-time, operations deal directly with
5665 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5666 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5667 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5668 they were not being used by the framework as there is no concept of a
5669 BIGNUM_METHOD and they could not be generalised to the new
5670 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5671 ENGINE_cpy() has been removed as it cannot be consistently defined in
5675 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5678 *) Change mkdef.pl to sort symbols that get the same entry number,
5679 and make sure the automatically generated functions ERR_load_*
5680 become part of libeay.num as well.
5683 *) New function SSL_renegotiate_pending(). This returns true once
5684 renegotiation has been requested (either SSL_renegotiate() call
5685 or HelloRequest/ClientHello receveived from the peer) and becomes
5686 false once a handshake has been completed.
5687 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5688 sends a HelloRequest, but does not ensure that a handshake takes
5689 place. SSL_renegotiate_pending() is useful for checking if the
5690 client has followed the request.)
5693 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5694 By default, clients may request session resumption even during
5695 renegotiation (if session ID contexts permit); with this option,
5696 session resumption is possible only in the first handshake.
5698 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5699 more bits available for options that should not be part of
5700 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5703 *) Add some demos for certificate and certificate request creation.
5706 *) Make maximum certificate chain size accepted from the peer application
5707 settable (SSL*_get/set_max_cert_list()), as proposed by
5708 "Douglas E. Engert" <deengert@anl.gov>.
5711 *) Add support for shared libraries for Unixware-7
5712 (Boyd Lynn Gerber <gerberb@zenez.com>).
5715 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5716 be done prior to destruction. Use this to unload error strings from
5717 ENGINEs that load their own error strings. NB: This adds two new API
5718 functions to "get" and "set" this destroy handler in an ENGINE.
5721 *) Alter all existing ENGINE implementations (except "openssl" and
5722 "openbsd") to dynamically instantiate their own error strings. This
5723 makes them more flexible to be built both as statically-linked ENGINEs
5724 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5725 Also, add stub code to each that makes building them as self-contained
5726 shared-libraries easier (see README.ENGINE).
5729 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5730 implementations into applications that are completely implemented in
5731 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5732 commands that can be used to configure what shared-library to load and
5733 to control aspects of the way it is handled. Also, made an update to
5734 the README.ENGINE file that brings its information up-to-date and
5735 provides some information and instructions on the "dynamic" ENGINE
5736 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5739 *) Make it possible to unload ranges of ERR strings with a new
5740 "ERR_unload_strings" function.
5743 *) Add a copy() function to EVP_MD.
5746 *) Make EVP_MD routines take a context pointer instead of just the
5747 md_data void pointer.
5750 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5751 that the digest can only process a single chunk of data
5752 (typically because it is provided by a piece of
5753 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5754 is only going to provide a single chunk of data, and hence the
5755 framework needn't accumulate the data for oneshot drivers.
5758 *) As with "ERR", make it possible to replace the underlying "ex_data"
5759 functions. This change also alters the storage and management of global
5760 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5761 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5762 index counters. The API functions that use this state have been changed
5763 to take a "class_index" rather than pointers to the class's local STACK
5764 and counter, and there is now an API function to dynamically create new
5765 classes. This centralisation allows us to (a) plug a lot of the
5766 thread-safety problems that existed, and (b) makes it possible to clean
5767 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5768 such data would previously have always leaked in application code and
5769 workarounds were in place to make the memory debugging turn a blind eye
5770 to it. Application code that doesn't use this new function will still
5771 leak as before, but their memory debugging output will announce it now
5772 rather than letting it slide.
5774 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5775 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5776 has a return value to indicate success or failure.
5779 *) Make it possible to replace the underlying "ERR" functions such that the
5780 global state (2 LHASH tables and 2 locks) is only used by the "default"
5781 implementation. This change also adds two functions to "get" and "set"
5782 the implementation prior to it being automatically set the first time
5783 any other ERR function takes place. Ie. an application can call "get",
5784 pass the return value to a module it has just loaded, and that module
5785 can call its own "set" function using that value. This means the
5786 module's "ERR" operations will use (and modify) the error state in the
5787 application and not in its own statically linked copy of OpenSSL code.
5790 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5791 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5792 the operation, and provides a more encapsulated way for external code
5793 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5794 to use these functions rather than manually incrementing the counts.
5796 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5799 *) Add EVP test program.
5802 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5805 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5806 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5807 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5808 These allow a CRL to be built without having to access X509_CRL fields
5809 directly. Modify 'ca' application to use new functions.
5812 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5813 bug workarounds. Rollback attack detection is a security feature.
5814 The problem will only arise on OpenSSL servers when TLSv1 is not
5815 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5816 Software authors not wanting to support TLSv1 will have special reasons
5817 for their choice and can explicitly enable this option.
5818 [Bodo Moeller, Lutz Jaenicke]
5820 *) Rationalise EVP so it can be extended: don't include a union of
5821 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5822 (similar to those existing for EVP_CIPHER_CTX).
5827 EVP_MD_CTX_init(&md); /* new function call */
5828 EVP_DigestInit(&md, EVP_sha1());
5829 EVP_DigestUpdate(&md, in, len);
5830 EVP_DigestFinal(&md, out, NULL);
5831 EVP_MD_CTX_cleanup(&md); /* new function call */
5835 *) Make DES key schedule conform to the usual scheme, as well as
5836 correcting its structure. This means that calls to DES functions
5837 now have to pass a pointer to a des_key_schedule instead of a
5838 plain des_key_schedule (which was actually always a pointer
5841 des_key_schedule ks;
5843 des_set_key_checked(..., &ks);
5844 des_ncbc_encrypt(..., &ks, ...);
5846 (Note that a later change renames 'des_...' into 'DES_...'.)
5849 *) Initial reduction of linker bloat: the use of some functions, such as
5850 PEM causes large amounts of unused functions to be linked in due to
5851 poor organisation. For example pem_all.c contains every PEM function
5852 which has a knock on effect of linking in large amounts of (unused)
5853 ASN1 code. Grouping together similar functions and splitting unrelated
5854 functions prevents this.
5857 *) Cleanup of EVP macros.
5860 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5861 correct _ecb suffix.
5864 *) Add initial OCSP responder support to ocsp application. The
5865 revocation information is handled using the text based index
5866 use by the ca application. The responder can either handle
5867 requests generated internally, supplied in files (for example
5868 via a CGI script) or using an internal minimal server.
5871 *) Add configuration choices to get zlib compression for TLS.
5874 *) Changes to Kerberos SSL for RFC 2712 compliance:
5875 1. Implemented real KerberosWrapper, instead of just using
5876 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5877 2. Implemented optional authenticator field of KerberosWrapper.
5879 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5880 and authenticator structs; see crypto/krb5/.
5882 Generalized Kerberos calls to support multiple Kerberos libraries.
5883 [Vern Staats <staatsvr@asc.hpc.mil>,
5884 Jeffrey Altman <jaltman@columbia.edu>
5885 via Richard Levitte]
5887 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5888 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5889 values for each of the key sizes rather than having just
5890 parameters (and 'speed' generating keys each time).
5893 *) Speed up EVP routines.
5896 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5897 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5898 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5899 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5901 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5902 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5903 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5906 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5908 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5911 *) Added the OS2-EMX target.
5912 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5914 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5915 to support NCONF routines in extension code. New function CONF_set_nconf()
5916 to allow functions which take an NCONF to also handle the old LHASH
5917 structure: this means that the old CONF compatible routines can be
5918 retained (in particular wrt extensions) without having to duplicate the
5919 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5922 *) Enhance the general user interface with mechanisms for inner control
5923 and with possibilities to have yes/no kind of prompts.
5926 *) Change all calls to low level digest routines in the library and
5927 applications to use EVP. Add missing calls to HMAC_cleanup() and
5928 don't assume HMAC_CTX can be copied using memcpy().
5929 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5931 *) Add the possibility to control engines through control names but with
5932 arbitrary arguments instead of just a string.
5933 Change the key loaders to take a UI_METHOD instead of a callback
5934 function pointer. NOTE: this breaks binary compatibility with earlier
5935 versions of OpenSSL [engine].
5936 Adapt the nCipher code for these new conditions and add a card insertion
5940 *) Enhance the general user interface with mechanisms to better support
5941 dialog box interfaces, application-defined prompts, the possibility
5942 to use defaults (for example default passwords from somewhere else)
5943 and interrupts/cancellations.
5946 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5947 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5950 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5951 tidy up some unnecessarily weird code in 'sk_new()').
5952 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5954 *) Change the key loading routines for ENGINEs to use the same kind
5955 callback (pem_password_cb) as all other routines that need this
5959 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5960 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5961 than this minimum value is recommended.
5964 *) New random seeder for OpenVMS, using the system process statistics
5965 that are easily reachable.
5968 *) Windows apparently can't transparently handle global
5969 variables defined in DLLs. Initialisations such as:
5971 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5973 wont compile. This is used by the any applications that need to
5974 declare their own ASN1 modules. This was fixed by adding the option
5975 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5976 needed for static libraries under Win32.
5979 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5980 setting of purpose and trust fields. New X509_STORE trust and
5981 purpose functions and tidy up setting in other SSL functions.
5984 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5985 structure. These are inherited by X509_STORE_CTX when it is
5986 initialised. This allows various defaults to be set in the
5987 X509_STORE structure (such as flags for CRL checking and custom
5988 purpose or trust settings) for functions which only use X509_STORE_CTX
5989 internally such as S/MIME.
5991 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5992 trust settings if they are not set in X509_STORE. This allows X509_STORE
5993 purposes and trust (in S/MIME for example) to override any set by default.
5995 Add command line options for CRL checking to smime, s_client and s_server
5999 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6000 are set then the CRL is looked up in the X509_STORE structure and
6001 its validity and signature checked, then if the certificate is found
6002 in the CRL the verify fails with a revoked error.
6004 Various new CRL related callbacks added to X509_STORE_CTX structure.
6006 Command line options added to 'verify' application to support this.
6008 This needs some additional work, such as being able to handle multiple
6009 CRLs with different times, extension based lookup (rather than just
6010 by subject name) and ultimately more complete V2 CRL extension
6014 *) Add a general user interface API (crypto/ui/). This is designed
6015 to replace things like des_read_password and friends (backward
6016 compatibility functions using this new API are provided).
6017 The purpose is to remove prompting functions from the DES code
6018 section as well as provide for prompting through dialog boxes in
6019 a window system and the like.
6022 *) Add "ex_data" support to ENGINE so implementations can add state at a
6023 per-structure level rather than having to store it globally.
6026 *) Make it possible for ENGINE structures to be copied when retrieved by
6027 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6028 This causes the "original" ENGINE structure to act like a template,
6029 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6030 operational state can be localised to each ENGINE structure, despite the
6031 fact they all share the same "methods". New ENGINE structures returned in
6032 this case have no functional references and the return value is the single
6033 structural reference. This matches the single structural reference returned
6034 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6038 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6039 needs to match any other type at all we need to manually clear the
6043 *) Changes to the "openssl engine" utility to include;
6044 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6045 about an ENGINE's available control commands.
6046 - executing control commands from command line arguments using the
6047 '-pre' and '-post' switches. '-post' is only used if '-t' is
6048 specified and the ENGINE is successfully initialised. The syntax for
6049 the individual commands are colon-separated, for example;
6050 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6053 *) New dynamic control command support for ENGINEs. ENGINEs can now
6054 declare their own commands (numbers), names (strings), descriptions,
6055 and input types for run-time discovery by calling applications. A
6056 subset of these commands are implicitly classed as "executable"
6057 depending on their input type, and only these can be invoked through
6058 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6059 can be based on user input, config files, etc). The distinction is
6060 that "executable" commands cannot return anything other than a boolean
6061 result and can only support numeric or string input, whereas some
6062 discoverable commands may only be for direct use through
6063 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6064 pointers, or other custom uses. The "executable" commands are to
6065 support parameterisations of ENGINE behaviour that can be
6066 unambiguously defined by ENGINEs and used consistently across any
6067 OpenSSL-based application. Commands have been added to all the
6068 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6069 control over shared-library paths without source code alterations.
6072 *) Changed all ENGINE implementations to dynamically allocate their
6073 ENGINEs rather than declaring them statically. Apart from this being
6074 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6075 this also allows the implementations to compile without using the
6076 internal engine_int.h header.
6079 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6080 'const' value. Any code that should be able to modify a RAND_METHOD
6081 should already have non-const pointers to it (ie. they should only
6082 modify their own ones).
6085 *) Made a variety of little tweaks to the ENGINE code.
6086 - "atalla" and "ubsec" string definitions were moved from header files
6087 to C code. "nuron" string definitions were placed in variables
6088 rather than hard-coded - allowing parameterisation of these values
6089 later on via ctrl() commands.
6090 - Removed unused "#if 0"'d code.
6091 - Fixed engine list iteration code so it uses ENGINE_free() to release
6092 structural references.
6093 - Constified the RAND_METHOD element of ENGINE structures.
6094 - Constified various get/set functions as appropriate and added
6095 missing functions (including a catch-all ENGINE_cpy that duplicates
6096 all ENGINE values onto a new ENGINE except reference counts/state).
6097 - Removed NULL parameter checks in get/set functions. Setting a method
6098 or function to NULL is a way of cancelling out a previously set
6099 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6100 and doesn't justify the extra error symbols and code.
6101 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6102 flags from engine_int.h to engine.h.
6103 - Changed prototypes for ENGINE handler functions (init(), finish(),
6104 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6107 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6108 to the algorithm using long division. The binary algorithm can be
6109 used only if the modulus is odd. On 32-bit systems, it is faster
6110 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6111 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6112 up to 450 bits. In 64-bit environments, the binary algorithm
6113 appears to be advantageous for much longer moduli; here we use it
6114 for moduli up to 2048 bits.
6117 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6118 could not support the combine flag in choice fields.
6121 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6122 extensions from a certificate request to the certificate.
6125 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6126 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6127 file: this allows the display of the certificate about to be
6128 signed to be customised, to allow certain fields to be included
6129 or excluded and extension details. The old system didn't display
6130 multicharacter strings properly, omitted fields not in the policy
6131 and couldn't display additional details such as extensions.
6134 *) Function EC_POINTs_mul for multiple scalar multiplication
6135 of an arbitrary number of elliptic curve points
6136 \sum scalars[i]*points[i],
6137 optionally including the generator defined for the EC_GROUP:
6138 scalar*generator + \sum scalars[i]*points[i].
6140 EC_POINT_mul is a simple wrapper function for the typical case
6141 that the point list has just one item (besides the optional
6145 *) First EC_METHODs for curves over GF(p):
6147 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6148 operations and provides various method functions that can also
6149 operate with faster implementations of modular arithmetic.
6151 EC_GFp_mont_method() reuses most functions that are part of
6152 EC_GFp_simple_method, but uses Montgomery arithmetic.
6154 [Bodo Moeller; point addition and point doubling
6155 implementation directly derived from source code provided by
6156 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6158 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6159 crypto/ec/ec_lib.c):
6161 Curves are EC_GROUP objects (with an optional group generator)
6162 based on EC_METHODs that are built into the library.
6164 Points are EC_POINT objects based on EC_GROUP objects.
6166 Most of the framework would be able to handle curves over arbitrary
6167 finite fields, but as there are no obvious types for fields other
6168 than GF(p), some functions are limited to that for now.
6171 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6172 that the file contains a complete HTTP response.
6175 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6176 change the def and num file printf format specifier from "%-40sXXX"
6177 to "%-39s XXX". The latter will always guarantee a space after the
6178 field while the former will cause them to run together if the field
6179 is 40 of more characters long.
6182 *) Constify the cipher and digest 'method' functions and structures
6183 and modify related functions to take constant EVP_MD and EVP_CIPHER
6187 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6188 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6191 *) Modify EVP_Digest*() routines so they now return values. Although the
6192 internal software routines can never fail additional hardware versions
6196 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6198 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6199 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6202 ERR_R_NESTED_ASN1_ERROR
6204 ERR_R_MISSING_ASN1_EOS
6205 were 4 .. 9, conflicting with
6206 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6208 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6209 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6211 Add new error code 'ERR_R_INTERNAL_ERROR'.
6214 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6218 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6219 sets the subject name for a new request or supersedes the
6220 subject name in a given request. Formats that can be parsed are
6221 'CN=Some Name, OU=myOU, C=IT'
6223 'CN=Some Name/OU=myOU/C=IT'.
6225 Add options '-batch' and '-verbose' to 'openssl req'.
6226 [Massimiliano Pala <madwolf@hackmasters.net>]
6228 *) Introduce the possibility to access global variables through
6229 functions on platform were that's the best way to handle exporting
6230 global variables in shared libraries. To enable this functionality,
6231 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6232 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6233 is normally done by Configure or something similar).
6235 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6236 in the source file (foo.c) like this:
6238 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6239 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6241 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6242 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6244 OPENSSL_DECLARE_GLOBAL(int,foo);
6245 #define foo OPENSSL_GLOBAL_REF(foo)
6246 OPENSSL_DECLARE_GLOBAL(double,bar);
6247 #define bar OPENSSL_GLOBAL_REF(bar)
6249 The #defines are very important, and therefore so is including the
6250 header file everywhere where the defined globals are used.
6252 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6253 of ASN.1 items, but that structure is a bit different.
6255 The largest change is in util/mkdef.pl which has been enhanced with
6256 better and easier to understand logic to choose which symbols should
6257 go into the Windows .def files as well as a number of fixes and code
6258 cleanup (among others, algorithm keywords are now sorted
6259 lexicographically to avoid constant rewrites).
6262 *) In BN_div() keep a copy of the sign of 'num' before writing the
6263 result to 'rm' because if rm==num the value will be overwritten
6264 and produce the wrong result if 'num' is negative: this caused
6265 problems with BN_mod() and BN_nnmod().
6268 *) Function OCSP_request_verify(). This checks the signature on an
6269 OCSP request and verifies the signer certificate. The signer
6270 certificate is just checked for a generic purpose and OCSP request
6274 *) Add OCSP_check_validity() function to check the validity of OCSP
6275 responses. OCSP responses are prepared in real time and may only
6276 be a few seconds old. Simply checking that the current time lies
6277 between thisUpdate and nextUpdate max reject otherwise valid responses
6278 caused by either OCSP responder or client clock inaccuracy. Instead
6279 we allow thisUpdate and nextUpdate to fall within a certain period of
6280 the current time. The age of the response can also optionally be
6281 checked. Two new options -validity_period and -status_age added to
6285 *) If signature or public key algorithm is unrecognized print out its
6286 OID rather that just UNKNOWN.
6289 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6290 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6291 ID to be generated from the issuer certificate alone which can then be
6292 passed to OCSP_id_issuer_cmp().
6295 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6296 ASN1 modules to export functions returning ASN1_ITEM pointers
6297 instead of the ASN1_ITEM structures themselves. This adds several
6298 new macros which allow the underlying ASN1 function/structure to
6299 be accessed transparently. As a result code should not use ASN1_ITEM
6300 references directly (such as &X509_it) but instead use the relevant
6301 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6302 use of the new ASN1 code on platforms where exporting structures
6303 is problematical (for example in shared libraries) but exporting
6304 functions returning pointers to structures is not.
6307 *) Add support for overriding the generation of SSL/TLS session IDs.
6308 These callbacks can be registered either in an SSL_CTX or per SSL.
6309 The purpose of this is to allow applications to control, if they wish,
6310 the arbitrary values chosen for use as session IDs, particularly as it
6311 can be useful for session caching in multiple-server environments. A
6312 command-line switch for testing this (and any client code that wishes
6313 to use such a feature) has been added to "s_server".
6314 [Geoff Thorpe, Lutz Jaenicke]
6316 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6317 of the form '#if defined(...) || defined(...) || ...' and
6318 '#if !defined(...) && !defined(...) && ...'. This also avoids
6319 the growing number of special cases it was previously handling.
6322 *) Make all configuration macros available for application by making
6323 sure they are available in opensslconf.h, by giving them names starting
6324 with "OPENSSL_" to avoid conflicts with other packages and by making
6325 sure e_os2.h will cover all platform-specific cases together with
6327 Additionally, it is now possible to define configuration/platform-
6328 specific names (called "system identities"). In the C code, these
6329 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6330 macro with the name beginning with "OPENSSL_SYS_", which is determined
6331 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6335 *) New option -set_serial to 'req' and 'x509' this allows the serial
6336 number to use to be specified on the command line. Previously self
6337 signed certificates were hard coded with serial number 0 and the
6338 CA options of 'x509' had to use a serial number in a file which was
6342 *) New options to 'ca' utility to support V2 CRL entry extensions.
6343 Currently CRL reason, invalidity date and hold instruction are
6344 supported. Add new CRL extensions to V3 code and some new objects.
6347 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6348 disable standard block padding (aka PKCS#5 padding) in the EVP
6349 API, which was previously mandatory. This means that the data is
6350 not padded in any way and so the total length much be a multiple
6351 of the block size, otherwise an error occurs.
6354 *) Initial (incomplete) OCSP SSL support.
6357 *) New function OCSP_parse_url(). This splits up a URL into its host,
6358 port and path components: primarily to parse OCSP URLs. New -url
6359 option to ocsp utility.
6362 *) New nonce behavior. The return value of OCSP_check_nonce() now
6363 reflects the various checks performed. Applications can decide
6364 whether to tolerate certain situations such as an absent nonce
6365 in a response when one was present in a request: the ocsp application
6366 just prints out a warning. New function OCSP_add1_basic_nonce()
6367 this is to allow responders to include a nonce in a response even if
6368 the request is nonce-less.
6371 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6372 skipped when using openssl x509 multiple times on a single input file,
6373 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6376 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6377 set string type: to handle setting ASN1_TIME structures. Fix ca
6378 utility to correctly initialize revocation date of CRLs.
6381 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6382 the clients preferred ciphersuites and rather use its own preferences.
6383 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6384 Internet Explorer by ensuring unchanged hash method during stepup.
6385 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6388 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6389 to aes and add a new 'exist' option to print out symbols that don't
6393 *) Additional options to ocsp utility to allow flags to be set and
6394 additional certificates supplied.
6397 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6398 OCSP client a number of certificate to only verify the response
6402 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6403 handle the new API. Currently only ECB, CBC modes supported. Add new
6406 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6407 Encryption Standard (AES) Ciphersuites for Transport Layer
6408 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6409 not enabled by default and were not part of the "ALL" ciphersuite
6410 alias because they were not yet official; they could be
6411 explicitly requested by specifying the "AESdraft" ciphersuite
6412 group alias. In the final release of OpenSSL 0.9.7, the group
6413 alias is called "AES" and is part of "ALL".)
6414 [Ben Laurie, Steve Henson, Bodo Moeller]
6416 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6417 request to response.
6420 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6421 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6422 extract information from a certificate request. OCSP_response_create()
6423 creates a response and optionally adds a basic response structure.
6424 OCSP_basic_add1_status() adds a complete single response to a basic
6425 response and returns the OCSP_SINGLERESP structure just added (to allow
6426 extensions to be included for example). OCSP_basic_add1_cert() adds a
6427 certificate to a basic response and OCSP_basic_sign() signs a basic
6428 response with various flags. New helper functions ASN1_TIME_check()
6429 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6430 (converts ASN1_TIME to GeneralizedTime).
6433 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6434 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6435 structure from a certificate. X509_pubkey_digest() digests the public_key
6436 contents: this is used in various key identifiers.
6439 *) Make sk_sort() tolerate a NULL argument.
6440 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6442 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6443 passed by the function are trusted implicitly. If any of them signed the
6444 response then it is assumed to be valid and is not verified.
6447 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6448 to data. This was previously part of the PKCS7 ASN1 code. This
6449 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6450 [Steve Henson, reported by Kenneth R. Robinette
6451 <support@securenetterm.com>]
6453 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6454 routines: without these tracing memory leaks is very painful.
6455 Fix leaks in PKCS12 and PKCS7 routines.
6458 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6459 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6460 effectively meant GeneralizedTime would never be used. Now it
6461 is initialised to -1 but X509_time_adj() now has to check the value
6462 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6463 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6464 [Steve Henson, reported by Kenneth R. Robinette
6465 <support@securenetterm.com>]
6467 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6468 result in a zero length in the ASN1_INTEGER structure which was
6469 not consistent with the structure when d2i_ASN1_INTEGER() was used
6470 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6471 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6472 where it did not print out a minus for negative ASN1_INTEGER.
6475 *) Add summary printout to ocsp utility. The various functions which
6476 convert status values to strings have been renamed to:
6477 OCSP_response_status_str(), OCSP_cert_status_str() and
6478 OCSP_crl_reason_str() and are no longer static. New options
6479 to verify nonce values and to disable verification. OCSP response
6480 printout format cleaned up.
6483 *) Add additional OCSP certificate checks. These are those specified
6484 in RFC2560. This consists of two separate checks: the CA of the
6485 certificate being checked must either be the OCSP signer certificate
6486 or the issuer of the OCSP signer certificate. In the latter case the
6487 OCSP signer certificate must contain the OCSP signing extended key
6488 usage. This check is performed by attempting to match the OCSP
6489 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6490 in the OCSP_CERTID structures of the response.
6493 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6494 and related routines. This uses the standard OpenSSL certificate
6495 verify routines to perform initial checks (just CA validity) and
6496 to obtain the certificate chain. Then additional checks will be
6497 performed on the chain. Currently the root CA is checked to see
6498 if it is explicitly trusted for OCSP signing. This is used to set
6499 a root CA as a global signing root: that is any certificate that
6500 chains to that CA is an acceptable OCSP signing certificate.
6503 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6504 extensions from a separate configuration file.
6505 As when reading extensions from the main configuration file,
6506 the '-extensions ...' option may be used for specifying the
6508 [Massimiliano Pala <madwolf@comune.modena.it>]
6510 *) New OCSP utility. Allows OCSP requests to be generated or
6511 read. The request can be sent to a responder and the output
6512 parsed, outputed or printed in text form. Not complete yet:
6513 still needs to check the OCSP response validity.
6516 *) New subcommands for 'openssl ca':
6517 'openssl ca -status <serial>' prints the status of the cert with
6518 the given serial number (according to the index file).
6519 'openssl ca -updatedb' updates the expiry status of certificates
6521 [Massimiliano Pala <madwolf@comune.modena.it>]
6523 *) New '-newreq-nodes' command option to CA.pl. This is like
6524 '-newreq', but calls 'openssl req' with the '-nodes' option
6525 so that the resulting key is not encrypted.
6526 [Damien Miller <djm@mindrot.org>]
6528 *) New configuration for the GNU Hurd.
6529 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6531 *) Initial code to implement OCSP basic response verify. This
6532 is currently incomplete. Currently just finds the signer's
6533 certificate and verifies the signature on the response.
6536 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6537 value of OPENSSLDIR. This is available via the new '-d' option
6538 to 'openssl version', and is also included in 'openssl version -a'.
6541 *) Allowing defining memory allocation callbacks that will be given
6542 file name and line number information in additional arguments
6543 (a const char* and an int). The basic functionality remains, as
6544 well as the original possibility to just replace malloc(),
6545 realloc() and free() by functions that do not know about these
6546 additional arguments. To register and find out the current
6547 settings for extended allocation functions, the following
6548 functions are provided:
6550 CRYPTO_set_mem_ex_functions
6551 CRYPTO_set_locked_mem_ex_functions
6552 CRYPTO_get_mem_ex_functions
6553 CRYPTO_get_locked_mem_ex_functions
6555 These work the same way as CRYPTO_set_mem_functions and friends.
6556 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6557 extended allocation function is enabled.
6558 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6559 a conventional allocation function is enabled.
6560 [Richard Levitte, Bodo Moeller]
6562 *) Finish off removing the remaining LHASH function pointer casts.
6563 There should no longer be any prototype-casting required when using
6564 the LHASH abstraction, and any casts that remain are "bugs". See
6565 the callback types and macros at the head of lhash.h for details
6566 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6569 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6570 If /dev/[u]random devices are not available or do not return enough
6571 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6573 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6574 /etc/entropy will be queried once each in this sequence, quering stops
6575 when enough entropy was collected without querying more sockets.
6578 *) Change the Unix RAND_poll() variant to be able to poll several
6579 random devices, as specified by DEVRANDOM, until a sufficient amount
6580 of data has been collected. We spend at most 10 ms on each file
6581 (select timeout) and read in non-blocking mode. DEVRANDOM now
6582 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6583 (previously it was just the string "/dev/urandom"), so on typical
6584 platforms the 10 ms delay will never occur.
6585 Also separate out the Unix variant to its own file, rand_unix.c.
6586 For VMS, there's a currently-empty rand_vms.c.
6589 *) Move OCSP client related routines to ocsp_cl.c. These
6590 provide utility functions which an application needing
6591 to issue a request to an OCSP responder and analyse the
6592 response will typically need: as opposed to those which an
6593 OCSP responder itself would need which will be added later.
6595 OCSP_request_sign() signs an OCSP request with an API similar
6596 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6597 response. OCSP_response_get1_basic() extracts basic response
6598 from response. OCSP_resp_find_status(): finds and extracts status
6599 information from an OCSP_CERTID structure (which will be created
6600 when the request structure is built). These are built from lower
6601 level functions which work on OCSP_SINGLERESP structures but
6602 wont normally be used unless the application wishes to examine
6603 extensions in the OCSP response for example.
6605 Replace nonce routines with a pair of functions.
6606 OCSP_request_add1_nonce() adds a nonce value and optionally
6607 generates a random value. OCSP_check_nonce() checks the
6608 validity of the nonce in an OCSP response.
6611 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6612 This doesn't copy the supplied OCSP_CERTID and avoids the
6613 need to free up the newly created id. Change return type
6614 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6615 This can then be used to add extensions to the request.
6616 Deleted OCSP_request_new(), since most of its functionality
6617 is now in OCSP_REQUEST_new() (and the case insensitive name
6618 clash) apart from the ability to set the request name which
6619 will be added elsewhere.
6622 *) Update OCSP API. Remove obsolete extensions argument from
6623 various functions. Extensions are now handled using the new
6624 OCSP extension code. New simple OCSP HTTP function which
6625 can be used to send requests and parse the response.
6628 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6629 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6630 uses the special reorder version of SET OF to sort the attributes
6631 and reorder them to match the encoded order. This resolves a long
6632 standing problem: a verify on a PKCS7 structure just after signing
6633 it used to fail because the attribute order did not match the
6634 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6635 it uses the received order. This is necessary to tolerate some broken
6636 software that does not order SET OF. This is handled by encoding
6637 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6638 to produce the required SET OF.
6641 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6642 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6643 files to get correct declarations of the ASN.1 item variables.
6646 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6647 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6648 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6649 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6650 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6651 ASN1_ITEM and no wrapper functions.
6654 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6655 replace the old function pointer based I/O routines. Change most of
6656 the *_d2i_bio() and *_d2i_fp() functions to use these.
6659 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6660 lines, recognice more "algorithms" that can be deselected, and make
6661 it complain about algorithm deselection that isn't recognised.
6664 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6665 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6666 to use new functions. Add NO_ASN1_OLD which can be set to remove
6667 some old style ASN1 functions: this can be used to determine if old
6668 code will still work when these eventually go away.
6671 *) New extension functions for OCSP structures, these follow the
6672 same conventions as certificates and CRLs.
6675 *) New function X509V3_add1_i2d(). This automatically encodes and
6676 adds an extension. Its behaviour can be customised with various
6677 flags to append, replace or delete. Various wrappers added for
6678 certifcates and CRLs.
6681 *) Fix to avoid calling the underlying ASN1 print routine when
6682 an extension cannot be parsed. Correct a typo in the
6683 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6686 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
6687 entries for variables.
6690 *) Add functionality to apps/openssl.c for detecting locking
6691 problems: As the program is single-threaded, all we have
6692 to do is register a locking callback using an array for
6693 storing which locks are currently held by the program.
6696 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6697 SSL_get_ex_data_X509_STORE_idx(), which is used in
6698 ssl_verify_cert_chain() and thus can be called at any time
6699 during TLS/SSL handshakes so that thread-safety is essential.
6700 Unfortunately, the ex_data design is not at all suited
6701 for multi-threaded use, so it probably should be abolished.
6704 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6705 [Broadcom, tweaked and integrated by Geoff Thorpe]
6707 *) Move common extension printing code to new function
6708 X509V3_print_extensions(). Reorganise OCSP print routines and
6709 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6712 *) New function X509_signature_print() to remove duplication in some
6716 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6717 set (this was treated exactly the same as SET OF previously). This
6718 is used to reorder the STACK representing the structure to match the
6719 encoding. This will be used to get round a problem where a PKCS7
6720 structure which was signed could not be verified because the STACK
6721 order did not reflect the encoded order.
6724 *) Reimplement the OCSP ASN1 module using the new code.
6727 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6728 for its ASN1 operations. The old style function pointers still exist
6729 for now but they will eventually go away.
6732 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6733 completely replaces the old ASN1 functionality with a table driven
6734 encoder and decoder which interprets an ASN1_ITEM structure describing
6735 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6736 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6737 has also been converted to the new form.
6740 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6741 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6742 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6743 for negative moduli.
6746 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6747 of not touching the result's sign bit.
6750 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6754 *) Changed the LHASH code to use prototypes for callbacks, and created
6755 macros to declare and implement thin (optionally static) functions
6756 that provide type-safety and avoid function pointer casting for the
6757 type-specific callbacks.
6760 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6762 [Veers Staats <staatsvr@asc.hpc.mil>,
6763 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6765 *) Reformat the FAQ so the different questions and answers can be divided
6766 in sections depending on the subject.
6769 *) Have the zlib compression code load ZLIB.DLL dynamically under
6773 *) New function BN_mod_sqrt for computing square roots modulo a prime
6774 (using the probabilistic Tonelli-Shanks algorithm unless
6775 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6776 be handled deterministically).
6777 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6779 *) Make BN_mod_inverse faster by explicitly handling small quotients
6780 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6781 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6784 *) New function BN_kronecker.
6787 *) Fix BN_gcd so that it works on negative inputs; the result is
6788 positive unless both parameters are zero.
6789 Previously something reasonably close to an infinite loop was
6790 possible because numbers could be growing instead of shrinking
6791 in the implementation of Euclid's algorithm.
6794 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6795 sign of the number in question.
6797 Fix BN_is_word(a,w) to work correctly for w == 0.
6799 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6800 because its test if the absolute value of 'a' equals 'w'.
6801 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6802 it exists mostly for use in the implementations of BN_is_zero(),
6803 BN_is_one(), and BN_is_word().
6806 *) New function BN_swap.
6809 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6810 the exponentiation functions are more likely to produce reasonable
6811 results on negative inputs.
6814 *) Change BN_mod_mul so that the result is always non-negative.
6815 Previously, it could be negative if one of the factors was negative;
6816 I don't think anyone really wanted that behaviour.
6819 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6820 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6821 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6822 and add new functions:
6831 BN_mod_lshift1_quick
6835 These functions always generate non-negative results.
6837 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6838 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6840 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6841 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6842 be reduced modulo m.
6843 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6846 The following entry accidentily appeared in the CHANGES file
6847 distributed with OpenSSL 0.9.7. The modifications described in
6848 it do *not* apply to OpenSSL 0.9.7.
6850 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6851 was actually never needed) and in BN_mul(). The removal in BN_mul()
6852 required a small change in bn_mul_part_recursive() and the addition
6853 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6854 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6855 bn_sub_words() and bn_add_words() except they take arrays with
6860 *) In 'openssl passwd', verify passwords read from the terminal
6861 unless the '-salt' option is used (which usually means that
6862 verification would just waste user's time since the resulting
6863 hash is going to be compared with some given password hash)
6864 or the new '-noverify' option is used.
6866 This is an incompatible change, but it does not affect
6867 non-interactive use of 'openssl passwd' (passwords on the command
6868 line, '-stdin' option, '-in ...' option) and thus should not
6872 *) Remove all references to RSAref, since there's no more need for it.
6875 *) Make DSO load along a path given through an environment variable
6876 (SHLIB_PATH) with shl_load().
6879 *) Constify the ENGINE code as a result of BIGNUM constification.
6880 Also constify the RSA code and most things related to it. In a
6881 few places, most notable in the depth of the ASN.1 code, ugly
6882 casts back to non-const were required (to be solved at a later
6886 *) Make it so the openssl application has all engines loaded by default.
6889 *) Constify the BIGNUM routines a little more.
6892 *) Add the following functions:
6894 ENGINE_load_cswift()
6896 ENGINE_load_atalla()
6898 ENGINE_load_builtin_engines()
6900 That way, an application can itself choose if external engines that
6901 are built-in in OpenSSL shall ever be used or not. The benefit is
6902 that applications won't have to be linked with libdl or other dso
6903 libraries unless it's really needed.
6905 Changed 'openssl engine' to load all engines on demand.
6906 Changed the engine header files to avoid the duplication of some
6907 declarations (they differed!).
6910 *) 'openssl engine' can now list capabilities.
6913 *) Better error reporting in 'openssl engine'.
6916 *) Never call load_dh_param(NULL) in s_server.
6919 *) Add engine application. It can currently list engines by name and
6920 identity, and test if they are actually available.
6923 *) Improve RPM specification file by forcing symbolic linking and making
6924 sure the installed documentation is also owned by root.root.
6925 [Damien Miller <djm@mindrot.org>]
6927 *) Give the OpenSSL applications more possibilities to make use of
6928 keys (public as well as private) handled by engines.
6931 *) Add OCSP code that comes from CertCo.
6934 *) Add VMS support for the Rijndael code.
6937 *) Added untested support for Nuron crypto accelerator.
6940 *) Add support for external cryptographic devices. This code was
6941 previously distributed separately as the "engine" branch.
6942 [Geoff Thorpe, Richard Levitte]
6944 *) Rework the filename-translation in the DSO code. It is now possible to
6945 have far greater control over how a "name" is turned into a filename
6946 depending on the operating environment and any oddities about the
6947 different shared library filenames on each system.
6950 *) Support threads on FreeBSD-elf in Configure.
6953 *) Fix for SHA1 assembly problem with MASM: it produces
6954 warnings about corrupt line number information when assembling
6955 with debugging information. This is caused by the overlapping
6957 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6960 NCONF_get_number() has no error checking at all. As a replacement,
6961 NCONF_get_number_e() is defined (_e for "error checking") and is
6962 promoted strongly. The old NCONF_get_number is kept around for
6963 binary backward compatibility.
6964 Make it possible for methods to load from something other than a BIO,
6965 by providing a function pointer that is given a name instead of a BIO.
6966 For example, this could be used to load configuration data from an
6970 *) Fix for non blocking accept BIOs. Added new I/O special reason
6971 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6972 with non blocking I/O was not possible because no retry code was
6973 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6977 *) Added the beginnings of Rijndael support.
6980 *) Fix for bug in DirectoryString mask setting. Add support for
6981 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6982 to allow certificate printing to more controllable, additional
6983 'certopt' option to 'x509' to allow new printing options to be
6987 *) Clean old EAY MD5 hack from e_os.h.
6990 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6992 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6993 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6994 [Joe Orton, Steve Henson]
6996 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6998 *) Fix additional bug revealed by the NISCC test suite:
7000 Stop bug triggering large recursion when presented with
7001 certain ASN.1 tags (CVE-2003-0851)
7004 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7006 *) Fix various bugs revealed by running the NISCC test suite:
7008 Stop out of bounds reads in the ASN1 code when presented with
7009 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7011 If verify callback ignores invalid public key errors don't try to check
7012 certificate signature with the NULL public key.
7016 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7017 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7021 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7022 extra data after the compression methods not only for TLS 1.0
7023 but also for SSL 3.0 (as required by the specification).
7024 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7026 *) Change X509_certificate_type() to mark the key as exported/exportable
7027 when it's 512 *bits* long, not 512 bytes.
7030 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7032 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7033 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7034 a protocol version number mismatch like a decryption error
7035 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7038 *) Turn on RSA blinding by default in the default implementation
7039 to avoid a timing attack. Applications that don't want it can call
7040 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7041 They would be ill-advised to do so in most cases.
7042 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7044 *) Change RSA blinding code so that it works when the PRNG is not
7045 seeded (in this case, the secret RSA exponent is abused as
7046 an unpredictable seed -- if it is not unpredictable, there
7047 is no point in blinding anyway). Make RSA blinding thread-safe
7048 by remembering the creator's thread ID in rsa->blinding and
7049 having all other threads use local one-time blinding factors
7050 (this requires more computation than sharing rsa->blinding, but
7051 avoids excessive locking; and if an RSA object is not shared
7052 between threads, blinding will still be very fast).
7055 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7057 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7058 via timing by performing a MAC computation even if incorrrect
7059 block cipher padding has been found. This is a countermeasure
7060 against active attacks where the attacker has to distinguish
7061 between bad padding and a MAC verification error. (CVE-2003-0078)
7063 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7064 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7065 Martin Vuagnoux (EPFL, Ilion)]
7067 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7069 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7070 memory from it's contents. This is done with a counter that will
7071 place alternating values in each byte. This can be used to solve
7072 two issues: 1) the removal of calls to memset() by highly optimizing
7073 compilers, and 2) cleansing with other values than 0, since those can
7074 be read through on certain media, for example a swap space on disk.
7077 *) Bugfix: client side session caching did not work with external caching,
7078 because the session->cipher setting was not restored when reloading
7079 from the external cache. This problem was masked, when
7080 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7081 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7084 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7085 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7086 [Zeev Lieber <zeev-l@yahoo.com>]
7088 *) Undo an undocumented change introduced in 0.9.6e which caused
7089 repeated calls to OpenSSL_add_all_ciphers() and
7090 OpenSSL_add_all_digests() to be ignored, even after calling
7094 *) Change the default configuration reader to deal with last line not
7095 being properly terminated.
7098 *) Change X509_NAME_cmp() so it applies the special rules on handling
7099 DN values that are of type PrintableString, as well as RDNs of type
7100 emailAddress where the value has the type ia5String.
7101 [stefank@valicert.com via Richard Levitte]
7103 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7104 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7105 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7106 the bitwise-OR of the two for use by the majority of applications
7107 wanting this behaviour, and update the docs. The documented
7108 behaviour and actual behaviour were inconsistent and had been
7109 changing anyway, so this is more a bug-fix than a behavioural
7111 [Geoff Thorpe, diagnosed by Nadav Har'El]
7113 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7114 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7117 *) Fix initialization code race conditions in
7118 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7119 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7120 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7121 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7122 ssl2_get_cipher_by_char(),
7123 ssl3_get_cipher_by_char().
7124 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7126 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7127 the cached sessions are flushed, as the remove_cb() might use ex_data
7128 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7129 (see [openssl.org #212]).
7130 [Geoff Thorpe, Lutz Jaenicke]
7132 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7133 length, instead of the encoding length to d2i_ASN1_OBJECT.
7136 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7138 *) [In 0.9.6g-engine release:]
7139 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7140 [Lynn Gazis <lgazis@rainbow.com>]
7142 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7144 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7145 and get fix the header length calculation.
7146 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7147 Alon Kantor <alonk@checkpoint.com> (and others),
7150 *) Use proper error handling instead of 'assertions' in buffer
7151 overflow checks added in 0.9.6e. This prevents DoS (the
7152 assertions could call abort()).
7153 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7155 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7157 *) Add various sanity checks to asn1_get_length() to reject
7158 the ASN1 length bytes if they exceed sizeof(long), will appear
7159 negative or the content length exceeds the length of the
7161 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7163 *) Fix cipher selection routines: ciphers without encryption had no flags
7164 for the cipher strength set and where therefore not handled correctly
7165 by the selection routines (PR #130).
7168 *) Fix EVP_dsa_sha macro.
7172 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7173 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7174 that was added in OpenSSL 0.9.6d.
7176 As the countermeasure turned out to be incompatible with some
7177 broken SSL implementations, the new option is part of SSL_OP_ALL.
7178 SSL_OP_ALL is usually employed when compatibility with weird SSL
7179 implementations is desired (e.g. '-bugs' option to 's_client' and
7180 's_server'), so the new option is automatically set in many
7184 *) Changes in security patch:
7186 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7187 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7188 Air Force Materiel Command, USAF, under agreement number
7191 *) Add various sanity checks to asn1_get_length() to reject
7192 the ASN1 length bytes if they exceed sizeof(long), will appear
7193 negative or the content length exceeds the length of the
7194 supplied buffer. (CVE-2002-0659)
7195 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7197 *) Assertions for various potential buffer overflows, not known to
7199 [Ben Laurie (CHATS)]
7201 *) Various temporary buffers to hold ASCII versions of integers were
7202 too small for 64 bit platforms. (CVE-2002-0655)
7203 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7205 *) Remote buffer overflow in SSL3 protocol - an attacker could
7206 supply an oversized session ID to a client. (CVE-2002-0656)
7207 [Ben Laurie (CHATS)]
7209 *) Remote buffer overflow in SSL2 protocol - an attacker could
7210 supply an oversized client master key. (CVE-2002-0656)
7211 [Ben Laurie (CHATS)]
7213 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7215 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7216 encoded as NULL) with id-dsa-with-sha1.
7217 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7219 *) Check various X509_...() return values in apps/req.c.
7220 [Nils Larsch <nla@trustcenter.de>]
7222 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7223 an end-of-file condition would erronously be flagged, when the CRLF
7224 was just at the end of a processed block. The bug was discovered when
7225 processing data through a buffering memory BIO handing the data to a
7226 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7227 <ptsekov@syntrex.com> and Nedelcho Stanev.
7230 *) Implement a countermeasure against a vulnerability recently found
7231 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7232 before application data chunks to avoid the use of known IVs
7233 with data potentially chosen by the attacker.
7236 *) Fix length checks in ssl3_get_client_hello().
7239 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7240 to prevent ssl3_read_internal() from incorrectly assuming that
7241 ssl3_read_bytes() found application data while handshake
7242 processing was enabled when in fact s->s3->in_read_app_data was
7243 merely automatically cleared during the initial handshake.
7244 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7246 *) Fix object definitions for Private and Enterprise: they were not
7247 recognized in their shortname (=lowercase) representation. Extend
7248 obj_dat.pl to issue an error when using undefined keywords instead
7249 of silently ignoring the problem (Svenning Sorensen
7250 <sss@sss.dnsalias.net>).
7253 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7254 generators, i.e. generators other than 2 and 5. (Previously, the
7255 code did not properly initialise the 'add' and 'rem' values to
7256 BN_generate_prime().)
7258 In the new general case, we do not insist that 'generator' is
7259 actually a primitive root: This requirement is rather pointless;
7260 a generator of the order-q subgroup is just as good, if not
7264 *) Map new X509 verification errors to alerts. Discovered and submitted by
7265 Tom Wu <tom@arcot.com>.
7268 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7269 returning non-zero before the data has been completely received
7270 when using non-blocking I/O.
7271 [Bodo Moeller; problem pointed out by John Hughes]
7273 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7274 [Ben Laurie, Lutz Jaenicke]
7276 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7277 Yoram Zahavi <YoramZ@gilian.com>).
7280 *) Add information about CygWin 1.3 and on, and preserve proper
7281 configuration for the versions before that.
7282 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7284 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7285 check whether we deal with a copy of a session and do not delete from
7286 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7287 <izhar@checkpoint.com>.
7290 *) Do not store session data into the internal session cache, if it
7291 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7292 flag is set). Proposed by Aslam <aslam@funk.com>.
7295 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7299 *) [In 0.9.6d-engine release:]
7300 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7301 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7303 *) Add the configuration target linux-s390x.
7304 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7306 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7307 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7308 variable as an indication that a ClientHello message has been
7309 received. As the flag value will be lost between multiple
7310 invocations of ssl3_accept when using non-blocking I/O, the
7311 function may not be aware that a handshake has actually taken
7312 place, thus preventing a new session from being added to the
7315 To avoid this problem, we now set s->new_session to 2 instead of
7316 using a local variable.
7317 [Lutz Jaenicke, Bodo Moeller]
7319 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7320 if the SSL_R_LENGTH_MISMATCH error is detected.
7321 [Geoff Thorpe, Bodo Moeller]
7323 *) New 'shared_ldflag' column in Configure platform table.
7326 *) Fix EVP_CIPHER_mode macro.
7327 ["Dan S. Camper" <dan@bti.net>]
7329 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7330 type, we must throw them away by setting rr->length to 0.
7331 [D P Chang <dpc@qualys.com>]
7333 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7335 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7336 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7337 worked incorrectly for those cases where range = 10..._2 and
7338 3*range is two bits longer than range.)
7341 *) Only add signing time to PKCS7 structures if it is not already
7345 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7346 OBJ_ld_ce should be OBJ_id_ce.
7347 Also some ip-pda OIDs in crypto/objects/objects.txt were
7348 incorrect (cf. RFC 3039).
7349 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7351 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7352 returns early because it has nothing to do.
7353 [Andy Schneider <andy.schneider@bjss.co.uk>]
7355 *) [In 0.9.6c-engine release:]
7356 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7357 [Andy Schneider <andy.schneider@bjss.co.uk>]
7359 *) [In 0.9.6c-engine release:]
7360 Add support for Cryptographic Appliance's keyserver technology.
7361 (Use engine 'keyclient')
7362 [Cryptographic Appliances and Geoff Thorpe]
7364 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7365 is called via tools/c89.sh because arguments have to be
7366 rearranged (all '-L' options must appear before the first object
7368 [Richard Shapiro <rshapiro@abinitio.com>]
7370 *) [In 0.9.6c-engine release:]
7371 Add support for Broadcom crypto accelerator cards, backported
7373 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7375 *) [In 0.9.6c-engine release:]
7376 Add support for SureWare crypto accelerator cards from
7377 Baltimore Technologies. (Use engine 'sureware')
7378 [Baltimore Technologies and Mark Cox]
7380 *) [In 0.9.6c-engine release:]
7381 Add support for crypto accelerator cards from Accelerated
7382 Encryption Processing, www.aep.ie. (Use engine 'aep')
7383 [AEP Inc. and Mark Cox]
7385 *) Add a configuration entry for gcc on UnixWare.
7386 [Gary Benson <gbenson@redhat.com>]
7388 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7389 messages are stored in a single piece (fixed-length part and
7390 variable-length part combined) and fix various bugs found on the way.
7393 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7394 instead. BIO_gethostbyname() does not know what timeouts are
7395 appropriate, so entries would stay in cache even when they have
7397 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7399 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7400 faced with a pathologically small ClientHello fragment that does
7401 not contain client_version: Instead of aborting with an error,
7402 simply choose the highest available protocol version (i.e.,
7403 TLS 1.0 unless it is disabled). In practice, ClientHello
7404 messages are never sent like this, but this change gives us
7405 strictly correct behaviour at least for TLS.
7408 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7409 never resets s->method to s->ctx->method when called from within
7410 one of the SSL handshake functions.
7411 [Bodo Moeller; problem pointed out by Niko Baric]
7413 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7414 (sent using the client's version number) if client_version is
7415 smaller than the protocol version in use. Also change
7416 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7417 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7418 the client will at least see that alert.
7421 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7425 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7426 client receives HelloRequest while in a handshake.
7427 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7429 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7430 should end in 'break', not 'goto end' which circuments various
7431 cleanups done in state SSL_ST_OK. But session related stuff
7432 must be disabled for SSL_ST_OK in the case that we just sent a
7435 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7436 before just sending a HelloRequest.
7437 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7439 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7440 reveal whether illegal block cipher padding was found or a MAC
7441 verification error occured. (Neither SSLerr() codes nor alerts
7442 are directly visible to potential attackers, but the information
7443 may leak via logfiles.)
7445 Similar changes are not required for the SSL 2.0 implementation
7446 because the number of padding bytes is sent in clear for SSL 2.0,
7447 and the extra bytes are just ignored. However ssl/s2_pkt.c
7448 failed to verify that the purported number of padding bytes is in
7452 *) Add OpenUNIX-8 support including shared libraries
7453 (Boyd Lynn Gerber <gerberb@zenez.com>).
7456 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7457 'wristwatch attack' using huge encoding parameters (cf.
7458 James H. Manger's CRYPTO 2001 paper). Note that the
7459 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7460 encoding parameters and hence was not vulnerable.
7463 *) BN_sqr() bug fix.
7464 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7466 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7467 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7468 followed by modular reduction.
7469 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7471 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7472 equivalent based on BN_pseudo_rand() instead of BN_rand().
7475 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7476 This function was broken, as the check for a new client hello message
7477 to handle SGC did not allow these large messages.
7478 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7481 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7484 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7485 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7488 *) Rework the configuration and shared library support for Tru64 Unix.
7489 The configuration part makes use of modern compiler features and
7490 still retains old compiler behavior for those that run older versions
7491 of the OS. The shared library support part includes a variant that
7492 uses the RPATH feature, and is available through the special
7493 configuration target "alpha-cc-rpath", which will never be selected
7495 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7497 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7498 with the same message size as in ssl3_get_certificate_request().
7499 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7500 messages might inadvertently be reject as too long.
7501 [Petr Lampa <lampa@fee.vutbr.cz>]
7503 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7506 *) Modified SSL library such that the verify_callback that has been set
7507 specificly for an SSL object with SSL_set_verify() is actually being
7508 used. Before the change, a verify_callback set with this function was
7509 ignored and the verify_callback() set in the SSL_CTX at the time of
7510 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7511 to allow the necessary settings.
7514 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7515 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7516 done automatically (in contradiction to the requirements of the C
7517 standard). This made problems when used from OpenSSH.
7520 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7521 dh->length and always used
7523 BN_rand_range(priv_key, dh->p).
7525 BN_rand_range() is not necessary for Diffie-Hellman, and this
7526 specific range makes Diffie-Hellman unnecessarily inefficient if
7527 dh->length (recommended exponent length) is much smaller than the
7528 length of dh->p. We could use BN_rand_range() if the order of
7529 the subgroup was stored in the DH structure, but we only have
7534 BN_rand(priv_key, l, ...)
7536 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7542 RSA_eay_public_encrypt
7543 RSA_eay_private_decrypt
7544 RSA_eay_private_encrypt (signing)
7545 RSA_eay_public_decrypt (signature verification)
7547 (default implementations for RSA_public_encrypt,
7548 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7549 always reject numbers >= n.
7552 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7553 to synchronize access to 'locking_thread'. This is necessary on
7554 systems where access to 'locking_thread' (an 'unsigned long'
7555 variable) is not atomic.
7558 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7559 *before* setting the 'crypto_lock_rand' flag. The previous code had
7560 a race condition if 0 is a valid thread ID.
7561 [Travis Vitek <vitek@roguewave.com>]
7563 *) Add support for shared libraries under Irix.
7564 [Albert Chin-A-Young <china@thewrittenword.com>]
7566 *) Add configuration option to build on Linux on both big-endian and
7568 [Ralf Baechle <ralf@uni-koblenz.de>]
7570 *) Add the possibility to create shared libraries on HP-UX.
7573 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7575 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7576 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7577 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7578 PRNG state recovery was possible based on the output of
7579 one PRNG request appropriately sized to gain knowledge on
7580 'md' followed by enough consecutive 1-byte PRNG requests
7581 to traverse all of 'state'.
7583 1. When updating 'md_local' (the current thread's copy of 'md')
7584 during PRNG output generation, hash all of the previous
7585 'md_local' value, not just the half used for PRNG output.
7587 2. Make the number of bytes from 'state' included into the hash
7588 independent from the number of PRNG bytes requested.
7590 The first measure alone would be sufficient to avoid
7591 Markku-Juhani's attack. (Actually it had never occurred
7592 to me that the half of 'md_local' used for chaining was the
7593 half from which PRNG output bytes were taken -- I had always
7594 assumed that the secret half would be used.) The second
7595 measure makes sure that additional data from 'state' is never
7596 mixed into 'md_local' in small portions; this heuristically
7597 further strengthens the PRNG.
7600 *) Fix crypto/bn/asm/mips3.s.
7603 *) When only the key is given to "enc", the IV is undefined. Print out
7604 an error message in this case.
7607 *) Handle special case when X509_NAME is empty in X509 printing routines.
7610 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7611 positive and less than q.
7614 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7615 used: it isn't thread safe and the add_lock_callback should handle
7617 [Paul Rose <Paul.Rose@bridge.com>]
7619 *) Verify that incoming data obeys the block size in
7620 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7624 [Ulf Möller, Bodo Möller]
7626 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7627 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7628 when fixing the server behaviour for backwards-compatible 'client
7629 hello' messages. (Note that the attack is impractical against
7630 SSL 3.0 and TLS 1.0 anyway because length and version checking
7631 means that the probability of guessing a valid ciphertext is
7632 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7635 Before 0.9.5, the countermeasure (hide the error by generating a
7636 random 'decryption result') did not work properly because
7637 ERR_clear_error() was missing, meaning that SSL_get_error() would
7638 detect the supposedly ignored error.
7640 Both problems are now fixed.
7643 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7644 (previously it was 1024).
7647 *) Fix for compatibility mode trust settings: ignore trust settings
7648 unless some valid trust or reject settings are present.
7651 *) Fix for blowfish EVP: its a variable length cipher.
7654 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7655 parameters in DSA public key structures and return an error in the
7656 DSA routines if parameters are absent.
7659 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7660 in the current directory if neither $RANDFILE nor $HOME was set.
7661 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7662 caused some confusion to Windows users who haven't defined $HOME.
7663 Thus RAND_file_name() is changed again: e_os.h can define a
7664 DEFAULT_HOME, which will be used if $HOME is not set.
7665 For Windows, we use "C:"; on other platforms, we still require
7666 environment variables.
7668 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7669 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7670 having multiple threads call RAND_poll() concurrently.
7673 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7674 combination of a flag and a thread ID variable.
7675 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7676 flag), *other* threads can enter ssleay_add_bytes without obeying
7677 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7678 that they do not hold after the first thread unsets add_do_not_lock).
7681 *) Change bctest again: '-x' expressions are not available in all
7685 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7687 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7688 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7690 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7691 the default extension for executables, if any. Also, make the perl
7692 scripts that use symlink() to test if it really exists and use "cp"
7693 if it doesn't. All this made OpenSSL compilable and installable in
7697 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7698 If SEQUENCE is length is indefinite just set c->slen to the total
7699 amount of data available.
7700 [Steve Henson, reported by shige@FreeBSD.org]
7701 [This change does not apply to 0.9.7.]
7703 *) Change bctest to avoid here-documents inside command substitution
7704 (workaround for FreeBSD /bin/sh bug).
7705 For compatibility with Ultrix, avoid shell functions (introduced
7706 in the bctest version that searches along $PATH).
7709 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7710 with des_encrypt() defined on some operating systems, like Solaris
7714 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7715 On the Importance of Eliminating Errors in Cryptographic
7716 Computations, J. Cryptology 14 (2001) 2, 101-119,
7717 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7720 *) MIPS assembler BIGNUM division bug fix.
7723 *) Disabled incorrect Alpha assembler code.
7726 *) Fix PKCS#7 decode routines so they correctly update the length
7727 after reading an EOC for the EXPLICIT tag.
7729 [This change does not apply to 0.9.7.]
7731 *) Fix bug in PKCS#12 key generation routines. This was triggered
7732 if a 3DES key was generated with a 0 initial byte. Include
7733 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7734 (but broken) behaviour.
7737 *) Enhance bctest to search for a working bc along $PATH and print
7739 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7741 *) Fix memory leaks in err.c: free err_data string if necessary;
7742 don't write to the wrong index in ERR_set_error_data.
7745 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7749 *) Replace rdtsc with _emit statements for VC++ version 5.
7750 [Jeremy Cooper <jeremy@baymoo.org>]
7752 *) Make it possible to reuse SSLv2 sessions.
7755 *) In copy_email() check for >= 0 as a return value for
7756 X509_NAME_get_index_by_NID() since 0 is a valid index.
7757 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7759 *) Avoid coredump with unsupported or invalid public keys by checking if
7760 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7761 PKCS7_verify() fails with non detached data.
7764 *) Don't use getenv in library functions when run as setuid/setgid.
7765 New function OPENSSL_issetugid().
7768 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7769 due to incorrect handling of multi-threading:
7771 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7773 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7775 3. Count how many times MemCheck_off() has been called so that
7776 nested use can be treated correctly. This also avoids
7777 inband-signalling in the previous code (which relied on the
7778 assumption that thread ID 0 is impossible).
7781 *) Add "-rand" option also to s_client and s_server.
7784 *) Fix CPU detection on Irix 6.x.
7785 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7786 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7788 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7791 [This change does not apply to 0.9.7.]
7793 *) Use the cached encoding of an X509_NAME structure rather than
7794 copying it. This is apparently the reason for the libsafe "errors"
7795 but the code is actually correct.
7798 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7799 Bleichenbacher's DSA attack.
7800 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7801 to be set and top=0 forces the highest bit to be set; top=-1 is new
7802 and leaves the highest bit random.
7803 [Ulf Moeller, Bodo Moeller]
7805 *) In the NCONF_...-based implementations for CONF_... queries
7806 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7807 a temporary CONF structure with the data component set to NULL
7808 (which gives segmentation faults in lh_retrieve).
7809 Instead, use NULL for the CONF pointer in CONF_get_string and
7810 CONF_get_number (which may use environment variables) and directly
7811 return NULL from CONF_get_section.
7814 *) Fix potential buffer overrun for EBCDIC.
7817 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7818 keyUsage if basicConstraints absent for a CA.
7821 *) Make SMIME_write_PKCS7() write mail header values with a format that
7822 is more generally accepted (no spaces before the semicolon), since
7823 some programs can't parse those values properly otherwise. Also make
7824 sure BIO's that break lines after each write do not create invalid
7828 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7829 macros previously used would not encode an empty SEQUENCE OF
7830 and break the signature.
7832 [This change does not apply to 0.9.7.]
7834 *) Zero the premaster secret after deriving the master secret in
7838 *) Add some EVP_add_digest_alias registrations (as found in
7839 OpenSSL_add_all_digests()) to SSL_library_init()
7840 aka OpenSSL_add_ssl_algorithms(). This provides improved
7841 compatibility with peers using X.509 certificates
7842 with unconventional AlgorithmIdentifier OIDs.
7845 *) Fix for Irix with NO_ASM.
7846 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7848 *) ./config script fixes.
7849 [Ulf Moeller, Richard Levitte]
7851 *) Fix 'openssl passwd -1'.
7854 *) Change PKCS12_key_gen_asc() so it can cope with non null
7855 terminated strings whose length is passed in the passlen
7856 parameter, for example from PEM callbacks. This was done
7857 by adding an extra length parameter to asc2uni().
7858 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7860 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7861 call failed, free the DSA structure.
7864 *) Fix to uni2asc() to cope with zero length Unicode strings.
7865 These are present in some PKCS#12 files.
7868 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7869 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7870 when writing a 32767 byte record.
7871 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7873 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7874 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7876 (RSA objects have a reference count access to which is protected
7877 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7878 so they are meant to be shared between threads.)
7879 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7880 "Reddie, Steven" <Steven.Reddie@ca.com>]
7882 *) Fix a deadlock in CRYPTO_mem_leaks().
7885 *) Use better test patterns in bntest.
7888 *) rand_win.c fix for Borland C.
7891 *) BN_rshift bugfix for n == 0.
7894 *) Add a 'bctest' script that checks for some known 'bc' bugs
7895 so that 'make test' does not abort just because 'bc' is broken.
7898 *) Store verify_result within SSL_SESSION also for client side to
7899 avoid potential security hole. (Re-used sessions on the client side
7900 always resulted in verify_result==X509_V_OK, not using the original
7901 result of the server certificate verification.)
7904 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7905 SSL3_RT_APPLICATION_DATA, return 0.
7906 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7910 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7911 releases, have been re-implemented by renaming the previous
7912 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7913 and ssl3_read_internal, respectively, and adding 'peek' parameters
7914 to them. The new ssl[23]_{read,peek} functions are calls to
7915 ssl[23]_read_internal with the 'peek' flag set appropriately.
7916 A 'peek' parameter has also been added to ssl3_read_bytes, which
7917 does the actual work for ssl3_read_internal.
7920 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7921 the method-specific "init()" handler. Also clean up ex_data after
7922 calling the method-specific "finish()" handler. Previously, this was
7923 happening the other way round.
7926 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7927 The previous value, 12, was not always sufficient for BN_mod_exp().
7930 *) Make sure that shared libraries get the internal name engine with
7931 the full version number and not just 0. This should mark the
7932 shared libraries as not backward compatible. Of course, this should
7933 be changed again when we can guarantee backward binary compatibility.
7936 *) Fix typo in get_cert_by_subject() in by_dir.c
7937 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7939 *) Rework the system to generate shared libraries:
7941 - Make note of the expected extension for the shared libraries and
7942 if there is a need for symbolic links from for example libcrypto.so.0
7943 to libcrypto.so.0.9.7. There is extended info in Configure for
7946 - Make as few rebuilds of the shared libraries as possible.
7948 - Still avoid linking the OpenSSL programs with the shared libraries.
7950 - When installing, install the shared libraries separately from the
7954 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7956 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7957 and not in SSL_clear because the latter is also used by the
7958 accept/connect functions; previously, the settings made by
7959 SSL_set_read_ahead would be lost during the handshake.
7960 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7962 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7963 Previously, it would create entries for disableed algorithms no
7967 *) Added several new manual pages for SSL_* function.
7970 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7972 *) In ssl23_get_client_hello, generate an error message when faced
7973 with an initial SSL 3.0/TLS record that is too small to contain the
7974 first two bytes of the ClientHello message, i.e. client_version.
7975 (Note that this is a pathologic case that probably has never happened
7976 in real life.) The previous approach was to use the version number
7977 from the record header as a substitute; but our protocol choice
7978 should not depend on that one because it is not authenticated
7979 by the Finished messages.
7982 *) More robust randomness gathering functions for Windows.
7983 [Jeffrey Altman <jaltman@columbia.edu>]
7985 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7986 not set then we don't setup the error code for issuer check errors
7987 to avoid possibly overwriting other errors which the callback does
7988 handle. If an application does set the flag then we assume it knows
7989 what it is doing and can handle the new informational codes
7993 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7994 a general "ANY" type, as such it should be able to decode anything
7995 including tagged types. However it didn't check the class so it would
7996 wrongly interpret tagged types in the same way as their universal
7997 counterpart and unknown types were just rejected. Changed so that the
7998 tagged and unknown types are handled in the same way as a SEQUENCE:
7999 that is the encoding is stored intact. There is also a new type
8000 "V_ASN1_OTHER" which is used when the class is not universal, in this
8001 case we have no idea what the actual type is so we just lump them all
8005 *) On VMS, stdout may very well lead to a file that is written to
8006 in a record-oriented fashion. That means that every write() will
8007 write a separate record, which will be read separately by the
8008 programs trying to read from it. This can be very confusing.
8010 The solution is to put a BIO filter in the way that will buffer
8011 text until a linefeed is reached, and then write everything a
8012 line at a time, so every record written will be an actual line,
8013 not chunks of lines and not (usually doesn't happen, but I've
8014 seen it once) several lines in one record. BIO_f_linebuffer() is
8017 Currently, it's a VMS-only method, because that's where it has
8018 been tested well enough.
8021 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8022 it can return incorrect results.
8023 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8024 but it was in 0.9.6-beta[12].)
8027 *) Disable the check for content being present when verifying detached
8028 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8029 include zero length content when signing messages.
8032 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8033 BIO_ctrl (for BIO pairs).
8036 *) Add DSO method for VMS.
8039 *) Bug fix: Montgomery multiplication could produce results with the
8043 *) Add RPM specification openssl.spec and modify it to build three
8044 packages. The default package contains applications, application
8045 documentation and run-time libraries. The devel package contains
8046 include files, static libraries and function documentation. The
8047 doc package contains the contents of the doc directory. The original
8048 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8051 *) Add a large number of documentation files for many SSL routines.
8052 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8054 *) Add a configuration entry for Sony News 4.
8055 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8057 *) Don't set the two most significant bits to one when generating a
8058 random number < q in the DSA library.
8061 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8062 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8063 the underlying transport is blocking) if a handshake took place.
8064 (The default behaviour is needed by applications such as s_client
8065 and s_server that use select() to determine when to use SSL_read;
8066 but for applications that know in advance when to expect data, it
8067 just makes things more complicated.)
8070 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8074 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8075 work better on such systems.
8076 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8078 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8079 Update PKCS12_parse() so it copies the friendlyName and the
8080 keyid to the certificates aux info.
8083 *) Fix bug in PKCS7_verify() which caused an infinite loop
8084 if there was more than one signature.
8085 [Sven Uszpelkat <su@celocom.de>]
8087 *) Major change in util/mkdef.pl to include extra information
8088 about each symbol, as well as presentig variables as well
8089 as functions. This change means that there's n more need
8090 to rebuild the .num files when some algorithms are excluded.
8093 *) Allow the verify time to be set by an application,
8094 rather than always using the current time.
8097 *) Phase 2 verify code reorganisation. The certificate
8098 verify code now looks up an issuer certificate by a
8099 number of criteria: subject name, authority key id
8100 and key usage. It also verifies self signed certificates
8101 by the same criteria. The main comparison function is
8102 X509_check_issued() which performs these checks.
8104 Lot of changes were necessary in order to support this
8105 without completely rewriting the lookup code.
8107 Authority and subject key identifier are now cached.
8109 The LHASH 'certs' is X509_STORE has now been replaced
8110 by a STACK_OF(X509_OBJECT). This is mainly because an
8111 LHASH can't store or retrieve multiple objects with
8112 the same hash value.
8114 As a result various functions (which were all internal
8115 use only) have changed to handle the new X509_STORE
8116 structure. This will break anything that messed round
8117 with X509_STORE internally.
8119 The functions X509_STORE_add_cert() now checks for an
8120 exact match, rather than just subject name.
8122 The X509_STORE API doesn't directly support the retrieval
8123 of multiple certificates matching a given criteria, however
8124 this can be worked round by performing a lookup first
8125 (which will fill the cache with candidate certificates)
8126 and then examining the cache for matches. This is probably
8127 the best we can do without throwing out X509_LOOKUP
8128 entirely (maybe later...).
8130 The X509_VERIFY_CTX structure has been enhanced considerably.
8132 All certificate lookup operations now go via a get_issuer()
8133 callback. Although this currently uses an X509_STORE it
8134 can be replaced by custom lookups. This is a simple way
8135 to bypass the X509_STORE hackery necessary to make this
8136 work and makes it possible to use more efficient techniques
8137 in future. A very simple version which uses a simple
8138 STACK for its trusted certificate store is also provided
8139 using X509_STORE_CTX_trusted_stack().
8141 The verify_cb() and verify() callbacks now have equivalents
8142 in the X509_STORE_CTX structure.
8144 X509_STORE_CTX also has a 'flags' field which can be used
8145 to customise the verify behaviour.
8148 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8149 excludes S/MIME capabilities.
8152 *) When a certificate request is read in keep a copy of the
8153 original encoding of the signed data and use it when outputing
8154 again. Signatures then use the original encoding rather than
8155 a decoded, encoded version which may cause problems if the
8156 request is improperly encoded.
8159 *) For consistency with other BIO_puts implementations, call
8160 buffer_write(b, ...) directly in buffer_puts instead of calling
8163 In BIO_puts, increment b->num_write as in BIO_write.
8164 [Peter.Sylvester@EdelWeb.fr]
8166 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8167 BN_zero, we may not return a BIGNUM with an array consisting of
8171 *) Avoid calling abort() from within the library when problems are
8172 detected, except if preprocessor symbols have been defined
8173 (such as REF_CHECK, BN_DEBUG etc.).
8176 *) New openssl application 'rsautl'. This utility can be
8177 used for low level RSA operations. DER public key
8178 BIO/fp routines also added.
8181 *) New Configure entry and patches for compiling on QNX 4.
8182 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8184 *) A demo state-machine implementation was sponsored by
8185 Nuron (http://www.nuron.com/) and is now available in
8186 demos/state_machine.
8189 *) New options added to the 'dgst' utility for signature
8190 generation and verification.
8193 *) Unrecognized PKCS#7 content types are now handled via a
8194 catch all ASN1_TYPE structure. This allows unsupported
8195 types to be stored as a "blob" and an application can
8196 encode and decode it manually.
8199 *) Fix various signed/unsigned issues to make a_strex.c
8201 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8203 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8204 length if passed a buffer. ASN1_INTEGER_to_BN failed
8205 if passed a NULL BN and its argument was negative.
8206 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8208 *) Modification to PKCS#7 encoding routines to output definite
8209 length encoding. Since currently the whole structures are in
8210 memory there's not real point in using indefinite length
8211 constructed encoding. However if OpenSSL is compiled with
8212 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8215 *) Added BIO_vprintf() and BIO_vsnprintf().
8218 *) Added more prefixes to parse for in the the strings written
8219 through a logging bio, to cover all the levels that are available
8220 through syslog. The prefixes are now:
8222 PANIC, EMERG, EMR => LOG_EMERG
8223 ALERT, ALR => LOG_ALERT
8224 CRIT, CRI => LOG_CRIT
8225 ERROR, ERR => LOG_ERR
8226 WARNING, WARN, WAR => LOG_WARNING
8227 NOTICE, NOTE, NOT => LOG_NOTICE
8228 INFO, INF => LOG_INFO
8229 DEBUG, DBG => LOG_DEBUG
8231 and as before, if none of those prefixes are present at the
8232 beginning of the string, LOG_ERR is chosen.
8234 On Win32, the LOG_* levels are mapped according to this:
8236 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8237 LOG_WARNING => EVENTLOG_WARNING_TYPE
8238 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8242 *) Made it possible to reconfigure with just the configuration
8243 argument "reconf" or "reconfigure". The command line arguments
8244 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8245 and are retrieved from there when reconfiguring.
8249 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8251 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8254 *) The obj_dat.pl script was messing up the sorting of object
8255 names. The reason was that it compared the quoted version
8256 of strings as a result "OCSP" > "OCSP Signing" because
8257 " > SPACE. Changed script to store unquoted versions of
8258 names and add quotes on output. It was also omitting some
8259 names from the lookup table if they were given a default
8260 value (that is if SN is missing it is given the same
8261 value as LN and vice versa), these are now added on the
8262 grounds that if an object has a name we should be able to
8263 look it up. Finally added warning output when duplicate
8264 short or long names are found.
8267 *) Changes needed for Tandem NSK.
8268 [Scott Uroff <scott@xypro.com>]
8270 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8271 RSA_padding_check_SSLv23(), special padding was never detected
8272 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8273 version rollback attacks was not effective.
8275 In s23_clnt.c, don't use special rollback-attack detection padding
8276 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8277 client; similarly, in s23_srvr.c, don't do the rollback check if
8278 SSL 2.0 is the only protocol enabled in the server.
8281 *) Make it possible to get hexdumps of unprintable data with 'openssl
8282 asn1parse'. By implication, the functions ASN1_parse_dump() and
8283 BIO_dump_indent() are added.
8286 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8287 these print out strings and name structures based on various
8288 flags including RFC2253 support and proper handling of
8289 multibyte characters. Added options to the 'x509' utility
8290 to allow the various flags to be set.
8293 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8294 Also change the functions X509_cmp_current_time() and
8295 X509_gmtime_adj() work with an ASN1_TIME structure,
8296 this will enable certificates using GeneralizedTime in validity
8297 dates to be checked.
8300 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8301 negative public key encodings) on by default,
8302 NO_NEG_PUBKEY_BUG can be set to disable it.
8305 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8306 content octets. An i2c_ASN1_OBJECT is unnecessary because
8307 the encoding can be trivially obtained from the structure.
8310 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8311 not read locks (CRYPTO_r_[un]lock).
8314 *) A first attempt at creating official support for shared
8315 libraries through configuration. I've kept it so the
8316 default is static libraries only, and the OpenSSL programs
8317 are always statically linked for now, but there are
8318 preparations for dynamic linking in place.
8319 This has been tested on Linux and Tru64.
8322 *) Randomness polling function for Win9x, as described in:
8323 Peter Gutmann, Software Generation of Practically Strong
8327 *) Fix so PRNG is seeded in req if using an already existing
8331 *) New options to smime application. -inform and -outform
8332 allow alternative formats for the S/MIME message including
8333 PEM and DER. The -content option allows the content to be
8334 specified separately. This should allow things like Netscape
8335 form signing output easier to verify.
8338 *) Fix the ASN1 encoding of tags using the 'long form'.
8341 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8342 STRING types. These convert content octets to and from the
8343 underlying type. The actual tag and length octets are
8344 already assumed to have been read in and checked. These
8345 are needed because all other string types have virtually
8346 identical handling apart from the tag. By having versions
8347 of the ASN1 functions that just operate on content octets
8348 IMPLICIT tagging can be handled properly. It also allows
8349 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8350 and ASN1_INTEGER are identical apart from the tag.
8353 *) Change the handling of OID objects as follows:
8355 - New object identifiers are inserted in objects.txt, following
8356 the syntax given in objects.README.
8357 - objects.pl is used to process obj_mac.num and create a new
8359 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8362 This is currently kind of a hack, and the perl code in objects.pl
8363 isn't very elegant, but it works as I intended. The simplest way
8364 to check that it worked correctly is to look in obj_dat.h and
8365 check the array nid_objs and make sure the objects haven't moved
8366 around (this is important!). Additions are OK, as well as
8367 consistent name changes.
8370 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8373 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8374 The given file adds to whatever has already been seeded into the
8375 random pool through the RANDFILE configuration file option or
8376 environment variable, or the default random state file.
8379 *) mkstack.pl now sorts each macro group into lexical order.
8380 Previously the output order depended on the order the files
8381 appeared in the directory, resulting in needless rewriting
8385 *) Patches to make OpenSSL compile under Win32 again. Mostly
8386 work arounds for the VC++ problem that it treats func() as
8387 func(void). Also stripped out the parts of mkdef.pl that
8388 added extra typesafe functions: these no longer exist.
8391 *) Reorganisation of the stack code. The macros are now all
8392 collected in safestack.h . Each macro is defined in terms of
8393 a "stack macro" of the form SKM_<name>(type, a, b). The
8394 DEBUG_SAFESTACK is now handled in terms of function casts,
8395 this has the advantage of retaining type safety without the
8396 use of additional functions. If DEBUG_SAFESTACK is not defined
8397 then the non typesafe macros are used instead. Also modified the
8398 mkstack.pl script to handle the new form. Needs testing to see
8399 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8400 the default if no major problems. Similar behaviour for ASN1_SET_OF
8401 and PKCS12_STACK_OF.
8404 *) When some versions of IIS use the 'NET' form of private key the
8405 key derivation algorithm is different. Normally MD5(password) is
8406 used as a 128 bit RC4 key. In the modified case
8407 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
8408 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8409 as the old Netscape_RSA functions except they have an additional
8410 'sgckey' parameter which uses the modified algorithm. Also added
8411 an -sgckey command line option to the rsa utility. Thanks to
8412 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8413 algorithm to openssl-dev.
8416 *) The evp_local.h macros were using 'c.##kname' which resulted in
8417 invalid expansion on some systems (SCO 5.0.5 for example).
8418 Corrected to 'c.kname'.
8419 [Phillip Porch <root@theporch.com>]
8421 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8422 a STACK of email addresses from a certificate or request, these look
8423 in the subject name and the subject alternative name extensions and
8424 omit any duplicate addresses.
8427 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8428 This makes DSA verification about 2 % faster.
8431 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8432 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8433 plus overhead for 1024 bit moduli).
8434 This makes exponentiations about 0.5 % faster for 1024 bit
8435 exponents (as measured by "openssl speed rsa2048").
8438 *) Rename memory handling macros to avoid conflicts with other
8440 Malloc => OPENSSL_malloc
8441 Malloc_locked => OPENSSL_malloc_locked
8442 Realloc => OPENSSL_realloc
8443 Free => OPENSSL_free
8446 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8447 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8450 *) CygWin32 support.
8451 [John Jarvie <jjarvie@newsguy.com>]
8453 *) The type-safe stack code has been rejigged. It is now only compiled
8454 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8455 by default all type-specific stack functions are "#define"d back to
8456 standard stack functions. This results in more streamlined output
8457 but retains the type-safety checking possibilities of the original
8461 *) The STACK code has been cleaned up, and certain type declarations
8462 that didn't make a lot of sense have been brought in line. This has
8463 also involved a cleanup of sorts in safestack.h to more correctly
8464 map type-safe stack functions onto their plain stack counterparts.
8465 This work has also resulted in a variety of "const"ifications of
8466 lots of the code, especially "_cmp" operations which should normally
8467 be prototyped with "const" parameters anyway.
8470 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8471 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8472 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8473 where all of 'md' is used each time the PRNG is used, but 'state'
8474 is used only indexed by a cyclic counter. As entropy may not be
8475 well distributed from the beginning, 'md' is important as a
8476 chaining variable. However, the output function chains only half
8477 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8478 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8479 in all of 'state' being rewritten, with the new values depending
8480 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8483 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8484 the handshake is continued after ssl_verify_cert_chain();
8485 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8486 can lead to 'unexplainable' connection aborts later.
8487 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8489 *) Major EVP API cipher revision.
8490 Add hooks for extra EVP features. This allows various cipher
8491 parameters to be set in the EVP interface. Support added for variable
8492 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8493 setting of RC2 and RC5 parameters.
8495 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8498 Remove lots of duplicated code from the EVP library. For example *every*
8499 cipher init() function handles the 'iv' in the same way according to the
8500 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8501 for CFB and OFB modes they zero ctx->num.
8503 New functionality allows removal of S/MIME code RC2 hack.
8505 Most of the routines have the same form and so can be declared in terms
8508 By shifting this to the top level EVP_CipherInit() it can be removed from
8509 all individual ciphers. If the cipher wants to handle IVs or keys
8510 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8513 Change lots of functions like EVP_EncryptUpdate() to now return a
8514 value: although software versions of the algorithms cannot fail
8515 any installed hardware versions can.
8518 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8519 this option is set, tolerate broken clients that send the negotiated
8520 protocol version number instead of the requested protocol version
8524 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8525 i.e. non-zero for export ciphersuites, zero otherwise.
8526 Previous versions had this flag inverted, inconsistent with
8527 rsa_tmp_cb (..._TMP_RSA_CB).
8528 [Bodo Moeller; problem reported by Amit Chopra]
8530 *) Add missing DSA library text string. Work around for some IIS
8531 key files with invalid SEQUENCE encoding.
8534 *) Add a document (doc/standards.txt) that list all kinds of standards
8535 and so on that are implemented in OpenSSL.
8538 *) Enhance c_rehash script. Old version would mishandle certificates
8539 with the same subject name hash and wouldn't handle CRLs at all.
8540 Added -fingerprint option to crl utility, to support new c_rehash
8544 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8547 *) Fix for SSL server purpose checking. Server checking was
8548 rejecting certificates which had extended key usage present
8549 but no ssl client purpose.
8550 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8552 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8553 is a little unclear about how a blank password is handled.
8554 Since the password in encoded as a BMPString with terminating
8555 double NULL a zero length password would end up as just the
8556 double NULL. However no password at all is different and is
8557 handled differently in the PKCS#12 key generation code. NS
8558 treats a blank password as zero length. MSIE treats it as no
8559 password on export: but it will try both on import. We now do
8560 the same: PKCS12_parse() tries zero length and no password if
8561 the password is set to "" or NULL (NULL is now a valid password:
8562 it wasn't before) as does the pkcs12 application.
8565 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8566 perror when PEM_read_bio_X509_REQ fails, the error message must
8567 be obtained from the error queue.
8570 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8571 it in ERR_remove_state if appropriate, and change ERR_get_state
8572 accordingly to avoid race conditions (this is necessary because
8573 thread_hash is no longer constant once set).
8576 *) Bugfix for linux-elf makefile.one.
8579 *) RSA_get_default_method() will now cause a default
8580 RSA_METHOD to be chosen if one doesn't exist already.
8581 Previously this was only set during a call to RSA_new()
8582 or RSA_new_method(NULL) meaning it was possible for
8583 RSA_get_default_method() to return NULL.
8586 *) Added native name translation to the existing DSO code
8587 that will convert (if the flag to do so is set) filenames
8588 that are sufficiently small and have no path information
8589 into a canonical native form. Eg. "blah" converted to
8590 "libblah.so" or "blah.dll" etc.
8593 *) New function ERR_error_string_n(e, buf, len) which is like
8594 ERR_error_string(e, buf), but writes at most 'len' bytes
8595 including the 0 terminator. For ERR_error_string_n, 'buf'
8597 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8599 *) CONF library reworked to become more general. A new CONF
8600 configuration file reader "class" is implemented as well as a
8601 new functions (NCONF_*, for "New CONF") to handle it. The now
8602 old CONF_* functions are still there, but are reimplemented to
8603 work in terms of the new functions. Also, a set of functions
8604 to handle the internal storage of the configuration data is
8605 provided to make it easier to write new configuration file
8606 reader "classes" (I can definitely see something reading a
8607 configuration file in XML format, for example), called _CONF_*,
8608 or "the configuration storage API"...
8610 The new configuration file reading functions are:
8612 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8613 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8615 NCONF_default, NCONF_WIN32
8617 NCONF_dump_fp, NCONF_dump_bio
8619 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8620 NCONF_new creates a new CONF object. This works in the same way
8621 as other interfaces in OpenSSL, like the BIO interface.
8622 NCONF_dump_* dump the internal storage of the configuration file,
8623 which is useful for debugging. All other functions take the same
8624 arguments as the old CONF_* functions wth the exception of the
8625 first that must be a `CONF *' instead of a `LHASH *'.
8627 To make it easer to use the new classes with the old CONF_* functions,
8628 the function CONF_set_default_method is provided.
8631 *) Add '-tls1' option to 'openssl ciphers', which was already
8632 mentioned in the documentation but had not been implemented.
8633 (This option is not yet really useful because even the additional
8634 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8637 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8638 OpenSSL-based applications) load shared libraries and bind to
8639 them in a portable way.
8640 [Geoff Thorpe, with contributions from Richard Levitte]
8642 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8644 *) Make sure _lrotl and _lrotr are only used with MSVC.
8646 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8647 (the default implementation of RAND_status).
8649 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8650 to '-clrext' (= clear extensions), as intended and documented.
8651 [Bodo Moeller; inconsistency pointed out by Michael Attili
8654 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8655 was larger than the MD block size.
8656 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8658 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8659 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8660 using the passed key: if the passed key was a private key the result
8661 of X509_print(), for example, would be to print out all the private key
8665 *) des_quad_cksum() byte order bug fix.
8666 [Ulf Möller, using the problem description in krb4-0.9.7, where
8667 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8669 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8671 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8673 *) For easily testing in shell scripts whether some command
8674 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8675 returns with exit code 0 iff no command of the given name is available.
8676 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8677 the output goes to stdout and nothing is printed to stderr.
8678 Additional arguments are always ignored.
8680 Since for each cipher there is a command of the same name,
8681 the 'no-cipher' compilation switches can be tested this way.
8683 ('openssl no-XXX' is not able to detect pseudo-commands such
8684 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8687 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8690 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8691 is set; it will be thrown away anyway because each handshake creates
8693 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8694 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8695 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
8696 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8699 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8700 'Q' and 'R' lose their special meanings (quit/renegotiate).
8701 This is part of what -quiet does; unlike -quiet, -ign_eof
8702 does not suppress any output.
8705 *) Add compatibility options to the purpose and trust code. The
8706 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8707 accepts a certificate or CA, this was the previous behaviour,
8708 with all the associated security issues.
8710 X509_TRUST_COMPAT is the old trust behaviour: only and
8711 automatically trust self signed roots in certificate store. A
8712 new trust setting X509_TRUST_DEFAULT is used to specify that
8713 a purpose has no associated trust setting and it should instead
8714 use the value in the default purpose.
8717 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8718 and fix a memory leak.
8721 *) In util/mkerr.pl (which implements 'make errors'), preserve
8722 reason strings from the previous version of the .c file, as
8723 the default to have only downcase letters (and digits) in
8724 automatically generated reasons codes is not always appropriate.
8727 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8728 using strerror. Previously, ERR_reason_error_string() returned
8729 library names as reason strings for SYSerr; but SYSerr is a special
8730 case where small numbers are errno values, not library numbers.
8733 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8734 converts DSA parameters into DH parameters. (When creating parameters,
8735 DSA_generate_parameters is used.)
8738 *) Include 'length' (recommended exponent length) in C code generated
8739 by 'openssl dhparam -C'.
8742 *) The second argument to set_label in perlasm was already being used
8743 so couldn't be used as a "file scope" flag. Moved to third argument
8747 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8748 instead of RAND_bytes for encryption IVs and salts.
8751 *) Include RAND_status() into RAND_METHOD instead of implementing
8752 it only for md_rand.c Otherwise replacing the PRNG by calling
8753 RAND_set_rand_method would be impossible.
8756 *) Don't let DSA_generate_key() enter an infinite loop if the random
8757 number generation fails.
8760 *) New 'rand' application for creating pseudo-random output.
8763 *) Added configuration support for Linux/IA64
8764 [Rolf Haberrecker <rolf@suse.de>]
8766 *) Assembler module support for Mingw32.
8769 *) Shared library support for HPUX (in shlib/).
8770 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8772 *) Shared library support for Solaris gcc.
8773 [Lutz Behnke <behnke@trustcenter.de>]
8775 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8777 *) PKCS7_encrypt() was adding text MIME headers twice because they
8778 were added manually and by SMIME_crlf_copy().
8781 *) In bntest.c don't call BN_rand with zero bits argument.
8782 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8784 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8785 case was implemented. This caused BN_div_recp() to fail occasionally.
8788 *) Add an optional second argument to the set_label() in the perl
8789 assembly language builder. If this argument exists and is set
8790 to 1 it signals that the assembler should use a symbol whose
8791 scope is the entire file, not just the current function. This
8792 is needed with MASM which uses the format label:: for this scope.
8793 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8795 *) Change the ASN1 types so they are typedefs by default. Before
8796 almost all types were #define'd to ASN1_STRING which was causing
8797 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8801 *) Change names of new functions to the new get1/get0 naming
8802 convention: After 'get1', the caller owns a reference count
8803 and has to call ..._free; 'get0' returns a pointer to some
8804 data structure without incrementing reference counters.
8805 (Some of the existing 'get' functions increment a reference
8806 counter, some don't.)
8807 Similarly, 'set1' and 'add1' functions increase reference
8808 counters or duplicate objects.
8811 *) Allow for the possibility of temp RSA key generation failure:
8812 the code used to assume it always worked and crashed on failure.
8815 *) Fix potential buffer overrun problem in BIO_printf().
8816 [Ulf Möller, using public domain code by Patrick Powell; problem
8817 pointed out by David Sacerdote <das33@cornell.edu>]
8819 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8820 RAND_egd() and RAND_status(). In the command line application,
8821 the EGD socket can be specified like a seed file using RANDFILE
8825 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8826 Some CAs (e.g. Verisign) distribute certificates in this form.
8829 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8830 list to exclude them. This means that no special compilation option
8831 is needed to use anonymous DH: it just needs to be included in the
8835 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8836 EVP_MD_type. The old functionality is available in a new macro called
8837 EVP_MD_md(). Change code that uses it and update docs.
8840 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8841 where the 'void *' argument is replaced by a function pointer argument.
8842 Previously 'void *' was abused to point to functions, which works on
8843 many platforms, but is not correct. As these functions are usually
8844 called by macros defined in OpenSSL header files, most source code
8845 should work without changes.
8848 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8849 sections with information on -D... compiler switches used for
8850 compiling the library so that applications can see them. To enable
8851 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8852 must be defined. E.g.,
8853 #define OPENSSL_ALGORITHM_DEFINES
8854 #include <openssl/opensslconf.h>
8855 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8856 [Richard Levitte, Ulf and Bodo Möller]
8858 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8862 *) Change the 'other' type in certificate aux info to a STACK_OF
8863 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8864 the required ASN1 format: arbitrary types determined by an OID.
8867 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8868 argument to 'req'. This is not because the function is newer or
8869 better than others it just uses the work 'NEW' in the certificate
8870 request header lines. Some software needs this.
8873 *) Reorganise password command line arguments: now passwords can be
8874 obtained from various sources. Delete the PEM_cb function and make
8875 it the default behaviour: i.e. if the callback is NULL and the
8876 usrdata argument is not NULL interpret it as a null terminated pass
8877 phrase. If usrdata and the callback are NULL then the pass phrase
8878 is prompted for as usual.
8881 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8882 the support is automatically enabled. The resulting binaries will
8883 autodetect the card and use it if present.
8884 [Ben Laurie and Compaq Inc.]
8886 *) Work around for Netscape hang bug. This sends certificate request
8887 and server done in one record. Since this is perfectly legal in the
8888 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8889 the bugs/SSLv3 entry for more info.
8892 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8895 *) Add -rand argument to smime and pkcs12 applications and read/write
8899 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8902 *) Add command line password options to the remaining applications.
8905 *) Bug fix for BN_div_recp() for numerators with an even number of
8909 *) More tests in bntest.c, and changed test_bn output.
8912 *) ./config recognizes MacOS X now.
8915 *) Bug fix for BN_div() when the first words of num and divsor are
8916 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8919 *) Add support for various broken PKCS#8 formats, and command line
8920 options to produce them.
8923 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8924 get temporary BIGNUMs from a BN_CTX.
8927 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8931 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8932 include a #define from the old name to the new. The original intent
8933 was that statically linked binaries could for example just call
8934 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8935 link with digests. This never worked becayse SSLeay_add_all_digests()
8936 and SSLeay_add_all_ciphers() were in the same source file so calling
8937 one would link with the other. They are now in separate source files.
8940 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8943 *) Use a less unusual form of the Miller-Rabin primality test (it used
8944 a binary algorithm for exponentiation integrated into the Miller-Rabin
8945 loop, our standard modexp algorithms are faster).
8948 *) Support for the EBCDIC character set completed.
8949 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8951 *) Source code cleanups: use const where appropriate, eliminate casts,
8952 use void * instead of char * in lhash.
8955 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8956 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8957 this the server could overwrite ephemeral keys that the client
8961 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8962 using 50 iterations of the Rabin-Miller test.
8964 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8965 iterations of the Rabin-Miller test as required by the appendix
8966 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8967 As BN_is_prime_fasttest includes trial division, DSA parameter
8968 generation becomes much faster.
8970 This implies a change for the callback functions in DSA_is_prime
8971 and DSA_generate_parameters: The callback function is called once
8972 for each positive witness in the Rabin-Miller test, not just
8973 occasionally in the inner loop; and the parameters to the
8974 callback function now provide an iteration count for the outer
8975 loop rather than for the current invocation of the inner loop.
8976 DSA_generate_parameters additionally can call the callback
8977 function with an 'iteration count' of -1, meaning that a
8978 candidate has passed the trial division test (when q is generated
8979 from an application-provided seed, trial division is skipped).
8982 *) New function BN_is_prime_fasttest that optionally does trial
8983 division before starting the Rabin-Miller test and has
8984 an additional BN_CTX * argument (whereas BN_is_prime always
8985 has to allocate at least one BN_CTX).
8986 'callback(1, -1, cb_arg)' is called when a number has passed the
8987 trial division stage.
8990 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8994 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8997 *) New function BN_pseudo_rand().
9000 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9001 bignum version of BN_from_montgomery() with the working code from
9002 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9006 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9007 made it impossible to use the same SSL_SESSION data structure in
9008 SSL2 clients in multiple threads.
9011 *) The return value of RAND_load_file() no longer counts bytes obtained
9012 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9013 to seed the PRNG (previously an explicit byte count was required).
9014 [Ulf Möller, Bodo Möller]
9016 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9017 used (char *) instead of (void *) and had casts all over the place.
9020 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9023 *) Retain source code compatibility for BN_prime_checks macro:
9024 BN_is_prime(..., BN_prime_checks, ...) now uses
9025 BN_prime_checks_for_size to determine the appropriate number of
9026 Rabin-Miller iterations.
9029 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9030 DH_CHECK_P_NOT_SAFE_PRIME.
9031 (Check if this is true? OpenPGP calls them "strong".)
9034 *) Merge the functionality of "dh" and "gendh" programs into a new program
9035 "dhparam". The old programs are retained for now but will handle DH keys
9036 (instead of parameters) in future.
9039 *) Make the ciphers, s_server and s_client programs check the return values
9040 when a new cipher list is set.
9043 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9044 ciphers. Before when the 56bit ciphers were enabled the sorting was
9047 The syntax for the cipher sorting has been extended to support sorting by
9048 cipher-strength (using the strength_bits hard coded in the tables).
9049 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9051 Fix a bug in the cipher-command parser: when supplying a cipher command
9052 string with an "undefined" symbol (neither command nor alphanumeric
9053 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9054 an error is flagged.
9056 Due to the strength-sorting extension, the code of the
9057 ssl_create_cipher_list() function was completely rearranged. I hope that
9058 the readability was also increased :-)
9059 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9061 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9062 for the first serial number and places 2 in the serial number file. This
9063 avoids problems when the root CA is created with serial number zero and
9064 the first user certificate has the same issuer name and serial number
9068 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9069 the new code. Add documentation for this stuff.
9072 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9073 X509_*() to X509at_*() on the grounds that they don't handle X509
9074 structures and behave in an analagous way to the X509v3 functions:
9075 they shouldn't be called directly but wrapper functions should be used
9078 So we also now have some wrapper functions that call the X509at functions
9079 when passed certificate requests. (TO DO: similar things can be done with
9080 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9081 things. Some of these need some d2i or i2d and print functionality
9082 because they handle more complex structures.)
9085 *) Add missing #ifndefs that caused missing symbols when building libssl
9086 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9087 NO_RSA in ssl/s2*.c.
9088 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9090 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9091 has a return value which indicates the quality of the random data
9092 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9093 error queue. New function RAND_pseudo_bytes() generates output that is
9094 guaranteed to be unique but not unpredictable. RAND_add is like
9095 RAND_seed, but takes an extra argument for an entropy estimate
9096 (RAND_seed always assumes full entropy).
9099 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9100 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9101 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9102 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9103 false-positive rate of at most 2^-80 for random input.
9106 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9109 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9110 in the 0.9.5 release), this returns the chain
9111 from an X509_CTX structure with a dup of the stack and all
9112 the X509 reference counts upped: so the stack will exist
9113 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9116 Also make SSL_SESSION_print() print out the verify return
9120 *) Add manpage for the pkcs12 command. Also change the default
9121 behaviour so MAC iteration counts are used unless the new
9122 -nomaciter option is used. This improves file security and
9123 only older versions of MSIE (4.0 for example) need it.
9126 *) Honor the no-xxx Configure options when creating .DEF files.
9129 *) Add PKCS#10 attributes to field table: challengePassword,
9130 unstructuredName and unstructuredAddress. These are taken from
9131 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9132 international characters are used.
9134 More changes to X509_ATTRIBUTE code: allow the setting of types
9135 based on strings. Remove the 'loc' parameter when adding
9136 attributes because these will be a SET OF encoding which is sorted
9140 *) Initial changes to the 'req' utility to allow request generation
9141 automation. This will allow an application to just generate a template
9142 file containing all the field values and have req construct the
9145 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9146 used all over the place including certificate requests and PKCS#7
9147 structures. They are currently handled manually where necessary with
9148 some primitive wrappers for PKCS#7. The new functions behave in a
9149 manner analogous to the X509 extension functions: they allow
9150 attributes to be looked up by NID and added.
9152 Later something similar to the X509V3 code would be desirable to
9153 automatically handle the encoding, decoding and printing of the
9154 more complex types. The string types like challengePassword can
9155 be handled by the string table functions.
9157 Also modified the multi byte string table handling. Now there is
9158 a 'global mask' which masks out certain types. The table itself
9159 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9160 is useful when for example there is only one permissible type
9161 (as in countryName) and using the mask might result in no valid
9165 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9166 SSL_get_peer_finished to allow applications to obtain the latest
9167 Finished messages sent to the peer or expected from the peer,
9168 respectively. (SSL_get_peer_finished is usually the Finished message
9169 actually received from the peer, otherwise the protocol will be aborted.)
9171 As the Finished message are message digests of the complete handshake
9172 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9173 be used for external authentication procedures when the authentication
9174 provided by SSL/TLS is not desired or is not enough.
9177 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9178 the host supports BWX extension and if Compaq C is present on the
9179 $PATH. Just exploiting of the BWX extension results in 20-30%
9180 performance kick for some algorithms, e.g. DES and RC4 to mention
9181 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9185 *) Add support for MS "fast SGC". This is arguably a violation of the
9186 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9187 weak crypto and after checking the certificate is SGC a second one
9188 with strong crypto. MS SGC stops the first handshake after receiving
9189 the server certificate message and sends a second client hello. Since
9190 a server will typically do all the time consuming operations before
9191 expecting any further messages from the client (server key exchange
9192 is the most expensive) there is little difference between the two.
9194 To get OpenSSL to support MS SGC we have to permit a second client
9195 hello message after we have sent server done. In addition we have to
9196 reset the MAC if we do get this second client hello.
9199 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9200 if a DER encoded private key is RSA or DSA traditional format. Changed
9201 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9202 format DER encoded private key. Newer code should use PKCS#8 format which
9203 has the key type encoded in the ASN1 structure. Added DER private key
9204 support to pkcs8 application.
9207 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9208 ciphersuites has been selected (as required by the SSL 3/TLS 1
9209 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9210 is set, we interpret this as a request to violate the specification
9211 (the worst that can happen is a handshake failure, and 'correct'
9212 behaviour would result in a handshake failure anyway).
9215 *) In SSL_CTX_add_session, take into account that there might be multiple
9216 SSL_SESSION structures with the same session ID (e.g. when two threads
9217 concurrently obtain them from an external cache).
9218 The internal cache can handle only one SSL_SESSION with a given ID,
9219 so if there's a conflict, we now throw out the old one to achieve
9223 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9224 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9225 some routines that use cipher OIDs: some ciphers do not have OIDs
9226 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9230 *) Simplify the trust setting structure and code. Now we just have
9231 two sequences of OIDs for trusted and rejected settings. These will
9232 typically have values the same as the extended key usage extension
9233 and any application specific purposes.
9235 The trust checking code now has a default behaviour: it will just
9236 check for an object with the same NID as the passed id. Functions can
9237 be provided to override either the default behaviour or the behaviour
9238 for a given id. SSL client, server and email already have functions
9239 in place for compatibility: they check the NID and also return "trusted"
9240 if the certificate is self signed.
9243 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9244 traditional format into an EVP_PKEY structure.
9247 *) Add a password callback function PEM_cb() which either prompts for
9248 a password if usr_data is NULL or otherwise assumes it is a null
9249 terminated password. Allow passwords to be passed on command line
9250 environment or config files in a few more utilities.
9253 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9254 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9255 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9256 Update documentation.
9259 *) Support for ASN1 "NULL" type. This could be handled before by using
9260 ASN1_TYPE but there wasn't any function that would try to read a NULL
9261 and produce an error if it couldn't. For compatibility we also have
9262 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9263 don't allocate anything because they don't need to.
9266 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9268 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9270 *) Rebuild of the memory allocation routines used by OpenSSL code and
9271 possibly others as well. The purpose is to make an interface that
9272 provide hooks so anyone can build a separate set of allocation and
9273 deallocation routines to be used by OpenSSL, for example memory
9274 pool implementations, or something else, which was previously hard
9275 since Malloc(), Realloc() and Free() were defined as macros having
9276 the values malloc, realloc and free, respectively (except for Win32
9277 compilations). The same is provided for memory debugging code.
9278 OpenSSL already comes with functionality to find memory leaks, but
9279 this gives people a chance to debug other memory problems.
9281 With these changes, a new set of functions and macros have appeared:
9283 CRYPTO_set_mem_debug_functions() [F]
9284 CRYPTO_get_mem_debug_functions() [F]
9285 CRYPTO_dbg_set_options() [F]
9286 CRYPTO_dbg_get_options() [F]
9287 CRYPTO_malloc_debug_init() [M]
9289 The memory debug functions are NULL by default, unless the library
9290 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9291 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9292 gives the standard debugging functions that come with OpenSSL) or
9293 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9294 provided by the library user) must be used. When the standard
9295 debugging functions are used, CRYPTO_dbg_set_options can be used to
9296 request additional information:
9297 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9298 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9300 Also, things like CRYPTO_set_mem_functions will always give the
9301 expected result (the new set of functions is used for allocation
9302 and deallocation) at all times, regardless of platform and compiler
9305 To finish it up, some functions that were never use in any other
9306 way than through macros have a new API and new semantic:
9309 CRYPTO_dbg_realloc()
9312 All macros of value have retained their old syntax.
9313 [Richard Levitte and Bodo Moeller]
9315 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9316 ordering of SMIMECapabilities wasn't in "strength order" and there
9317 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9321 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9322 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9323 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9325 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9326 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9327 functionality to handle multipart/signed properly) and a utility
9328 called 'smime' to call all this stuff. This is based on code I
9329 originally wrote for Celo who have kindly allowed it to be
9330 included in OpenSSL.
9333 *) Add variants des_set_key_checked and des_set_key_unchecked of
9334 des_set_key (aka des_key_sched). Global variable des_check_key
9335 decides which of these is called by des_set_key; this way
9336 des_check_key behaves as it always did, but applications and
9337 the library itself, which was buggy for des_check_key == 1,
9338 have a cleaner way to pick the version they need.
9341 *) New function PKCS12_newpass() which changes the password of a
9345 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9346 dynamic mix. In both cases the ids can be used as an index into the
9347 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9348 functions so they accept a list of the field values and the
9349 application doesn't need to directly manipulate the X509_TRUST
9353 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9357 *) Modify the way the V3 extension code looks up extensions. This now
9358 works in a similar way to the object code: we have some "standard"
9359 extensions in a static table which is searched with OBJ_bsearch()
9360 and the application can add dynamic ones if needed. The file
9361 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9362 updated whenever a new extension is added to the core code and kept
9363 in ext_nid order. There is a simple program 'tabtest.c' which checks
9364 this. New extensions are not added too often so this file can readily
9365 be maintained manually.
9367 There are two big advantages in doing things this way. The extensions
9368 can be looked up immediately and no longer need to be "added" using
9369 X509V3_add_standard_extensions(): this function now does nothing.
9370 [Side note: I get *lots* of email saying the extension code doesn't
9371 work because people forget to call this function]
9372 Also no dynamic allocation is done unless new extensions are added:
9373 so if we don't add custom extensions there is no need to call
9374 X509V3_EXT_cleanup().
9377 *) Modify enc utility's salting as follows: make salting the default. Add a
9378 magic header, so unsalted files fail gracefully instead of just decrypting
9379 to garbage. This is because not salting is a big security hole, so people
9380 should be discouraged from doing it.
9383 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9384 digest to be passed on the command line but it only used this
9385 parameter when signing a certificate. Modified so all relevant
9386 operations are affected by the digest parameter including the
9387 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9388 DSA key was used because it didn't fix the digest.
9391 *) Initial certificate chain verify code. Currently tests the untrusted
9392 certificates for consistency with the verify purpose (which is set
9393 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9395 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9396 this is because it will reject chains with invalid extensions whereas
9397 every previous version of OpenSSL and SSLeay made no checks at all.
9399 Trust code: checks the root CA for the relevant trust settings. Trust
9400 settings have an initial value consistent with the verify purpose: e.g.
9401 if the verify purpose is for SSL client use it expects the CA to be
9402 trusted for SSL client use. However the default value can be changed to
9403 permit custom trust settings: one example of this would be to only trust
9404 certificates from a specific "secure" set of CAs.
9406 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9407 which should be used for version portability: especially since the
9408 verify structure is likely to change more often now.
9410 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9411 to set them. If not set then assume SSL clients will verify SSL servers
9414 Two new options to the verify program: -untrusted allows a set of
9415 untrusted certificates to be passed in and -purpose which sets the
9416 intended purpose of the certificate. If a purpose is set then the
9417 new chain verify code is used to check extension consistency.
9420 *) Support for the authority information access extension.
9423 *) Modify RSA and DSA PEM read routines to transparently handle
9424 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9425 public keys in a format compatible with certificate
9426 SubjectPublicKeyInfo structures. Unfortunately there were already
9427 functions called *_PublicKey_* which used various odd formats so
9428 these are retained for compatibility: however the DSA variants were
9429 never in a public release so they have been deleted. Changed dsa/rsa
9430 utilities to handle the new format: note no releases ever handled public
9431 keys so we should be OK.
9433 The primary motivation for this change is to avoid the same fiasco
9434 that dogs private keys: there are several incompatible private key
9435 formats some of which are standard and some OpenSSL specific and
9436 require various evil hacks to allow partial transparent handling and
9437 even then it doesn't work with DER formats. Given the option anything
9438 other than PKCS#8 should be dumped: but the other formats have to
9439 stay in the name of compatibility.
9441 With public keys and the benefit of hindsight one standard format
9442 is used which works with EVP_PKEY, RSA or DSA structures: though
9443 it clearly returns an error if you try to read the wrong kind of key.
9445 Added a -pubkey option to the 'x509' utility to output the public key.
9446 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9447 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9448 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9449 that do the same as the EVP_PKEY_assign_*() except they up the
9450 reference count of the added key (they don't "swallow" the
9454 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9455 CRLs would fail if the file contained no certificates or no CRLs:
9456 added a new function to read in both types and return the number
9457 read: this means that if none are read it will be an error. The
9458 DER versions of the certificate and CRL reader would always fail
9459 because it isn't possible to mix certificates and CRLs in DER format
9460 without choking one or the other routine. Changed this to just read
9461 a certificate: this is the best we can do. Also modified the code
9462 in apps/verify.c to take notice of return codes: it was previously
9463 attempting to read in certificates from NULL pointers and ignoring
9464 any errors: this is one reason why the cert and CRL reader seemed
9465 to work. It doesn't check return codes from the default certificate
9466 routines: these may well fail if the certificates aren't installed.
9469 *) Code to support otherName option in GeneralName.
9472 *) First update to verify code. Change the verify utility
9473 so it warns if it is passed a self signed certificate:
9474 for consistency with the normal behaviour. X509_verify
9475 has been modified to it will now verify a self signed
9476 certificate if *exactly* the same certificate appears
9477 in the store: it was previously impossible to trust a
9478 single self signed certificate. This means that:
9479 openssl verify ss.pem
9480 now gives a warning about a self signed certificate but
9481 openssl verify -CAfile ss.pem ss.pem
9485 *) For servers, store verify_result in SSL_SESSION data structure
9486 (and add it to external session representation).
9487 This is needed when client certificate verifications fails,
9488 but an application-provided verification callback (set by
9489 SSL_CTX_set_cert_verify_callback) allows accepting the session
9490 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9491 but returns 1): When the session is reused, we have to set
9492 ssl->verify_result to the appropriate error code to avoid
9494 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9496 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9497 case in PKCS7_dataInit() where the signed PKCS7 structure
9498 didn't contain any existing data because it was being created.
9499 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9501 *) Add a salt to the key derivation routines in enc.c. This
9502 forms the first 8 bytes of the encrypted file. Also add a
9503 -S option to allow a salt to be input on the command line.
9506 *) New function X509_cmp(). Oddly enough there wasn't a function
9507 to compare two certificates. We do this by working out the SHA1
9508 hash and comparing that. X509_cmp() will be needed by the trust
9512 *) SSL_get1_session() is like SSL_get_session(), but increments
9513 the reference count in the SSL_SESSION returned.
9514 [Geoff Thorpe <geoff@eu.c2.net>]
9516 *) Fix for 'req': it was adding a null to request attributes.
9517 Also change the X509_LOOKUP and X509_INFO code to handle
9518 certificate auxiliary information.
9521 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9525 *) Add the possibility to add extra information to the memory leak
9526 detecting output, to form tracebacks, showing from where each
9527 allocation was originated: CRYPTO_push_info("constant string") adds
9528 the string plus current file name and line number to a per-thread
9529 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9530 is like calling CYRPTO_pop_info() until the stack is empty.
9531 Also updated memory leak detection code to be multi-thread-safe.
9534 *) Add options -text and -noout to pkcs7 utility and delete the
9535 encryption options which never did anything. Update docs.
9538 *) Add options to some of the utilities to allow the pass phrase
9539 to be included on either the command line (not recommended on
9540 OSes like Unix) or read from the environment. Update the
9541 manpages and fix a few bugs.
9544 *) Add a few manpages for some of the openssl commands.
9547 *) Fix the -revoke option in ca. It was freeing up memory twice,
9548 leaking and not finding already revoked certificates.
9551 *) Extensive changes to support certificate auxiliary information.
9552 This involves the use of X509_CERT_AUX structure and X509_AUX
9553 functions. An X509_AUX function such as PEM_read_X509_AUX()
9554 can still read in a certificate file in the usual way but it
9555 will also read in any additional "auxiliary information". By
9556 doing things this way a fair degree of compatibility can be
9557 retained: existing certificates can have this information added
9558 using the new 'x509' options.
9560 Current auxiliary information includes an "alias" and some trust
9561 settings. The trust settings will ultimately be used in enhanced
9562 certificate chain verification routines: currently a certificate
9563 can only be trusted if it is self signed and then it is trusted
9567 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9568 The problem was that one of the replacement routines had not been working
9569 since SSLeay releases. For now the offending routine has been replaced
9570 with non-optimised assembler. Even so, this now gives around 95%
9571 performance improvement for 1024 bit RSA signs.
9574 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9575 handling. Most clients have the effective key size in bits equal to
9576 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9577 A few however don't do this and instead use the size of the decrypted key
9578 to determine the RC2 key length and the AlgorithmIdentifier to determine
9579 the effective key length. In this case the effective key length can still
9580 be 40 bits but the key length can be 168 bits for example. This is fixed
9581 by manually forcing an RC2 key into the EVP_PKEY structure because the
9582 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9583 the key length and effective key length are equal.
9586 *) Add a bunch of functions that should simplify the creation of
9587 X509_NAME structures. Now you should be able to do:
9588 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9589 and have it automatically work out the correct field type and fill in
9590 the structures. The more adventurous can try:
9591 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9592 and it will (hopefully) work out the correct multibyte encoding.
9595 *) Change the 'req' utility to use the new field handling and multibyte
9596 copy routines. Before the DN field creation was handled in an ad hoc
9597 way in req, ca, and x509 which was rather broken and didn't support
9598 BMPStrings or UTF8Strings. Since some software doesn't implement
9599 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9600 using the dirstring_type option. See the new comment in the default
9601 openssl.cnf for more info.
9604 *) Make crypto/rand/md_rand.c more robust:
9605 - Assure unique random numbers after fork().
9606 - Make sure that concurrent threads access the global counter and
9607 md serializably so that we never lose entropy in them
9608 or use exactly the same state in multiple threads.
9609 Access to the large state is not always serializable because
9610 the additional locking could be a performance killer, and
9611 md should be large enough anyway.
9614 *) New file apps/app_rand.c with commonly needed functionality
9615 for handling the random seed file.
9617 Use the random seed file in some applications that previously did not:
9619 dsaparam -genkey (which also ignored its '-rand' option),
9622 x509 (when signing).
9623 Except on systems with /dev/urandom, it is crucial to have a random
9624 seed file at least for key creation, DSA signing, and for DH exchanges;
9625 for RSA signatures we could do without one.
9627 gendh and gendsa (unlike genrsa) used to read only the first byte
9628 of each file listed in the '-rand' option. The function as previously
9629 found in genrsa is now in app_rand.c and is used by all programs
9630 that support '-rand'.
9633 *) In RAND_write_file, use mode 0600 for creating files;
9634 don't just chmod when it may be too late.
9637 *) Report an error from X509_STORE_load_locations
9638 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9641 *) New function ASN1_mbstring_copy() this copies a string in either
9642 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9643 into an ASN1_STRING type. A mask of permissible types is passed
9644 and it chooses the "minimal" type to use or an error if not type
9648 *) Add function equivalents to the various macros in asn1.h. The old
9649 macros are retained with an M_ prefix. Code inside the library can
9650 use the M_ macros. External code (including the openssl utility)
9651 should *NOT* in order to be "shared library friendly".
9654 *) Add various functions that can check a certificate's extensions
9655 to see if it usable for various purposes such as SSL client,
9656 server or S/MIME and CAs of these types. This is currently
9657 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9658 verification. Also added a -purpose flag to x509 utility to
9659 print out all the purposes.
9662 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9666 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9667 for, obtain and decode and extension and obtain its critical flag.
9668 This allows all the necessary extension code to be handled in a
9669 single function call.
9672 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9673 platforms. See crypto/rc4/rc4_enc.c for further details.
9676 *) New -noout option to asn1parse. This causes no output to be produced
9677 its main use is when combined with -strparse and -out to extract data
9678 from a file (which may not be in ASN.1 format).
9681 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9682 when producing the local key id.
9683 [Richard Levitte <levitte@stacken.kth.se>]
9685 *) New option -dhparam in s_server. This allows a DH parameter file to be
9686 stated explicitly. If it is not stated then it tries the first server
9687 certificate file. The previous behaviour hard coded the filename
9691 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9692 a public key to be input or output. For example:
9693 openssl rsa -in key.pem -pubout -out pubkey.pem
9694 Also added necessary DSA public key functions to handle this.
9697 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9698 in the message. This was handled by allowing
9699 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9700 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9702 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9703 to the end of the strings whereas this didn't. This would cause problems
9704 if strings read with d2i_ASN1_bytes() were later modified.
9705 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9707 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9708 data and it contains EOF it will end up returning an error. This is
9709 caused by input 46 bytes long. The cause is due to the way base64
9710 BIOs find the start of base64 encoded data. They do this by trying a
9711 trial decode on each line until they find one that works. When they
9712 do a flag is set and it starts again knowing it can pass all the
9713 data directly through the decoder. Unfortunately it doesn't reset
9714 the context it uses. This means that if EOF is reached an attempt
9715 is made to pass two EOFs through the context and this causes the
9716 resulting error. This can also cause other problems as well. As is
9717 usual with these problems it takes *ages* to find and the fix is
9718 trivial: move one line.
9719 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9721 *) Ugly workaround to get s_client and s_server working under Windows. The
9722 old code wouldn't work because it needed to select() on sockets and the
9723 tty (for keypresses and to see if data could be written). Win32 only
9724 supports select() on sockets so we select() with a 1s timeout on the
9725 sockets and then see if any characters are waiting to be read, if none
9726 are present then we retry, we also assume we can always write data to
9727 the tty. This isn't nice because the code then blocks until we've
9728 received a complete line of data and it is effectively polling the
9729 keyboard at 1s intervals: however it's quite a bit better than not
9730 working at all :-) A dedicated Windows application might handle this
9731 with an event loop for example.
9734 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9735 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9736 will be called when RSA_sign() and RSA_verify() are used. This is useful
9737 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9738 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9739 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9740 This necessitated the support of an extra signature type NID_md5_sha1
9741 for SSL signatures and modifications to the SSL library to use it instead
9742 of calling RSA_public_decrypt() and RSA_private_encrypt().
9745 *) Add new -verify -CAfile and -CApath options to the crl program, these
9746 will lookup a CRL issuers certificate and verify the signature in a
9747 similar way to the verify program. Tidy up the crl program so it
9748 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9749 less strict. It will now permit CRL extensions even if it is not
9750 a V2 CRL: this will allow it to tolerate some broken CRLs.
9753 *) Initialize all non-automatic variables each time one of the openssl
9754 sub-programs is started (this is necessary as they may be started
9755 multiple times from the "OpenSSL>" prompt).
9756 [Lennart Bang, Bodo Moeller]
9758 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9759 removing all other RSA functionality (this is what NO_RSA does). This
9760 is so (for example) those in the US can disable those operations covered
9761 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9765 *) Non-copying interface to BIO pairs.
9766 (still largely untested)
9769 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9770 ASCII string. This was handled independently in various places before.
9773 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9774 UTF8 strings a character at a time.
9777 *) Use client_version from client hello to select the protocol
9778 (s23_srvr.c) and for RSA client key exchange verification
9779 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9782 *) Add various utility functions to handle SPKACs, these were previously
9783 handled by poking round in the structure internals. Added new function
9784 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9785 print, verify and generate SPKACs. Based on an original idea from
9786 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9789 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9792 *) Allow the config file extension section to be overwritten on the
9793 command line. Based on an original idea from Massimiliano Pala
9794 <madwolf@comune.modena.it>. The new option is called -extensions
9795 and can be applied to ca, req and x509. Also -reqexts to override
9796 the request extensions in req and -crlexts to override the crl extensions
9800 *) Add new feature to the SPKAC handling in ca. Now you can include
9801 the same field multiple times by preceding it by "XXXX." for example:
9804 this is the same syntax as used in the req config file.
9807 *) Allow certificate extensions to be added to certificate requests. These
9808 are specified in a 'req_extensions' option of the req section of the
9809 config file. They can be printed out with the -text option to req but
9810 are otherwise ignored at present.
9813 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9814 data read consists of only the final block it would not decrypted because
9815 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9816 A misplaced 'break' also meant the decrypted final block might not be
9817 copied until the next read.
9820 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9821 a few extra parameters to the DH structure: these will be useful if
9822 for example we want the value of 'q' or implement X9.42 DH.
9825 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9826 provides hooks that allow the default DSA functions or functions on a
9827 "per key" basis to be replaced. This allows hardware acceleration and
9828 hardware key storage to be handled without major modification to the
9829 library. Also added low level modexp hooks and CRYPTO_EX structure and
9830 associated functions.
9833 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9834 as "read only": it can't be written to and the buffer it points to will
9835 not be freed. Reading from a read only BIO is much more efficient than
9836 a normal memory BIO. This was added because there are several times when
9837 an area of memory needs to be read from a BIO. The previous method was
9838 to create a memory BIO and write the data to it, this results in two
9839 copies of the data and an O(n^2) reading algorithm. There is a new
9840 function BIO_new_mem_buf() which creates a read only memory BIO from
9841 an area of memory. Also modified the PKCS#7 routines to use read only
9845 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9846 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9847 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9848 but a retry condition occured while trying to read the rest.
9851 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9852 NID_pkcs7_encrypted by default: this was wrong since this should almost
9853 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9854 the encrypted data type: this is a more sensible place to put it and it
9855 allows the PKCS#12 code to be tidied up that duplicated this
9859 *) Changed obj_dat.pl script so it takes its input and output files on
9860 the command line. This should avoid shell escape redirection problems
9864 *) Initial support for certificate extension requests, these are included
9865 in things like Xenroll certificate requests. Included functions to allow
9866 extensions to be obtained and added.
9869 *) -crlf option to s_client and s_server for sending newlines as
9870 CRLF (as required by many protocols).
9873 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9875 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9876 [Ralf S. Engelschall]
9878 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9879 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9881 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9885 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9886 DH parameters/keys (q is lost during that conversion, but the resulting
9887 DH parameters contain its length).
9889 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9890 much faster than DH_generate_parameters (which creates parameters
9891 where p = 2*q + 1), and also the smaller q makes DH computations
9892 much more efficient (160-bit exponentiation instead of 1024-bit
9893 exponentiation); so this provides a convenient way to support DHE
9894 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9895 utter importance to use
9896 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9898 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9899 when such DH parameters are used, because otherwise small subgroup
9900 attacks may become possible!
9903 *) Avoid memory leak in i2d_DHparams.
9906 *) Allow the -k option to be used more than once in the enc program:
9907 this allows the same encrypted message to be read by multiple recipients.
9910 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9911 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9912 it will always use the numerical form of the OID, even if it has a short
9916 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9917 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9918 otherwise bn_mod_exp was called. In the case of hardware keys for example
9919 no private key components need be present and it might store extra data
9920 in the RSA structure, which cannot be accessed from bn_mod_exp.
9921 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9922 private key operations.
9925 *) Added support for SPARC Linux.
9928 *) pem_password_cb function type incompatibly changed from
9929 typedef int pem_password_cb(char *buf, int size, int rwflag);
9931 ....(char *buf, int size, int rwflag, void *userdata);
9932 so that applications can pass data to their callbacks:
9933 The PEM[_ASN1]_{read,write}... functions and macros now take an
9934 additional void * argument, which is just handed through whenever
9935 the password callback is called.
9936 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9938 New function SSL_CTX_set_default_passwd_cb_userdata.
9940 Compatibility note: As many C implementations push function arguments
9941 onto the stack in reverse order, the new library version is likely to
9942 interoperate with programs that have been compiled with the old
9943 pem_password_cb definition (PEM_whatever takes some data that
9944 happens to be on the stack as its last argument, and the callback
9945 just ignores this garbage); but there is no guarantee whatsoever that
9948 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9949 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9950 problems not only on Windows, but also on some Unix platforms.
9951 To avoid problematic command lines, these definitions are now in an
9952 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9953 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9956 *) MIPS III/IV assembler module is reimplemented.
9959 *) More DES library cleanups: remove references to srand/rand and
9960 delete an unused file.
9963 *) Add support for the the free Netwide assembler (NASM) under Win32,
9964 since not many people have MASM (ml) and it can be hard to obtain.
9965 This is currently experimental but it seems to work OK and pass all
9966 the tests. Check out INSTALL.W32 for info.
9969 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9970 without temporary keys kept an extra copy of the server key,
9971 and connections with temporary keys did not free everything in case
9975 *) New function RSA_check_key and new openssl rsa option -check
9976 for verifying the consistency of RSA keys.
9977 [Ulf Moeller, Bodo Moeller]
9979 *) Various changes to make Win32 compile work:
9980 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9981 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9982 comparison" warnings.
9983 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9986 *) Add a debugging option to PKCS#5 v2 key generation function: when
9987 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9988 derived keys are printed to stderr.
9991 *) Copy the flags in ASN1_STRING_dup().
9992 [Roman E. Pavlov <pre@mo.msk.ru>]
9994 *) The x509 application mishandled signing requests containing DSA
9995 keys when the signing key was also DSA and the parameters didn't match.
9997 It was supposed to omit the parameters when they matched the signing key:
9998 the verifying software was then supposed to automatically use the CA's
9999 parameters if they were absent from the end user certificate.
10001 Omitting parameters is no longer recommended. The test was also
10002 the wrong way round! This was probably due to unusual behaviour in
10003 EVP_cmp_parameters() which returns 1 if the parameters match.
10004 This meant that parameters were omitted when they *didn't* match and
10005 the certificate was useless. Certificates signed with 'ca' didn't have
10007 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10009 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10010 The interface is as follows:
10011 Applications can use
10012 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10013 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10014 "off" is now the default.
10015 The library internally uses
10016 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10017 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10018 to disable memory-checking temporarily.
10020 Some inconsistent states that previously were possible (and were
10021 even the default) are now avoided.
10023 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10024 with each memory chunk allocated; this is occasionally more helpful
10025 than just having a counter.
10027 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10029 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10033 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10034 which largely parallels "options", but is for changing API behaviour,
10035 whereas "options" are about protocol behaviour.
10036 Initial "mode" flags are:
10038 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10039 a single record has been written.
10040 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10041 retries use the same buffer location.
10042 (But all of the contents must be
10046 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10049 *) Fix problems with no-hmac etc.
10050 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10052 *) New functions RSA_get_default_method(), RSA_set_method() and
10053 RSA_get_method(). These allows replacement of RSA_METHODs without having
10054 to mess around with the internals of an RSA structure.
10057 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10058 Also really enable memory leak checks in openssl.c and in some
10060 [Chad C. Mulligan, Bodo Moeller]
10062 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10063 up the length of negative integers. This has now been simplified to just
10064 store the length when it is first determined and use it later, rather
10065 than trying to keep track of where data is copied and updating it to
10067 [Steve Henson, reported by Brien Wheeler
10068 <bwheeler@authentica-security.com>]
10070 *) Add a new function PKCS7_signatureVerify. This allows the verification
10071 of a PKCS#7 signature but with the signing certificate passed to the
10072 function itself. This contrasts with PKCS7_dataVerify which assumes the
10073 certificate is present in the PKCS#7 structure. This isn't always the
10074 case: certificates can be omitted from a PKCS#7 structure and be
10075 distributed by "out of band" means (such as a certificate database).
10078 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10079 function prototypes in pem.h, also change util/mkdef.pl to add the
10080 necessary function names.
10083 *) mk1mf.pl (used by Windows builds) did not properly read the
10084 options set by Configure in the top level Makefile, and Configure
10085 was not even able to write more than one option correctly.
10086 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10089 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10090 file to be loaded from a BIO or FILE pointer. The BIO version will
10091 for example allow memory BIOs to contain config info.
10094 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10095 Whoever hopes to achieve shared-library compatibility across versions
10096 must use this, not the compile-time macro.
10097 (Exercise 0.9.4: Which is the minimum library version required by
10099 Note: All this applies only to multi-threaded programs, others don't
10103 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10104 through a BIO pair triggered the default case, i.e.
10105 SSLerr(...,SSL_R_UNKNOWN_STATE).
10108 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10109 can use the SSL library even if none of the specific BIOs is
10113 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10114 for the encoded length.
10115 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10117 *) Add initial documentation of the X509V3 functions.
10120 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10121 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10122 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10123 secure PKCS#8 private key format with a high iteration count.
10126 *) Fix determination of Perl interpreter: A perl or perl5
10127 _directory_ in $PATH was also accepted as the interpreter.
10128 [Ralf S. Engelschall]
10130 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10131 wrong with it but it was very old and did things like calling
10132 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10133 unusual formatting.
10136 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10137 to use the new extension code.
10140 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10141 with macros. This should make it easier to change their form, add extra
10142 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10146 *) Add to configuration table a new entry that can specify an alternative
10147 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10148 according to Mark Crispin <MRC@Panda.COM>.
10152 *) DES CBC did not update the IV. Weird.
10155 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10156 Changing the behaviour of the former might break existing programs --
10157 where IV updating is needed, des_ncbc_encrypt can be used.
10160 *) When bntest is run from "make test" it drives bc to check its
10161 calculations, as well as internally checking them. If an internal check
10162 fails, it needs to cause bc to give a non-zero result or make test carries
10163 on without noticing the failure. Fixed.
10166 *) DES library cleanups.
10169 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10170 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10171 ciphers. NOTE: although the key derivation function has been verified
10172 against some published test vectors it has not been extensively tested
10173 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10177 *) Instead of "mkdir -p", which is not fully portable, use new
10178 Perl script "util/mkdir-p.pl".
10181 *) Rewrite the way password based encryption (PBE) is handled. It used to
10182 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10183 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10184 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10185 the 'parameter' field of the AlgorithmIdentifier is passed to the
10186 underlying key generation function so it must do its own ASN1 parsing.
10187 This has also changed the EVP_PBE_CipherInit() function which now has a
10188 'parameter' argument instead of literal salt and iteration count values
10189 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10192 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10193 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10194 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10195 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10196 value was just used as a "magic string" and not used directly its
10197 value doesn't matter.
10200 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10204 *) "linux-sparc64" configuration (ultrapenguin).
10205 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10206 "linux-sparc" configuration.
10207 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10209 *) config now generates no-xxx options for missing ciphers.
10212 *) Support the EBCDIC character set (work in progress).
10213 File ebcdic.c not yet included because it has a different license.
10214 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10216 *) Support BS2000/OSD-POSIX.
10217 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10219 *) Make callbacks for key generation use void * instead of char *.
10222 *) Make S/MIME samples compile (not yet tested).
10225 *) Additional typesafe stacks.
10228 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10232 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10234 *) New configuration variant "sco5-gcc".
10236 *) Updated some demos.
10237 [Sean O Riordain, Wade Scholine]
10239 *) Add missing BIO_free at exit of pkcs12 application.
10242 *) Fix memory leak in conf.c.
10245 *) Updates for Win32 to assembler version of MD5.
10248 *) Set #! path to perl in apps/der_chop to where we found it
10249 instead of using a fixed path.
10252 *) SHA library changes for irix64-mips4-cc.
10255 *) Improvements for VMS support.
10259 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10261 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10262 This also avoids the problems with SC4.2 and unpatched SC5.
10263 [Andy Polyakov <appro@fy.chalmers.se>]
10265 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10266 These are required because of the typesafe stack would otherwise break
10267 existing code. If old code used a structure member which used to be STACK
10268 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10269 sk_num or sk_value it would produce an error because the num, data members
10270 are not present in STACK_OF. Now it just produces a warning. sk_set
10271 replaces the old method of assigning a value to sk_value
10272 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10273 that does this will no longer work (and should use sk_set instead) but
10274 this could be regarded as a "questionable" behaviour anyway.
10277 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10278 correctly handle encrypted S/MIME data.
10281 *) Change type of various DES function arguments from des_cblock
10282 (which means, in function argument declarations, pointer to char)
10283 to des_cblock * (meaning pointer to array with 8 char elements),
10284 which allows the compiler to do more typechecking; it was like
10285 that back in SSLeay, but with lots of ugly casts.
10287 Introduce new type const_des_cblock.
10290 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10291 problems: find RecipientInfo structure that matches recipient certificate
10292 and initialise the ASN1 structures properly based on passed cipher.
10295 *) Belatedly make the BN tests actually check the results.
10298 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10299 to and from BNs: it was completely broken. New compilation option
10300 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10301 key elements as negative integers.
10304 *) Reorganize and speed up MD5.
10305 [Andy Polyakov <appro@fy.chalmers.se>]
10308 [Richard Levitte <richard@levitte.org>]
10310 *) New option -out to asn1parse to allow the parsed structure to be
10311 output to a file. This is most useful when combined with the -strparse
10312 option to examine the output of things like OCTET STRINGS.
10315 *) Make SSL library a little more fool-proof by not requiring any longer
10316 that SSL_set_{accept,connect}_state be called before
10317 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10318 in many applications because usually everything *appeared* to work as
10319 intended anyway -- now it really works as intended).
10322 *) Move openssl.cnf out of lib/.
10325 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10326 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10327 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10328 [Ralf S. Engelschall]
10330 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10331 handle PKCS#7 enveloped data properly.
10332 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10334 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10335 copying pointers. The cert_st handling is changed by this in
10336 various ways (and thus what used to be known as ctx->default_cert
10337 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10338 any longer when s->cert does not give us what we need).
10339 ssl_cert_instantiate becomes obsolete by this change.
10340 As soon as we've got the new code right (possibly it already is?),
10341 we have solved a couple of bugs of the earlier code where s->cert
10342 was used as if it could not have been shared with other SSL structures.
10344 Note that using the SSL API in certain dirty ways now will result
10345 in different behaviour than observed with earlier library versions:
10346 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10347 does not influence s as it used to.
10349 In order to clean up things more thoroughly, inside SSL_SESSION
10350 we don't use CERT any longer, but a new structure SESS_CERT
10351 that holds per-session data (if available); currently, this is
10352 the peer's certificate chain and, for clients, the server's certificate
10353 and temporary key. CERT holds only those values that can have
10354 meaningful defaults in an SSL_CTX.
10357 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10358 from the internal representation. Various PKCS#7 fixes: remove some
10359 evil casts and set the enc_dig_alg field properly based on the signing
10363 *) Allow PKCS#12 password to be set from the command line or the
10364 environment. Let 'ca' get its config file name from the environment
10365 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10369 *) Allow certificate policies extension to use an IA5STRING for the
10370 organization field. This is contrary to the PKIX definition but
10371 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10375 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10376 without disallowing inline assembler and the like for non-pedantic builds.
10379 *) Support Borland C++ builder.
10380 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10382 *) Support Mingw32.
10385 *) SHA-1 cleanups and performance enhancements.
10386 [Andy Polyakov <appro@fy.chalmers.se>]
10388 *) Sparc v8plus assembler for the bignum library.
10389 [Andy Polyakov <appro@fy.chalmers.se>]
10391 *) Accept any -xxx and +xxx compiler options in Configure.
10394 *) Update HPUX configuration.
10397 *) Add missing sk_<type>_unshift() function to safestack.h
10398 [Ralf S. Engelschall]
10400 *) New function SSL_CTX_use_certificate_chain_file that sets the
10401 "extra_cert"s in addition to the certificate. (This makes sense
10402 only for "PEM" format files, as chains as a whole are not
10406 *) Support verify_depth from the SSL API.
10407 x509_vfy.c had what can be considered an off-by-one-error:
10408 Its depth (which was not part of the external interface)
10409 was actually counting the number of certificates in a chain;
10410 now it really counts the depth.
10413 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10414 instead of X509err, which often resulted in confusing error
10415 messages since the error codes are not globally unique
10416 (e.g. an alleged error in ssl3_accept when a certificate
10417 didn't match the private key).
10419 *) New function SSL_CTX_set_session_id_context that allows to set a default
10420 value (so that you don't need SSL_set_session_id_context for each
10421 connection using the SSL_CTX).
10424 *) OAEP decoding bug fix.
10427 *) Support INSTALL_PREFIX for package builders, as proposed by
10431 *) New Configure options "threads" and "no-threads". For systems
10432 where the proper compiler options are known (currently Solaris
10433 and Linux), "threads" is the default.
10436 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10439 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10440 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10441 such as /usr/local/bin.
10444 *) "make linux-shared" to build shared libraries.
10445 [Niels Poppe <niels@netbox.org>]
10447 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10450 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10451 extension adding in x509 utility.
10454 *) Remove NOPROTO sections and error code comments.
10457 *) Partial rewrite of the DEF file generator to now parse the ANSI
10461 *) New Configure options --prefix=DIR and --openssldir=DIR.
10464 *) Complete rewrite of the error code script(s). It is all now handled
10465 by one script at the top level which handles error code gathering,
10466 header rewriting and C source file generation. It should be much better
10467 than the old method: it now uses a modified version of Ulf's parser to
10468 read the ANSI prototypes in all header files (thus the old K&R definitions
10469 aren't needed for error creation any more) and do a better job of
10470 translating function codes into names. The old 'ASN1 error code imbedded
10471 in a comment' is no longer necessary and it doesn't use .err files which
10472 have now been deleted. Also the error code call doesn't have to appear all
10473 on one line (which resulted in some large lines...).
10476 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10479 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10480 0 (which usually indicates a closed connection), but continue reading.
10483 *) Fix some race conditions.
10486 *) Add support for CRL distribution points extension. Add Certificate
10487 Policies and CRL distribution points documentation.
10490 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10493 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10494 8 of keying material. Merlin has also confirmed interop with this fix
10495 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10496 [Merlin Hughes <merlin@baltimore.ie>]
10498 *) Fix lots of warnings.
10499 [Richard Levitte <levitte@stacken.kth.se>]
10501 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10502 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10503 [Richard Levitte <levitte@stacken.kth.se>]
10505 *) Fix problems with sizeof(long) == 8.
10506 [Andy Polyakov <appro@fy.chalmers.se>]
10508 *) Change functions to ANSI C.
10511 *) Fix typos in error codes.
10512 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10514 *) Remove defunct assembler files from Configure.
10517 *) SPARC v8 assembler BIGNUM implementation.
10518 [Andy Polyakov <appro@fy.chalmers.se>]
10520 *) Support for Certificate Policies extension: both print and set.
10521 Various additions to support the r2i method this uses.
10524 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10525 return a const string when you are expecting an allocated buffer.
10528 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10529 types DirectoryString and DisplayText.
10532 *) Add code to allow r2i extensions to access the configuration database,
10533 add an LHASH database driver and add several ctx helper functions.
10536 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10537 fail when they extended the size of a BIGNUM.
10540 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10541 support typesafe stack.
10544 *) Fix typo in SSL_[gs]et_options().
10545 [Nils Frostberg <nils@medcom.se>]
10547 *) Delete various functions and files that belonged to the (now obsolete)
10548 old X509V3 handling code.
10551 *) New Configure option "rsaref".
10554 *) Don't auto-generate pem.h.
10557 *) Introduce type-safe ASN.1 SETs.
10560 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10561 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10563 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10564 that links with OpenSSL (well at least cause lots of warnings), but fear
10565 not: the conversion is trivial, and it eliminates loads of evil casts. A
10566 few STACKed things have been converted already. Feel free to convert more.
10567 In the fullness of time, I'll do away with the STACK type altogether.
10570 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10571 specified in <certfile> by updating the entry in the index.txt file.
10572 This way one no longer has to edit the index.txt file manually for
10573 revoking a certificate. The -revoke option does the gory details now.
10574 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10576 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10577 `-text' option at all and this way the `-noout -text' combination was
10578 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10579 [Ralf S. Engelschall]
10581 *) Make sure a corresponding plain text error message exists for the
10582 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10583 verify callback function determined that a certificate was revoked.
10584 [Ralf S. Engelschall]
10586 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10587 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10588 all available cipers including rc5, which was forgotten until now.
10589 In order to let the testing shell script know which algorithms
10590 are available, a new (up to now undocumented) command
10591 "openssl list-cipher-commands" is used.
10594 *) Bugfix: s_client occasionally would sleep in select() when
10595 it should have checked SSL_pending() first.
10598 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10599 the raw DSA values prior to ASN.1 encoding.
10602 *) Tweaks to Configure
10603 [Niels Poppe <niels@netbox.org>]
10605 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10609 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10612 *) New config option to avoid instructions that are illegal on the 80386.
10613 The default code is faster, but requires at least a 486.
10616 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10617 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10618 same as SSL2_VERSION anyway.
10621 *) New "-showcerts" option for s_client.
10624 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10625 application. Various cleanups and fixes.
10628 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10629 modify error routines to work internally. Add error codes and PBE init
10630 to library startup routines.
10633 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10634 packing functions to asn1 and evp. Changed function names and error
10635 codes along the way.
10638 *) PKCS12 integration: and so it begins... First of several patches to
10639 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10640 objects to objects.h
10643 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10644 and display support for Thawte strong extranet extension.
10647 *) Add LinuxPPC support.
10648 [Jeff Dubrule <igor@pobox.org>]
10650 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10651 bn_div_words in alpha.s.
10652 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10654 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10655 OAEP isn't supported when OpenSSL is built with RSAref.
10656 [Ulf Moeller <ulf@fitug.de>]
10658 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10659 so they no longer are missing under -DNOPROTO.
10660 [Soren S. Jorvang <soren@t.dk>]
10663 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10665 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10666 doesn't work when the session is reused. Coming soon!
10669 *) Fix a security hole, that allows sessions to be reused in the wrong
10670 context thus bypassing client cert protection! All software that uses
10671 client certs and session caches in multiple contexts NEEDS PATCHING to
10672 allow session reuse! A fuller solution is in the works.
10673 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10675 *) Some more source tree cleanups (removed obsolete files
10676 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10677 permission on "config" script to be executable) and a fix for the INSTALL
10679 [Ulf Moeller <ulf@fitug.de>]
10681 *) Remove some legacy and erroneous uses of malloc, free instead of
10683 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10685 *) Make rsa_oaep_test return non-zero on error.
10686 [Ulf Moeller <ulf@fitug.de>]
10688 *) Add support for native Solaris shared libraries. Configure
10689 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10690 if someone would make that last step automatic.
10691 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10693 *) ctx_size was not built with the right compiler during "make links". Fixed.
10696 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10697 except NULL ciphers". This means the default cipher list will no longer
10698 enable NULL ciphers. They need to be specifically enabled e.g. with
10699 the string "DEFAULT:eNULL".
10702 *) Fix to RSA private encryption routines: if p < q then it would
10703 occasionally produce an invalid result. This will only happen with
10704 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10707 *) Be less restrictive and allow also `perl util/perlpath.pl
10708 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10709 because this way one can also use an interpreter named `perl5' (which is
10710 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10711 installed as `perl').
10712 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10714 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10715 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10717 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10718 advapi32.lib to Win32 build and change the pem test comparision
10719 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10720 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10721 and crypto/des/ede_cbcm_enc.c.
10724 *) DES quad checksum was broken on big-endian architectures. Fixed.
10727 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10728 Win32 test batch file so it (might) work again. The Win32 test batch file
10729 is horrible: I feel ill....
10732 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10733 in e_os.h. Audit of header files to check ANSI and non ANSI
10734 sections: 10 functions were absent from non ANSI section and not exported
10735 from Windows DLLs. Fixed up libeay.num for new functions.
10738 *) Make `openssl version' output lines consistent.
10739 [Ralf S. Engelschall]
10741 *) Fix Win32 symbol export lists for BIO functions: Added
10742 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10743 to ms/libeay{16,32}.def.
10744 [Ralf S. Engelschall]
10746 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10747 fine under Unix and passes some trivial tests I've now added. But the
10748 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10749 added to make sure no one expects that this stuff really works in the
10750 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10751 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10753 [Ralf S. Engelschall]
10755 *) Fix the generation of two part addresses in perl.
10756 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10758 *) Add config entry for Linux on MIPS.
10759 [John Tobey <jtobey@channel1.com>]
10761 *) Make links whenever Configure is run, unless we are on Windoze.
10764 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10765 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10769 *) Add a useful kludge to allow package maintainers to specify compiler and
10770 other platforms details on the command line without having to patch the
10771 Configure script everytime: One now can use ``perl Configure
10772 <id>:<details>'', i.e. platform ids are allowed to have details appended
10773 to them (seperated by colons). This is treated as there would be a static
10774 pre-configured entry in Configure's %table under key <id> with value
10775 <details> and ``perl Configure <id>'' is called. So, when you want to
10776 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10777 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10778 now, which overrides the FreeBSD-elf entry on-the-fly.
10779 [Ralf S. Engelschall]
10781 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10784 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10785 on the `perl Configure ...' command line. This way one can compile
10786 OpenSSL libraries with Position Independent Code (PIC) which is needed
10787 for linking it into DSOs.
10788 [Ralf S. Engelschall]
10790 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10794 *) Cleaned up the LICENSE document: The official contact for any license
10795 questions now is the OpenSSL core team under openssl-core@openssl.org.
10796 And add a paragraph about the dual-license situation to make sure people
10797 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10798 to the OpenSSL toolkit.
10799 [Ralf S. Engelschall]
10801 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10802 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10803 Additonally cleaned up the `make links' target: Remove unnecessary
10804 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10805 to speed processing and no longer clutter the display with confusing
10806 stuff. Instead only the actually done links are displayed.
10807 [Ralf S. Engelschall]
10809 *) Permit null encryption ciphersuites, used for authentication only. It used
10810 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10811 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10815 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10816 signed attributes when verifying signatures (this would break them),
10817 the detached data encoding was wrong and public keys obtained using
10818 X509_get_pubkey() weren't freed.
10821 *) Add text documentation for the BUFFER functions. Also added a work around
10822 to a Win95 console bug. This was triggered by the password read stuff: the
10823 last character typed gets carried over to the next fread(). If you were
10824 generating a new cert request using 'req' for example then the last
10825 character of the passphrase would be CR which would then enter the first
10829 *) Added the new `Includes OpenSSL Cryptography Software' button as
10830 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10831 button and can be used by applications based on OpenSSL to show the
10832 relationship to the OpenSSL project.
10833 [Ralf S. Engelschall]
10835 *) Remove confusing variables in function signatures in files
10836 ssl/ssl_lib.c and ssl/ssl.h.
10837 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10839 *) Don't install bss_file.c under PREFIX/include/
10840 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10842 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10843 functions that return function pointers and has support for NT specific
10844 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10845 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10846 unsigned to signed types: this was killing the Win32 compile.
10849 *) Add new certificate file to stack functions,
10850 SSL_add_dir_cert_subjects_to_stack() and
10851 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10852 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10853 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10854 This means that Apache-SSL and similar packages don't have to mess around
10855 to add as many CAs as they want to the preferred list.
10858 *) Experiment with doxygen documentation. Currently only partially applied to
10860 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10861 openssl.doxy as the configuration file.
10864 *) Get rid of remaining C++-style comments which strict C compilers hate.
10865 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10867 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10868 compiled in by default: it has problems with large keys.
10871 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10872 DH private keys and/or callback functions which directly correspond to
10873 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10874 is needed for applications which have to configure certificates on a
10875 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10877 For the RSA certificate situation is makes no difference, but
10878 for the DSA certificate situation this fixes the "no shared cipher"
10879 problem where the OpenSSL cipher selection procedure failed because the
10880 temporary keys were not overtaken from the context and the API provided
10881 no way to reconfigure them.
10882 The new functions now let applications reconfigure the stuff and they
10883 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10884 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10885 non-public-API function ssl_cert_instantiate() is used as a helper
10886 function and also to reduce code redundancy inside ssl_rsa.c.
10887 [Ralf S. Engelschall]
10889 *) Move s_server -dcert and -dkey options out of the undocumented feature
10890 area because they are useful for the DSA situation and should be
10891 recognized by the users.
10892 [Ralf S. Engelschall]
10894 *) Fix the cipher decision scheme for export ciphers: the export bits are
10895 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10896 SSL_EXP_MASK. So, the original variable has to be used instead of the
10897 already masked variable.
10898 [Richard Levitte <levitte@stacken.kth.se>]
10900 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10901 [Richard Levitte <levitte@stacken.kth.se>]
10903 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10904 from `int' to `unsigned int' because it's a length and initialized by
10905 EVP_DigestFinal() which expects an `unsigned int *'.
10906 [Richard Levitte <levitte@stacken.kth.se>]
10908 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10909 script. Instead use the usual Shell->Perl transition trick.
10910 [Ralf S. Engelschall]
10912 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10913 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10914 -noout -modulus' as it's already the case for `openssl rsa -noout
10915 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10916 currently the public key is printed (a decision which was already done by
10917 `openssl dsa -modulus' in the past) which serves a similar purpose.
10918 Additionally the NO_RSA no longer completely removes the whole -modulus
10919 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10921 [Ralf S. Engelschall]
10923 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10924 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10925 [Arne Ansper <arne@ats.cyber.ee>]
10927 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10928 to be added. Now both 'req' and 'ca' can use new objects defined in the
10932 *) Add cool BIO that does syslog (or event log on NT).
10933 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10935 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10936 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10937 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10938 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10941 *) Add preliminary config info for new extension code.
10944 *) Make RSA_NO_PADDING really use no padding.
10945 [Ulf Moeller <ulf@fitug.de>]
10947 *) Generate errors when private/public key check is done.
10950 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10951 for some CRL extensions and new objects added.
10954 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10955 key usage extension and fuller support for authority key id.
10958 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10959 padding method for RSA, which is recommended for new applications in PKCS
10960 #1 v2.0 (RFC 2437, October 1998).
10961 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10962 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10963 against Bleichbacher's attack on RSA.
10964 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10967 *) Updates to the new SSL compression code
10968 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10970 *) Fix so that the version number in the master secret, when passed
10971 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10972 (because the server will not accept higher), that the version number
10973 is 0x03,0x01, not 0x03,0x00
10974 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10976 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10977 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10978 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10981 *) Support for RAW extensions where an arbitrary extension can be
10982 created by including its DER encoding. See apps/openssl.cnf for
10986 *) Make sure latest Perl versions don't interpret some generated C array
10987 code as Perl array code in the crypto/err/err_genc.pl script.
10988 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10990 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10991 not many people have the assembler. Various Win32 compilation fixes and
10992 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10993 build instructions.
10996 *) Modify configure script 'Configure' to automatically create crypto/date.h
10997 file under Win32 and also build pem.h from pem.org. New script
10998 util/mkfiles.pl to create the MINFO file on environments that can't do a
10999 'make files': perl util/mkfiles.pl >MINFO should work.
11002 *) Major rework of DES function declarations, in the pursuit of correctness
11003 and purity. As a result, many evil casts evaporated, and some weirdness,
11004 too. You may find this causes warnings in your code. Zapping your evil
11005 casts will probably fix them. Mostly.
11008 *) Fix for a typo in asn1.h. Bug fix to object creation script
11009 obj_dat.pl. It considered a zero in an object definition to mean
11010 "end of object": none of the objects in objects.h have any zeros
11011 so it wasn't spotted.
11012 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11014 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11015 Masking (CBCM). In the absence of test vectors, the best I have been able
11016 to do is check that the decrypt undoes the encrypt, so far. Send me test
11017 vectors if you have them.
11020 *) Correct calculation of key length for export ciphers (too much space was
11021 allocated for null ciphers). This has not been tested!
11024 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11025 message is now correct (it understands "crypto" and "ssl" on its
11026 command line). There is also now an "update" option. This will update
11027 the util/ssleay.num and util/libeay.num files with any new functions.
11029 perl util/mkdef.pl crypto ssl update
11030 it will update them.
11033 *) Overhauled the Perl interface (perl/*):
11034 - ported BN stuff to OpenSSL's different BN library
11035 - made the perl/ source tree CVS-aware
11036 - renamed the package from SSLeay to OpenSSL (the files still contain
11037 their history because I've copied them in the repository)
11038 - removed obsolete files (the test scripts will be replaced
11039 by better Test::Harness variants in the future)
11040 [Ralf S. Engelschall]
11042 *) First cut for a very conservative source tree cleanup:
11043 1. merge various obsolete readme texts into doc/ssleay.txt
11044 where we collect the old documents and readme texts.
11045 2. remove the first part of files where I'm already sure that we no
11046 longer need them because of three reasons: either they are just temporary
11047 files which were left by Eric or they are preserved original files where
11048 I've verified that the diff is also available in the CVS via "cvs diff
11049 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11050 the crypto/md/ stuff).
11051 [Ralf S. Engelschall]
11053 *) More extension code. Incomplete support for subject and issuer alt
11054 name, issuer and authority key id. Change the i2v function parameters
11055 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11056 what that's for :-) Fix to ASN1 macro which messed up
11057 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11060 *) Preliminary support for ENUMERATED type. This is largely copied from the
11064 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11065 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11067 *) Make sure `make rehash' target really finds the `openssl' program.
11068 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11070 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11071 like to hear about it if this slows down other processors.
11074 *) Add CygWin32 platform information to Configure script.
11075 [Alan Batie <batie@aahz.jf.intel.com>]
11077 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11078 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11080 *) New program nseq to manipulate netscape certificate sequences
11083 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11087 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11088 but the BN code had some problems that would cause failures when
11089 doing certificate verification and some other functions.
11090 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11092 *) Add ASN1 and PEM code to support netscape certificate sequences.
11095 *) Add ASN1 and PEM code to support netscape certificate sequences.
11098 *) Add several PKIX and private extended key usage OIDs.
11101 *) Modify the 'ca' program to handle the new extension code. Modify
11102 openssl.cnf for new extension format, add comments.
11105 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11106 and add a sample to openssl.cnf so req -x509 now adds appropriate
11110 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11111 error code, add initial support to X509_print() and x509 application.
11114 *) Takes a deep breath and start addding X509 V3 extension support code. Add
11115 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11116 stuff is currently isolated and isn't even compiled yet.
11119 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11120 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11121 Removed the versions check from X509 routines when loading extensions:
11122 this allows certain broken certificates that don't set the version
11123 properly to be processed.
11126 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11127 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11128 can still be regenerated with "make depend".
11131 *) Spelling mistake in C version of CAST-128.
11132 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11134 *) Changes to the error generation code. The perl script err-code.pl
11135 now reads in the old error codes and retains the old numbers, only
11136 adding new ones if necessary. It also only changes the .err files if new
11137 codes are added. The makefiles have been modified to only insert errors
11138 when needed (to avoid needlessly modifying header files). This is done
11139 by only inserting errors if the .err file is newer than the auto generated
11140 C file. To rebuild all the error codes from scratch (the old behaviour)
11141 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11142 or delete all the .err files.
11145 *) CAST-128 was incorrectly implemented for short keys. The C version has
11146 been fixed, but is untested. The assembler versions are also fixed, but
11147 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11148 to regenerate it if needed.
11149 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11150 Hagino <itojun@kame.net>]
11152 *) File was opened incorrectly in randfile.c.
11153 [Ulf Möller <ulf@fitug.de>]
11155 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11156 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11157 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11158 al: it's just almost always a UTCTime. Note this patch adds new error
11159 codes so do a "make errors" if there are problems.
11162 *) Correct Linux 1 recognition in config.
11163 [Ulf Möller <ulf@fitug.de>]
11165 *) Remove pointless MD5 hash when using DSA keys in ca.
11166 [Anonymous <nobody@replay.com>]
11168 *) Generate an error if given an empty string as a cert directory. Also
11169 generate an error if handed NULL (previously returned 0 to indicate an
11170 error, but didn't set one).
11171 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11173 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11176 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11177 parameters. This was causing a warning which killed off the Win32 compile.
11180 *) Remove C++ style comments from crypto/bn/bn_local.h.
11181 [Neil Costigan <neil.costigan@celocom.com>]
11183 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11184 based on a text string, looking up short and long names and finally
11185 "dot" format. The "dot" format stuff didn't work. Added new function
11186 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11187 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11188 OID is not part of the table.
11191 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11192 X509_LOOKUP_by_alias().
11195 *) Sort openssl functions by name.
11198 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11199 encryption from sample DSA keys (in case anyone is interested the password
11203 *) Make _all_ *_free functions accept a NULL pointer.
11204 [Frans Heymans <fheymans@isaserver.be>]
11206 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11208 [Anonymous <nobody@replay.com>]
11210 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11211 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11213 *) Don't blow it for numeric -newkey arguments to apps/req.
11214 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11216 *) Temp key "for export" tests were wrong in s3_srvr.c.
11217 [Anonymous <nobody@replay.com>]
11219 *) Add prototype for temp key callback functions
11220 SSL_CTX_set_tmp_{rsa,dh}_callback().
11223 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11224 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11227 *) X509_name_add_entry() freed the wrong thing after an error.
11228 [Arne Ansper <arne@ats.cyber.ee>]
11230 *) rsa_eay.c would attempt to free a NULL context.
11231 [Arne Ansper <arne@ats.cyber.ee>]
11233 *) BIO_s_socket() had a broken should_retry() on Windoze.
11234 [Arne Ansper <arne@ats.cyber.ee>]
11236 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11237 [Arne Ansper <arne@ats.cyber.ee>]
11239 *) Make sure the already existing X509_STORE->depth variable is initialized
11240 in X509_STORE_new(), but document the fact that this variable is still
11241 unused in the certificate verification process.
11242 [Ralf S. Engelschall]
11244 *) Fix the various library and apps files to free up pkeys obtained from
11245 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11248 *) Fix reference counting in X509_PUBKEY_get(). This makes
11249 demos/maurice/example2.c work, amongst others, probably.
11250 [Steve Henson and Ben Laurie]
11252 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11253 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11254 are no longer created. This way we have a single and consistent command
11255 line interface `openssl <command>', similar to `cvs <command>'.
11256 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11258 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11259 BIT STRING wrapper always have zero unused bits.
11262 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11265 *) Make the top-level INSTALL documentation easier to understand.
11268 *) Makefiles updated to exit if an error occurs in a sub-directory
11269 make (including if user presses ^C) [Paul Sutton]
11271 *) Make Montgomery context stuff explicit in RSA data structure.
11274 *) Fix build order of pem and err to allow for generated pem.h.
11277 *) Fix renumbering bug in X509_NAME_delete_entry().
11280 *) Enhanced the err-ins.pl script so it makes the error library number
11281 global and can add a library name. This is needed for external ASN1 and
11282 other error libraries.
11285 *) Fixed sk_insert which never worked properly.
11288 *) Fix ASN1 macros so they can handle indefinite length construted
11289 EXPLICIT tags. Some non standard certificates use these: they can now
11293 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11294 into a single doc/ssleay.txt bundle. This way the information is still
11295 preserved but no longer messes up this directory. Now it's new room for
11296 the new set of documenation files.
11297 [Ralf S. Engelschall]
11299 *) SETs were incorrectly DER encoded. This was a major pain, because they
11300 shared code with SEQUENCEs, which aren't coded the same. This means that
11301 almost everything to do with SETs or SEQUENCEs has either changed name or
11302 number of arguments.
11303 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11305 *) Fix test data to work with the above.
11308 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11309 was already fixed by Eric for 0.9.1 it seems.
11310 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11312 *) Autodetect FreeBSD3.
11315 *) Fix various bugs in Configure. This affects the following platforms:
11319 unixware-2.0-pentium
11323 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11324 before they are needed.
11327 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11331 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11333 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11334 changed SSLeay to OpenSSL in version strings.
11335 [Ralf S. Engelschall]
11337 *) Some fixups to the top-level documents.
11340 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11341 because the symlink to include/ was missing.
11342 [Ralf S. Engelschall]
11344 *) Incorporated the popular no-RSA/DSA-only patches
11345 which allow to compile a RSA-free SSLeay.
11346 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11348 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11349 when "ssleay" is still not found.
11350 [Ralf S. Engelschall]
11352 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11353 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11355 *) Updated the README file.
11356 [Ralf S. Engelschall]
11358 *) Added various .cvsignore files in the CVS repository subdirs
11359 to make a "cvs update" really silent.
11360 [Ralf S. Engelschall]
11362 *) Recompiled the error-definition header files and added
11363 missing symbols to the Win32 linker tables.
11364 [Ralf S. Engelschall]
11366 *) Cleaned up the top-level documents;
11367 o new files: CHANGES and LICENSE
11368 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11369 o merged COPYRIGHT into LICENSE
11370 o removed obsolete TODO file
11371 o renamed MICROSOFT to INSTALL.W32
11372 [Ralf S. Engelschall]
11374 *) Removed dummy files from the 0.9.1b source tree:
11375 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11376 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11377 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11378 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11379 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11380 [Ralf S. Engelschall]
11382 *) Added various platform portability fixes.
11385 *) The Genesis of the OpenSSL rpject:
11386 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11387 Young and Tim J. Hudson created while they were working for C2Net until
11389 [The OpenSSL Project]
11392 Changes between 0.9.0b and 0.9.1b [not released]
11394 *) Updated a few CA certificates under certs/
11397 *) Changed some BIGNUM api stuff.
11400 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11401 DGUX x86, Linux Alpha, etc.
11404 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11405 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11409 *) Add -strparse option to asn1pars program which parses nested
11411 [Dr Stephen Henson <shenson@bigfoot.com>]
11413 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11416 *) DSA fix for "ca" program.
11419 *) Added "-genkey" option to "dsaparam" program.
11422 *) Added RIPE MD160 (rmd160) message digest.
11425 *) Added -a (all) option to "ssleay version" command.
11428 *) Added PLATFORM define which is the id given to Configure.
11431 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11434 *) Extended the ASN.1 parser routines.
11437 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11440 *) Added a BN_CTX to the BN library.
11443 *) Fixed the weak key values in DES library
11446 *) Changed API in EVP library for cipher aliases.
11449 *) Added support for RC2/64bit cipher.
11452 *) Converted the lhash library to the crypto/mem.c functions.
11455 *) Added more recognized ASN.1 object ids.
11458 *) Added more RSA padding checks for SSL/TLS.
11461 *) Added BIO proxy/filter functionality.
11464 *) Added extra_certs to SSL_CTX which can be used
11465 send extra CA certificates to the client in the CA cert chain sending
11466 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11469 *) Now Fortezza is denied in the authentication phase because
11470 this is key exchange mechanism is not supported by SSLeay at all.
11473 *) Additional PKCS1 checks.
11476 *) Support the string "TLSv1" for all TLS v1 ciphers.
11479 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11480 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11483 *) Fixed a few memory leaks.
11486 *) Fixed various code and comment typos.
11489 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11490 bytes sent in the client random.
11491 [Edward Bishop <ebishop@spyglass.com>]