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 1.0.2l and 1.0.2m [xx XXX xxxx]
14 Changes between 1.0.2k and 1.0.2l [25 May 2017]
16 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
17 platform rather than 'mingw'.
20 Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
22 *) Truncated packet could crash via OOB read
24 If one side of an SSL/TLS path is running on a 32-bit host and a specific
25 cipher is being used, then a truncated packet can cause that host to
26 perform an out-of-bounds read, usually resulting in a crash.
28 This issue was reported to OpenSSL by Robert Święcki of Google.
32 *) BN_mod_exp may produce incorrect results on x86_64
34 There is a carry propagating bug in the x86_64 Montgomery squaring
35 procedure. No EC algorithms are affected. Analysis suggests that attacks
36 against RSA and DSA as a result of this defect would be very difficult to
37 perform and are not believed likely. Attacks against DH are considered just
38 feasible (although very difficult) because most of the work necessary to
39 deduce information about a private key may be performed offline. The amount
40 of resources required for such an attack would be very significant and
41 likely only accessible to a limited number of attackers. An attacker would
42 additionally need online access to an unpatched system using the target
43 private key in a scenario with persistent DH parameters and a private
44 key that is shared between multiple clients. For example this can occur by
45 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
46 similar to CVE-2015-3193 but must be treated as a separate problem.
48 This issue was reported to OpenSSL by the OSS-Fuzz project.
52 *) Montgomery multiplication may produce incorrect results
54 There is a carry propagating bug in the Broadwell-specific Montgomery
55 multiplication procedure that handles input lengths divisible by, but
56 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
57 and DH private keys are impossible. This is because the subroutine in
58 question is not used in operations with the private key itself and an input
59 of the attacker's direct choice. Otherwise the bug can manifest itself as
60 transient authentication and key negotiation failures or reproducible
61 erroneous outcome of public-key operations with specially crafted input.
62 Among EC algorithms only Brainpool P-512 curves are affected and one
63 presumably can attack ECDH key negotiation. Impact was not analyzed in
64 detail, because pre-requisites for attack are considered unlikely. Namely
65 multiple clients have to choose the curve in question and the server has to
66 share the private key among them, neither of which is default behaviour.
67 Even then only clients that chose the curve will be affected.
69 This issue was publicly reported as transient failures and was not
70 initially recognized as a security issue. Thanks to Richard Morgan for
71 providing reproducible case.
75 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
76 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
77 prevent issues where no progress is being made and the peer continually
78 sends unrecognised record types, using up resources processing them.
81 Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
83 *) Missing CRL sanity check
85 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
86 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
87 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
89 This issue only affects the OpenSSL 1.0.2i
93 Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
95 *) OCSP Status Request extension unbounded memory growth
97 A malicious client can send an excessively large OCSP Status Request
98 extension. If that client continually requests renegotiation, sending a
99 large OCSP Status Request extension each time, then there will be unbounded
100 memory growth on the server. This will eventually lead to a Denial Of
101 Service attack through memory exhaustion. Servers with a default
102 configuration are vulnerable even if they do not support OCSP. Builds using
103 the "no-ocsp" build time option are not affected.
105 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
109 *) In order to mitigate the SWEET32 attack, the DES ciphers were moved from
112 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
117 *) OOB write in MDC2_Update()
119 An overflow can occur in MDC2_Update() either if called directly or
120 through the EVP_DigestUpdate() function using MDC2. If an attacker
121 is able to supply very large amounts of input data after a previous
122 call to EVP_EncryptUpdate() with a partial block then a length check
123 can overflow resulting in a heap corruption.
125 The amount of data needed is comparable to SIZE_MAX which is impractical
128 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
132 *) Malformed SHA512 ticket DoS
134 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
135 DoS attack where a malformed ticket will result in an OOB read which will
138 The use of SHA512 in TLS session tickets is comparatively rare as it requires
139 a custom server callback and ticket lookup mechanism.
141 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
145 *) OOB write in BN_bn2dec()
147 The function BN_bn2dec() does not check the return value of BN_div_word().
148 This can cause an OOB write if an application uses this function with an
149 overly large BIGNUM. This could be a problem if an overly large certificate
150 or CRL is printed out from an untrusted source. TLS is not affected because
151 record limits will reject an oversized certificate before it is parsed.
153 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
157 *) OOB read in TS_OBJ_print_bio()
159 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
160 the total length the OID text representation would use and not the amount
161 of data written. This will result in OOB reads when large OIDs are
164 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
168 *) Pointer arithmetic undefined behaviour
170 Avoid some undefined pointer arithmetic
172 A common idiom in the codebase is to check limits in the following manner:
175 Where "p" points to some malloc'd data of SIZE bytes and
178 "len" here could be from some externally supplied data (e.g. from a TLS
181 The rules of C pointer arithmetic are such that "p + len" is only well
182 defined where len <= SIZE. Therefore the above idiom is actually
185 For example this could cause problems if some malloc implementation
186 provides an address for "p" such that "p + len" actually overflows for
187 values of len that are too big and therefore p + len < limit.
189 This issue was reported to OpenSSL by Guido Vranken
193 *) Constant time flag not preserved in DSA signing
195 Operations in the DSA signing algorithm should run in constant time in
196 order to avoid side channel attacks. A flaw in the OpenSSL DSA
197 implementation means that a non-constant time codepath is followed for
198 certain operations. This has been demonstrated through a cache-timing
199 attack to be sufficient for an attacker to recover the private DSA key.
201 This issue was reported by César Pereida (Aalto University), Billy Brumley
202 (Tampere University of Technology), and Yuval Yarom (The University of
207 *) DTLS buffered message DoS
209 In a DTLS connection where handshake messages are delivered out-of-order
210 those messages that OpenSSL is not yet ready to process will be buffered
211 for later use. Under certain circumstances, a flaw in the logic means that
212 those messages do not get removed from the buffer even though the handshake
213 has been completed. An attacker could force up to approx. 15 messages to
214 remain in the buffer when they are no longer required. These messages will
215 be cleared when the DTLS connection is closed. The default maximum size for
216 a message is 100k. Therefore the attacker could force an additional 1500k
217 to be consumed per connection. By opening many simulataneous connections an
218 attacker could cause a DoS attack through memory exhaustion.
220 This issue was reported to OpenSSL by Quan Luo.
224 *) DTLS replay protection DoS
226 A flaw in the DTLS replay attack protection mechanism means that records
227 that arrive for future epochs update the replay protection "window" before
228 the MAC for the record has been validated. This could be exploited by an
229 attacker by sending a record for the next epoch (which does not have to
230 decrypt or have a valid MAC), with a very large sequence number. This means
231 that all subsequent legitimate packets are dropped causing a denial of
232 service for a specific DTLS connection.
234 This issue was reported to OpenSSL by the OCAP audit team.
238 *) Certificate message OOB reads
240 In OpenSSL 1.0.2 and earlier some missing message length checks can result
241 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
242 theoretical DoS risk but this has not been observed in practice on common
245 The messages affected are client certificate, client certificate request
246 and server certificate. As a result the attack can only be performed
247 against a client or a server which enables client authentication.
249 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
253 Changes between 1.0.2g and 1.0.2h [3 May 2016]
255 *) Prevent padding oracle in AES-NI CBC MAC check
257 A MITM attacker can use a padding oracle attack to decrypt traffic
258 when the connection uses an AES CBC cipher and the server support
261 This issue was introduced as part of the fix for Lucky 13 padding
262 attack (CVE-2013-0169). The padding check was rewritten to be in
263 constant time by making sure that always the same bytes are read and
264 compared against either the MAC or padding bytes. But it no longer
265 checked that there was enough data to have both the MAC and padding
268 This issue was reported by Juraj Somorovsky using TLS-Attacker.
272 *) Fix EVP_EncodeUpdate overflow
274 An overflow can occur in the EVP_EncodeUpdate() function which is used for
275 Base64 encoding of binary data. If an attacker is able to supply very large
276 amounts of input data then a length check can overflow resulting in a heap
279 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
280 the PEM_write_bio* family of functions. These are mainly used within the
281 OpenSSL command line applications, so any application which processes data
282 from an untrusted source and outputs it as a PEM file should be considered
283 vulnerable to this issue. User applications that call these APIs directly
284 with large amounts of untrusted data may also be vulnerable.
286 This issue was reported by Guido Vranken.
290 *) Fix EVP_EncryptUpdate overflow
292 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
293 is able to supply very large amounts of input data after a previous call to
294 EVP_EncryptUpdate() with a partial block then a length check can overflow
295 resulting in a heap corruption. Following an analysis of all OpenSSL
296 internal usage of the EVP_EncryptUpdate() function all usage is one of two
297 forms. The first form is where the EVP_EncryptUpdate() call is known to be
298 the first called function after an EVP_EncryptInit(), and therefore that
299 specific call must be safe. The second form is where the length passed to
300 EVP_EncryptUpdate() can be seen from the code to be some small value and
301 therefore there is no possibility of an overflow. Since all instances are
302 one of these two forms, it is believed that there can be no overflows in
303 internal code due to this problem. It should be noted that
304 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
305 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
306 of these calls have also been analysed too and it is believed there are no
307 instances in internal usage where an overflow could occur.
309 This issue was reported by Guido Vranken.
313 *) Prevent ASN.1 BIO excessive memory allocation
315 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
316 a short invalid encoding can casuse allocation of large amounts of memory
317 potentially consuming excessive resources or exhausting memory.
319 Any application parsing untrusted data through d2i BIO functions is
320 affected. The memory based functions such as d2i_X509() are *not* affected.
321 Since the memory based functions are used by the TLS library, TLS
322 applications are not affected.
324 This issue was reported by Brian Carpenter.
330 ASN1 Strings that are over 1024 bytes can cause an overread in applications
331 using the X509_NAME_oneline() function on EBCDIC systems. This could result
332 in arbitrary stack data being returned in the buffer.
334 This issue was reported by Guido Vranken.
338 *) Modify behavior of ALPN to invoke callback after SNI/servername
339 callback, such that updates to the SSL_CTX affect ALPN.
342 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
346 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
347 methods are enabled and ssl2 is disabled the methods return NULL.
350 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
352 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
353 Builds that are not configured with "enable-weak-ssl-ciphers" will not
354 provide any "EXPORT" or "LOW" strength ciphers.
357 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
358 is by default disabled at build-time. Builds that are not configured with
359 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
360 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
361 will need to explicitly call either of:
363 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
365 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
367 as appropriate. Even if either of those is used, or the application
368 explicitly uses the version-specific SSLv2_method() or its client and
369 server variants, SSLv2 ciphers vulnerable to exhaustive search key
370 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
371 ciphers, and SSLv2 56-bit DES are no longer available.
375 *) Fix a double-free in DSA code
377 A double free bug was discovered when OpenSSL parses malformed DSA private
378 keys and could lead to a DoS attack or memory corruption for applications
379 that receive DSA private keys from untrusted sources. This scenario is
382 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
387 *) Disable SRP fake user seed to address a server memory leak.
389 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
391 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
392 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
393 was changed to ignore the "fake user" SRP seed, even if the seed
396 Users should use SRP_VBASE_get1_by_user instead. Note that in
397 SRP_VBASE_get1_by_user, caller must free the returned value. Note
398 also that even though configuring the SRP seed attempts to hide
399 invalid usernames by continuing the handshake with fake
400 credentials, this behaviour is not constant time and no strong
401 guarantees are made that the handshake is indistinguishable from
402 that of a valid user.
406 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
408 In the BN_hex2bn function the number of hex digits is calculated using an
409 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
410 large values of |i| this can result in |bn_expand| not allocating any
411 memory because |i * 4| is negative. This can leave the internal BIGNUM data
412 field as NULL leading to a subsequent NULL ptr deref. For very large values
413 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
414 In this case memory is allocated to the internal BIGNUM data field, but it
415 is insufficiently sized leading to heap corruption. A similar issue exists
416 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
417 is ever called by user applications with very large untrusted hex/dec data.
418 This is anticipated to be a rare occurrence.
420 All OpenSSL internal usage of these functions use data that is not expected
421 to be untrusted, e.g. config file data or application command line
422 arguments. If user developed applications generate config file data based
423 on untrusted data then it is possible that this could also lead to security
424 consequences. This is also anticipated to be rare.
426 This issue was reported to OpenSSL by Guido Vranken.
430 *) Fix memory issues in BIO_*printf functions
432 The internal |fmtstr| function used in processing a "%s" format string in
433 the BIO_*printf functions could overflow while calculating the length of a
434 string and cause an OOB read when printing very long strings.
436 Additionally the internal |doapr_outch| function can attempt to write to an
437 OOB memory location (at an offset from the NULL pointer) in the event of a
438 memory allocation failure. In 1.0.2 and below this could be caused where
439 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
440 could be in processing a very long "%s" format string. Memory leaks can
443 The first issue may mask the second issue dependent on compiler behaviour.
444 These problems could enable attacks where large amounts of untrusted data
445 is passed to the BIO_*printf functions. If applications use these functions
446 in this way then they could be vulnerable. OpenSSL itself uses these
447 functions when printing out human-readable dumps of ASN.1 data. Therefore
448 applications that print this data could be vulnerable if the data is from
449 untrusted sources. OpenSSL command line applications could also be
450 vulnerable where they print out ASN.1 data, or if untrusted data is passed
451 as command line arguments.
453 Libssl is not considered directly vulnerable. Additionally certificates etc
454 received via remote connections via libssl are also unlikely to be able to
455 trigger these issues because of message size limits enforced within libssl.
457 This issue was reported to OpenSSL Guido Vranken.
461 *) Side channel attack on modular exponentiation
463 A side-channel attack was found which makes use of cache-bank conflicts on
464 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
465 of RSA keys. The ability to exploit this issue is limited as it relies on
466 an attacker who has control of code in a thread running on the same
467 hyper-threaded core as the victim thread which is performing decryptions.
469 This issue was reported to OpenSSL by Yuval Yarom, The University of
470 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
471 Nadia Heninger, University of Pennsylvania with more information at
472 http://cachebleed.info.
476 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
477 if no keysize is specified with default_bits. This fixes an
478 omission in an earlier change that changed all RSA/DSA key generation
479 apps to use 2048 bits by default.
482 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
484 *) DH small subgroups
486 Historically OpenSSL only ever generated DH parameters based on "safe"
487 primes. More recently (in version 1.0.2) support was provided for
488 generating X9.42 style parameter files such as those required for RFC 5114
489 support. The primes used in such files may not be "safe". Where an
490 application is using DH configured with parameters based on primes that are
491 not "safe" then an attacker could use this fact to find a peer's private
492 DH exponent. This attack requires that the attacker complete multiple
493 handshakes in which the peer uses the same private DH exponent. For example
494 this could be used to discover a TLS server's private DH exponent if it's
495 reusing the private DH exponent or it's using a static DH ciphersuite.
497 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
498 TLS. It is not on by default. If the option is not set then the server
499 reuses the same private DH exponent for the life of the server process and
500 would be vulnerable to this attack. It is believed that many popular
501 applications do set this option and would therefore not be at risk.
503 The fix for this issue adds an additional check where a "q" parameter is
504 available (as is the case in X9.42 based parameters). This detects the
505 only known attack, and is the only possible defense for static DH
506 ciphersuites. This could have some performance impact.
508 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
509 default and cannot be disabled. This could have some performance impact.
511 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
515 *) SSLv2 doesn't block disabled ciphers
517 A malicious client can negotiate SSLv2 ciphers that have been disabled on
518 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
519 been disabled, provided that the SSLv2 protocol was not also disabled via
522 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
523 and Sebastian Schinzel.
527 *) Reject DH handshakes with parameters shorter than 1024 bits.
530 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
532 *) BN_mod_exp may produce incorrect results on x86_64
534 There is a carry propagating bug in the x86_64 Montgomery squaring
535 procedure. No EC algorithms are affected. Analysis suggests that attacks
536 against RSA and DSA as a result of this defect would be very difficult to
537 perform and are not believed likely. Attacks against DH are considered just
538 feasible (although very difficult) because most of the work necessary to
539 deduce information about a private key may be performed offline. The amount
540 of resources required for such an attack would be very significant and
541 likely only accessible to a limited number of attackers. An attacker would
542 additionally need online access to an unpatched system using the target
543 private key in a scenario with persistent DH parameters and a private
544 key that is shared between multiple clients. For example this can occur by
545 default in OpenSSL DHE based SSL/TLS ciphersuites.
547 This issue was reported to OpenSSL by Hanno Böck.
551 *) Certificate verify crash with missing PSS parameter
553 The signature verification routines will crash with a NULL pointer
554 dereference if presented with an ASN.1 signature using the RSA PSS
555 algorithm and absent mask generation function parameter. Since these
556 routines are used to verify certificate signature algorithms this can be
557 used to crash any certificate verification operation and exploited in a
558 DoS attack. Any application which performs certificate verification is
559 vulnerable including OpenSSL clients and servers which enable client
562 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
566 *) X509_ATTRIBUTE memory leak
568 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
569 memory. This structure is used by the PKCS#7 and CMS routines so any
570 application which reads PKCS#7 or CMS data from untrusted sources is
571 affected. SSL/TLS is not affected.
573 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
578 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
579 This changes the decoding behaviour for some invalid messages,
580 though the change is mostly in the more lenient direction, and
581 legacy behaviour is preserved as much as possible.
584 *) In DSA_generate_parameters_ex, if the provided seed is too short,
585 use a random seed, as already documented.
586 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
588 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
590 *) Alternate chains certificate forgery
592 During certificate verfification, OpenSSL will attempt to find an
593 alternative certificate chain if the first attempt to build such a chain
594 fails. An error in the implementation of this logic can mean that an
595 attacker could cause certain checks on untrusted certificates to be
596 bypassed, such as the CA flag, enabling them to use a valid leaf
597 certificate to act as a CA and "issue" an invalid certificate.
599 This issue was reported to OpenSSL by Adam Langley/David Benjamin
604 *) Race condition handling PSK identify hint
606 If PSK identity hints are received by a multi-threaded client then
607 the values are wrongly updated in the parent SSL_CTX structure. This can
608 result in a race condition potentially leading to a double free of the
613 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
615 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
616 incompatibility in the handling of HMAC. The previous ABI has now been
619 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
621 *) Malformed ECParameters causes infinite loop
623 When processing an ECParameters structure OpenSSL enters an infinite loop
624 if the curve specified is over a specially malformed binary polynomial
627 This can be used to perform denial of service against any
628 system which processes public keys, certificate requests or
629 certificates. This includes TLS clients and TLS servers with
630 client authentication enabled.
632 This issue was reported to OpenSSL by Joseph Barr-Pixton.
636 *) Exploitable out-of-bounds read in X509_cmp_time
638 X509_cmp_time does not properly check the length of the ASN1_TIME
639 string and can read a few bytes out of bounds. In addition,
640 X509_cmp_time accepts an arbitrary number of fractional seconds in the
643 An attacker can use this to craft malformed certificates and CRLs of
644 various sizes and potentially cause a segmentation fault, resulting in
645 a DoS on applications that verify certificates or CRLs. TLS clients
646 that verify CRLs are affected. TLS clients and servers with client
647 authentication enabled may be affected if they use custom verification
650 This issue was reported to OpenSSL by Robert Swiecki (Google), and
651 independently by Hanno Böck.
655 *) PKCS7 crash with missing EnvelopedContent
657 The PKCS#7 parsing code does not handle missing inner EncryptedContent
658 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
659 with missing content and trigger a NULL pointer dereference on parsing.
661 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
662 structures from untrusted sources are affected. OpenSSL clients and
663 servers are not affected.
665 This issue was reported to OpenSSL by Michal Zalewski (Google).
669 *) CMS verify infinite loop with unknown hash function
671 When verifying a signedData message the CMS code can enter an infinite loop
672 if presented with an unknown hash function OID. This can be used to perform
673 denial of service against any system which verifies signedData messages using
675 This issue was reported to OpenSSL by Johannes Bauer.
679 *) Race condition handling NewSessionTicket
681 If a NewSessionTicket is received by a multi-threaded client when attempting to
682 reuse a previous ticket then a race condition can occur potentially leading to
683 a double free of the ticket data.
687 *) Removed support for the two export grade static DH ciphersuites
688 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
689 were newly added (along with a number of other static DH ciphersuites) to
690 1.0.2. However the two export ones have *never* worked since they were
691 introduced. It seems strange in any case to be adding new export
692 ciphersuites, and given "logjam" it also does not seem correct to fix them.
695 *) Only support 256-bit or stronger elliptic curves with the
696 'ecdh_auto' setting (server) or by default (client). Of supported
697 curves, prefer P-256 (both).
700 *) Reject DH handshakes with parameters shorter than 768 bits.
701 [Kurt Roeckx and Emilia Kasper]
703 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
705 *) ClientHello sigalgs DoS fix
707 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
708 invalid signature algorithms extension a NULL pointer dereference will
709 occur. This can be exploited in a DoS attack against the server.
711 This issue was was reported to OpenSSL by David Ramos of Stanford
714 [Stephen Henson and Matt Caswell]
716 *) Multiblock corrupted pointer fix
718 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
719 feature only applies on 64 bit x86 architecture platforms that support AES
720 NI instructions. A defect in the implementation of "multiblock" can cause
721 OpenSSL's internal write buffer to become incorrectly set to NULL when
722 using non-blocking IO. Typically, when the user application is using a
723 socket BIO for writing, this will only result in a failed connection.
724 However if some other BIO is used then it is likely that a segmentation
725 fault will be triggered, thus enabling a potential DoS attack.
727 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
731 *) Segmentation fault in DTLSv1_listen fix
733 The DTLSv1_listen function is intended to be stateless and processes the
734 initial ClientHello from many peers. It is common for user code to loop
735 over the call to DTLSv1_listen until a valid ClientHello is received with
736 an associated cookie. A defect in the implementation of DTLSv1_listen means
737 that state is preserved in the SSL object from one invocation to the next
738 that can lead to a segmentation fault. Errors processing the initial
739 ClientHello can trigger this scenario. An example of such an error could be
740 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
743 This issue was reported to OpenSSL by Per Allansson.
747 *) Segmentation fault in ASN1_TYPE_cmp fix
749 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
750 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
751 certificate signature algorithm consistency this can be used to crash any
752 certificate verification operation and exploited in a DoS attack. Any
753 application which performs certificate verification is vulnerable including
754 OpenSSL clients and servers which enable client authentication.
758 *) Segmentation fault for invalid PSS parameters fix
760 The signature verification routines will crash with a NULL pointer
761 dereference if presented with an ASN.1 signature using the RSA PSS
762 algorithm and invalid parameters. Since these routines are used to verify
763 certificate signature algorithms this can be used to crash any
764 certificate verification operation and exploited in a DoS attack. Any
765 application which performs certificate verification is vulnerable including
766 OpenSSL clients and servers which enable client authentication.
768 This issue was was reported to OpenSSL by Brian Carpenter.
772 *) ASN.1 structure reuse memory corruption fix
774 Reusing a structure in ASN.1 parsing may allow an attacker to cause
775 memory corruption via an invalid write. Such reuse is and has been
776 strongly discouraged and is believed to be rare.
778 Applications that parse structures containing CHOICE or ANY DEFINED BY
779 components may be affected. Certificate parsing (d2i_X509 and related
780 functions) are however not affected. OpenSSL clients and servers are
785 *) PKCS7 NULL pointer dereferences fix
787 The PKCS#7 parsing code does not handle missing outer ContentInfo
788 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
789 missing content and trigger a NULL pointer dereference on parsing.
791 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
792 otherwise parse PKCS#7 structures from untrusted sources are
793 affected. OpenSSL clients and servers are not affected.
795 This issue was reported to OpenSSL by Michal Zalewski (Google).
799 *) DoS via reachable assert in SSLv2 servers fix
801 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
802 servers that both support SSLv2 and enable export cipher suites by sending
803 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
805 This issue was discovered by Sean Burford (Google) and Emilia Käsper
806 (OpenSSL development team).
810 *) Empty CKE with client auth and DHE fix
812 If client auth is used then a server can seg fault in the event of a DHE
813 ciphersuite being selected and a zero length ClientKeyExchange message
814 being sent by the client. This could be exploited in a DoS attack.
818 *) Handshake with unseeded PRNG fix
820 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
821 with an unseeded PRNG. The conditions are:
822 - The client is on a platform where the PRNG has not been seeded
823 automatically, and the user has not seeded manually
824 - A protocol specific client method version has been used (i.e. not
825 SSL_client_methodv23)
826 - A ciphersuite is used that does not require additional random data from
827 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
829 If the handshake succeeds then the client random that has been used will
830 have been generated from a PRNG with insufficient entropy and therefore the
831 output may be predictable.
833 For example using the following command with an unseeded openssl will
834 succeed on an unpatched platform:
836 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
840 *) Use After Free following d2i_ECPrivatekey error fix
842 A malformed EC private key file consumed via the d2i_ECPrivateKey function
843 could cause a use after free condition. This, in turn, could cause a double
844 free in several private key parsing functions (such as d2i_PrivateKey
845 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
846 for applications that receive EC private keys from untrusted
847 sources. This scenario is considered rare.
849 This issue was discovered by the BoringSSL project and fixed in their
854 *) X509_to_X509_REQ NULL pointer deref fix
856 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
857 the certificate key is invalid. This function is rarely used in practice.
859 This issue was discovered by Brian Carpenter.
863 *) Removed the export ciphers from the DEFAULT ciphers
866 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
868 *) Change RSA and DH/DSA key generation apps to generate 2048-bit
872 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
873 ARMv5 through ARMv8, as opposite to "locking" it to single one.
874 So far those who have to target multiple plaforms would compromise
875 and argue that binary targeting say ARMv5 would still execute on
876 ARMv8. "Universal" build resolves this compromise by providing
877 near-optimal performance even on newer platforms.
880 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
881 (other platforms pending).
882 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
884 *) Add support for the SignedCertificateTimestampList certificate and
885 OCSP response extensions from RFC6962.
888 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
889 for corner cases. (Certain input points at infinity could lead to
890 bogus results, with non-infinity inputs mapped to infinity too.)
893 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
894 This covers AES, SHA256/512 and GHASH. "Initial" means that most
895 common cases are optimized and there still is room for further
896 improvements. Vector Permutation AES for Altivec is also added.
899 *) Add support for little-endian ppc64 Linux target.
900 [Marcelo Cerri (IBM)]
902 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
903 SHA1, SHA256 and GHASH. "Initial" means that most common cases
904 are optimized and there still is room for further improvements.
905 Both 32- and 64-bit modes are supported.
906 [Andy Polyakov, Ard Biesheuvel (Linaro)]
908 *) Improved ARMv7 NEON support.
911 *) Support for SPARC Architecture 2011 crypto extensions, first
912 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
913 SHA256/512, MD5, GHASH and modular exponentiation.
914 [Andy Polyakov, David Miller]
916 *) Accelerated modular exponentiation for Intel processors, a.k.a.
918 [Shay Gueron & Vlad Krasnov (Intel Corp)]
920 *) Support for new and upcoming Intel processors, including AVX2,
921 BMI and SHA ISA extensions. This includes additional "stitched"
922 implementations, AESNI-SHA256 and GCM, and multi-buffer support
925 This work was sponsored by Intel Corp.
928 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
929 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
930 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
933 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
934 this fixes a limiation in previous versions of OpenSSL.
937 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
938 MGF1 digest and OAEP label.
941 *) Add EVP support for key wrapping algorithms, to avoid problems with
942 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
943 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
944 algorithms and include tests cases.
947 *) Add functions to allocate and set the fields of an ECDSA_METHOD
949 [Douglas E. Engert, Steve Henson]
951 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
952 difference in days and seconds between two tm or ASN1_TIME structures.
955 *) Add -rev test option to s_server to just reverse order of characters
956 received by client and send back to server. Also prints an abbreviated
957 summary of the connection parameters.
960 *) New option -brief for s_client and s_server to print out a brief summary
961 of connection parameters.
964 *) Add callbacks for arbitrary TLS extensions.
965 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
967 *) New option -crl_download in several openssl utilities to download CRLs
968 from CRLDP extension in certificates.
971 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
974 *) New function X509_CRL_diff to generate a delta CRL from the difference
975 of two full CRLs. Add support to "crl" utility.
978 *) New functions to set lookup_crls function and to retrieve
979 X509_STORE from X509_STORE_CTX.
982 *) Print out deprecated issuer and subject unique ID fields in
986 *) Extend OCSP I/O functions so they can be used for simple general purpose
987 HTTP as well as OCSP. New wrapper function which can be used to download
988 CRLs using the OCSP API.
991 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
994 *) SSL_CONF* functions. These provide a common framework for application
995 configuration using configuration files or command lines.
998 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
999 message callback and prints the results. Needs compile time option
1000 "enable-ssl-trace". New options to s_client and s_server to enable
1004 *) New ctrl and macro to retrieve supported points extensions.
1005 Print out extension in s_server and s_client.
1008 *) New functions to retrieve certificate signature and signature
1012 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1016 *) New Suite B modes for TLS code. These use and enforce the requirements
1017 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1018 only use Suite B curves. The Suite B modes can be set by using the
1019 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1022 *) New chain verification flags for Suite B levels of security. Check
1023 algorithms are acceptable when flags are set in X509_verify_cert.
1026 *) Make tls1_check_chain return a set of flags indicating checks passed
1027 by a certificate chain. Add additional tests to handle client
1028 certificates: checks for matching certificate type and issuer name
1032 *) If an attempt is made to use a signature algorithm not in the peer
1033 preference list abort the handshake. If client has no suitable
1034 signature algorithms in response to a certificate request do not
1035 use the certificate.
1038 *) If server EC tmp key is not in client preference list abort handshake.
1041 *) Add support for certificate stores in CERT structure. This makes it
1042 possible to have different stores per SSL structure or one store in
1043 the parent SSL_CTX. Include distint stores for certificate chain
1044 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1045 to build and store a certificate chain in CERT structure: returing
1046 an error if the chain cannot be built: this will allow applications
1047 to test if a chain is correctly configured.
1049 Note: if the CERT based stores are not set then the parent SSL_CTX
1050 store is used to retain compatibility with existing behaviour.
1054 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1055 mask based on the current session, check mask when sending client
1056 hello and checking the requested ciphersuite.
1059 *) New ctrls to retrieve and set certificate types in a certificate
1060 request message. Print out received values in s_client. If certificate
1061 types is not set with custom values set sensible values based on
1062 supported signature algorithms.
1065 *) Support for distinct client and server supported signature algorithms.
1068 *) Add certificate callback. If set this is called whenever a certificate
1069 is required by client or server. An application can decide which
1070 certificate chain to present based on arbitrary criteria: for example
1071 supported signature algorithms. Add very simple example to s_server.
1072 This fixes many of the problems and restrictions of the existing client
1073 certificate callback: for example you can now clear an existing
1074 certificate and specify the whole chain.
1077 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1078 the certificate can be used for (if anything). Set valid_flags field
1079 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1080 to have similar checks in it.
1082 Add new "cert_flags" field to CERT structure and include a "strict mode".
1083 This enforces some TLS certificate requirements (such as only permitting
1084 certificate signature algorithms contained in the supported algorithms
1085 extension) which some implementations ignore: this option should be used
1086 with caution as it could cause interoperability issues.
1089 *) Update and tidy signature algorithm extension processing. Work out
1090 shared signature algorithms based on preferences and peer algorithms
1091 and print them out in s_client and s_server. Abort handshake if no
1092 shared signature algorithms.
1095 *) Add new functions to allow customised supported signature algorithms
1096 for SSL and SSL_CTX structures. Add options to s_client and s_server
1100 *) New function SSL_certs_clear() to delete all references to certificates
1101 from an SSL structure. Before this once a certificate had been added
1102 it couldn't be removed.
1105 *) Integrate hostname, email address and IP address checking with certificate
1106 verification. New verify options supporting checking in opensl utility.
1109 *) Fixes and wildcard matching support to hostname and email checking
1110 functions. Add manual page.
1111 [Florian Weimer (Red Hat Product Security Team)]
1113 *) New functions to check a hostname email or IP address against a
1114 certificate. Add options x509 utility to print results of checks against
1118 *) Fix OCSP checking.
1119 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1121 *) Initial experimental support for explicitly trusted non-root CAs.
1122 OpenSSL still tries to build a complete chain to a root but if an
1123 intermediate CA has a trust setting included that is used. The first
1124 setting is used: whether to trust (e.g., -addtrust option to the x509
1128 *) Add -trusted_first option which attempts to find certificates in the
1129 trusted store even if an untrusted chain is also supplied.
1132 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1133 platform support for Linux and Android.
1136 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1139 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1140 When in FIPS mode the approved implementations are used as normal,
1141 when not in FIPS mode the internal unapproved versions are used instead.
1142 This means that the FIPS capable OpenSSL isn't forced to use the
1143 (often lower perfomance) FIPS implementations outside FIPS mode.
1146 *) Transparently support X9.42 DH parameters when calling
1147 PEM_read_bio_DHparameters. This means existing applications can handle
1148 the new parameter format automatically.
1151 *) Initial experimental support for X9.42 DH parameter format: mainly
1152 to support use of 'q' parameter for RFC5114 parameters.
1155 *) Add DH parameters from RFC5114 including test data to dhtest.
1158 *) Support for automatic EC temporary key parameter selection. If enabled
1159 the most preferred EC parameters are automatically used instead of
1160 hardcoded fixed parameters. Now a server just has to call:
1161 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1162 support ECDH and use the most appropriate parameters.
1165 *) Enhance and tidy EC curve and point format TLS extension code. Use
1166 static structures instead of allocation if default values are used.
1167 New ctrls to set curves we wish to support and to retrieve shared curves.
1168 Print out shared curves in s_server. New options to s_server and s_client
1169 to set list of supported curves.
1172 *) New ctrls to retrieve supported signature algorithms and
1173 supported curve values as an array of NIDs. Extend openssl utility
1174 to print out received values.
1177 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1178 between NIDs and the more common NIST names such as "P-256". Enhance
1179 ecparam utility and ECC method to recognise the NIST names for curves.
1182 *) Enhance SSL/TLS certificate chain handling to support different
1183 chains for each certificate instead of one chain in the parent SSL_CTX.
1186 *) Support for fixed DH ciphersuite client authentication: where both
1187 server and client use DH certificates with common parameters.
1190 *) Support for fixed DH ciphersuites: those requiring DH server
1194 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1196 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1197 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1198 X509_CINF_get_signature were reverted post internal team review.
1200 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1202 *) Build fixes for the Windows and OpenVMS platforms
1203 [Matt Caswell and Richard Levitte]
1205 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1207 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1208 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1209 dereference. This could lead to a Denial Of Service attack. Thanks to
1210 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1214 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1215 dtls1_buffer_record function under certain conditions. In particular this
1216 could occur if an attacker sent repeated DTLS records with the same
1217 sequence number but for the next epoch. The memory leak could be exploited
1218 by an attacker in a Denial of Service attack through memory exhaustion.
1219 Thanks to Chris Mueller for reporting this issue.
1223 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1224 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1225 method would be set to NULL which could later result in a NULL pointer
1226 dereference. Thanks to Frank Schmirler for reporting this issue.
1230 *) Abort handshake if server key exchange message is omitted for ephemeral
1233 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1234 reporting this issue.
1238 *) Remove non-export ephemeral RSA code on client and server. This code
1239 violated the TLS standard by allowing the use of temporary RSA keys in
1240 non-export ciphersuites and could be used by a server to effectively
1241 downgrade the RSA key length used to a value smaller than the server
1242 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1243 INRIA or reporting this issue.
1247 *) Fixed issue where DH client certificates are accepted without verification.
1248 An OpenSSL server will accept a DH certificate for client authentication
1249 without the certificate verify message. This effectively allows a client to
1250 authenticate without the use of a private key. This only affects servers
1251 which trust a client certificate authority which issues certificates
1252 containing DH keys: these are extremely rare and hardly ever encountered.
1253 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1258 *) Ensure that the session ID context of an SSL is updated when its
1259 SSL_CTX is updated via SSL_set_SSL_CTX.
1261 The session ID context is typically set from the parent SSL_CTX,
1262 and can vary with the CTX.
1265 *) Fix various certificate fingerprint issues.
1267 By using non-DER or invalid encodings outside the signed portion of a
1268 certificate the fingerprint can be changed without breaking the signature.
1269 Although no details of the signed portion of the certificate can be changed
1270 this can cause problems with some applications: e.g. those using the
1271 certificate fingerprint for blacklists.
1273 1. Reject signatures with non zero unused bits.
1275 If the BIT STRING containing the signature has non zero unused bits reject
1276 the signature. All current signature algorithms require zero unused bits.
1278 2. Check certificate algorithm consistency.
1280 Check the AlgorithmIdentifier inside TBS matches the one in the
1281 certificate signature. NB: this will result in signature failure
1282 errors for some broken certificates.
1284 Thanks to Konrad Kraszewski from Google for reporting this issue.
1286 3. Check DSA/ECDSA signatures use DER.
1288 Reencode DSA/ECDSA signatures and compare with the original received
1289 signature. Return an error if there is a mismatch.
1291 This will reject various cases including garbage after signature
1292 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1293 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1294 (negative or with leading zeroes).
1296 Further analysis was conducted and fixes were developed by Stephen Henson
1297 of the OpenSSL core team.
1302 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1303 results on some platforms, including x86_64. This bug occurs at random
1304 with a very low probability, and is not known to be exploitable in any
1305 way, though its exact impact is difficult to determine. Thanks to Pieter
1306 Wuille (Blockstream) who reported this issue and also suggested an initial
1307 fix. Further analysis was conducted by the OpenSSL development team and
1308 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1309 the OpenSSL core team.
1313 *) Do not resume sessions on the server if the negotiated protocol
1314 version does not match the session's version. Resuming with a different
1315 version, while not strictly forbidden by the RFC, is of questionable
1316 sanity and breaks all known clients.
1317 [David Benjamin, Emilia Käsper]
1319 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1320 early CCS messages during renegotiation. (Note that because
1321 renegotiation is encrypted, this early CCS was not exploitable.)
1324 *) Tighten client-side session ticket handling during renegotiation:
1325 ensure that the client only accepts a session ticket if the server sends
1326 the extension anew in the ServerHello. Previously, a TLS client would
1327 reuse the old extension state and thus accept a session ticket if one was
1328 announced in the initial ServerHello.
1330 Similarly, ensure that the client requires a session ticket if one
1331 was advertised in the ServerHello. Previously, a TLS client would
1332 ignore a missing NewSessionTicket message.
1335 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1337 *) SRTP Memory Leak.
1339 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1340 sends a carefully crafted handshake message, to cause OpenSSL to fail
1341 to free up to 64k of memory causing a memory leak. This could be
1342 exploited in a Denial Of Service attack. This issue affects OpenSSL
1343 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1344 whether SRTP is used or configured. Implementations of OpenSSL that
1345 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1347 The fix was developed by the OpenSSL team.
1351 *) Session Ticket Memory Leak.
1353 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1354 integrity of that ticket is first verified. In the event of a session
1355 ticket integrity check failing, OpenSSL will fail to free memory
1356 causing a memory leak. By sending a large number of invalid session
1357 tickets an attacker could exploit this issue in a Denial Of Service
1362 *) Build option no-ssl3 is incomplete.
1364 When OpenSSL is configured with "no-ssl3" as a build option, servers
1365 could accept and complete a SSL 3.0 handshake, and clients could be
1366 configured to send them.
1368 [Akamai and the OpenSSL team]
1370 *) Add support for TLS_FALLBACK_SCSV.
1371 Client applications doing fallback retries should call
1372 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1374 [Adam Langley, Bodo Moeller]
1376 *) Add additional DigestInfo checks.
1378 Reencode DigestInto in DER and check against the original when
1379 verifying RSA signature: this will reject any improperly encoded
1380 DigestInfo structures.
1382 Note: this is a precautionary measure and no attacks are currently known.
1386 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1388 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1389 SRP code can be overrun an internal buffer. Add sanity check that
1390 g, A, B < N to SRP code.
1392 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1393 Group for discovering this issue.
1397 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1398 TLS 1.0 instead of higher protocol versions when the ClientHello message
1399 is badly fragmented. This allows a man-in-the-middle attacker to force a
1400 downgrade to TLS 1.0 even if both the server and the client support a
1401 higher protocol version, by modifying the client's TLS records.
1403 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1404 researching this issue.
1408 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1409 to a denial of service attack. A malicious server can crash the client
1410 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1411 ciphersuite and sending carefully crafted handshake messages.
1413 Thanks to Felix Gröbert (Google) for discovering and researching this
1418 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1419 to leak memory. This can be exploited through a Denial of Service attack.
1420 Thanks to Adam Langley for discovering and researching this issue.
1424 *) An attacker can force openssl to consume large amounts of memory whilst
1425 processing DTLS handshake messages. This can be exploited through a
1426 Denial of Service attack.
1427 Thanks to Adam Langley for discovering and researching this issue.
1431 *) An attacker can force an error condition which causes openssl to crash
1432 whilst processing DTLS packets due to memory being freed twice. This
1433 can be exploited through a Denial of Service attack.
1434 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1439 *) If a multithreaded client connects to a malicious server using a resumed
1440 session and the server sends an ec point format extension it could write
1441 up to 255 bytes to freed memory.
1443 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1448 *) A malicious server can crash an OpenSSL client with a null pointer
1449 dereference (read) by specifying an SRP ciphersuite even though it was not
1450 properly negotiated with the client. This can be exploited through a
1451 Denial of Service attack.
1453 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1454 discovering and researching this issue.
1458 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1459 X509_name_oneline, X509_name_print_ex et al. to leak some information
1460 from the stack. Applications may be affected if they echo pretty printing
1461 output to the attacker.
1463 Thanks to Ivan Fratric (Google) for discovering this issue.
1465 [Emilia Käsper, and Steve Henson]
1467 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1468 for corner cases. (Certain input points at infinity could lead to
1469 bogus results, with non-infinity inputs mapped to infinity too.)
1472 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1474 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1475 handshake can force the use of weak keying material in OpenSSL
1476 SSL/TLS clients and servers.
1478 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1479 researching this issue. (CVE-2014-0224)
1480 [KIKUCHI Masashi, Steve Henson]
1482 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1483 OpenSSL DTLS client the code can be made to recurse eventually crashing
1486 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1488 [Imre Rad, Steve Henson]
1490 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1491 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1492 client or server. This is potentially exploitable to run arbitrary
1493 code on a vulnerable client or server.
1495 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1496 [Jüri Aedla, Steve Henson]
1498 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1499 are subject to a denial of service attack.
1501 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1502 this issue. (CVE-2014-3470)
1503 [Felix Gröbert, Ivan Fratric, Steve Henson]
1505 *) Harmonize version and its documentation. -f flag is used to display
1507 [mancha <mancha1@zoho.com>]
1509 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1510 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1511 [mancha <mancha1@zoho.com>]
1513 *) Fix some double frees. These are not thought to be exploitable.
1514 [mancha <mancha1@zoho.com>]
1516 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1518 *) A missing bounds check in the handling of the TLS heartbeat extension
1519 can be used to reveal up to 64k of memory to a connected client or
1522 Thanks for Neel Mehta of Google Security for discovering this bug and to
1523 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1524 preparing the fix (CVE-2014-0160)
1525 [Adam Langley, Bodo Moeller]
1527 *) Fix for the attack described in the paper "Recovering OpenSSL
1528 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1529 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1530 http://eprint.iacr.org/2014/140
1532 Thanks to Yuval Yarom and Naomi Benger for discovering this
1533 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1534 [Yuval Yarom and Naomi Benger]
1536 *) TLS pad extension: draft-agl-tls-padding-03
1538 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1539 TLS client Hello record length value would otherwise be > 255 and
1540 less that 512 pad with a dummy extension containing zeroes so it
1541 is at least 512 bytes long.
1543 [Adam Langley, Steve Henson]
1545 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1547 *) Fix for TLS record tampering bug. A carefully crafted invalid
1548 handshake could crash OpenSSL with a NULL pointer exception.
1549 Thanks to Anton Johansson for reporting this issues.
1552 *) Keep original DTLS digest and encryption contexts in retransmission
1553 structures so we can use the previous session parameters if they need
1554 to be resent. (CVE-2013-6450)
1557 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1558 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1559 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1560 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1561 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1562 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1563 [Rob Stradling, Adam Langley]
1565 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1567 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1568 supporting platforms or when small records were transferred.
1569 [Andy Polyakov, Steve Henson]
1571 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1573 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1575 This addresses the flaw in CBC record processing discovered by
1576 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1577 at: http://www.isg.rhul.ac.uk/tls/
1579 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1580 Security Group at Royal Holloway, University of London
1581 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1582 Emilia Käsper for the initial patch.
1584 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1586 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1587 ciphersuites which can be exploited in a denial of service attack.
1588 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1589 and detecting this bug and to Wolfgang Ettlinger
1590 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1594 *) Return an error when checking OCSP signatures when key is NULL.
1595 This fixes a DoS attack. (CVE-2013-0166)
1598 *) Make openssl verify return errors.
1599 [Chris Palmer <palmer@google.com> and Ben Laurie]
1601 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1602 the right response is stapled. Also change SSL_get_certificate()
1603 so it returns the certificate actually sent.
1604 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1605 [Rob Stradling <rob.stradling@comodo.com>]
1607 *) Fix possible deadlock when decoding public keys.
1610 *) Don't use TLS 1.0 record version number in initial client hello
1614 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1616 *) Sanity check record length before skipping explicit IV in TLS
1617 1.2, 1.1 and DTLS to fix DoS attack.
1619 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1620 fuzzing as a service testing platform.
1624 *) Initialise tkeylen properly when encrypting CMS messages.
1625 Thanks to Solar Designer of Openwall for reporting this issue.
1628 *) In FIPS mode don't try to use composite ciphers as they are not
1632 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1634 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1635 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1636 mean any application compiled against OpenSSL 1.0.0 headers setting
1637 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1638 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1639 0x10000000L Any application which was previously compiled against
1640 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1641 will need to be recompiled as a result. Letting be results in
1642 inability to disable specifically TLS 1.1 and in client context,
1643 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1646 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1647 disable just protocol X, but all protocols above X *if* there are
1648 protocols *below* X still enabled. In more practical terms it means
1649 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1650 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1651 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1655 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1657 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1658 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1659 in CRYPTO_realloc_clean.
1661 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1662 issue and to Adam Langley <agl@chromium.org> for fixing it.
1664 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1666 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1669 *) Workarounds for some broken servers that "hang" if a client hello
1670 record length exceeds 255 bytes.
1672 1. Do not use record version number > TLS 1.0 in initial client
1673 hello: some (but not all) hanging servers will now work.
1674 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1675 the number of ciphers sent in the client hello. This should be
1676 set to an even number, such as 50, for example by passing:
1677 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1678 Most broken servers should now work.
1679 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1680 TLS 1.2 client support entirely.
1683 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1686 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1688 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1689 STRING form instead of a DigestInfo.
1692 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1693 and the RSA_sign/RSA_verify functions. This was made more apparent when
1694 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1695 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1696 the correct format in RSA_verify so both forms transparently work.
1699 *) Some servers which support TLS 1.0 can choke if we initially indicate
1700 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1701 encrypted premaster secret. As a workaround use the maximum pemitted
1702 client version in client hello, this should keep such servers happy
1703 and still work with previous versions of OpenSSL.
1706 *) Add support for TLS/DTLS heartbeats.
1707 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1709 *) Add support for SCTP.
1710 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1712 *) Improved PRNG seeding for VOS.
1713 [Paul Green <Paul.Green@stratus.com>]
1715 *) Extensive assembler packs updates, most notably:
1717 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1718 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1719 - x86_64: bit-sliced AES implementation;
1720 - ARM: NEON support, contemporary platforms optimizations;
1721 - s390x: z196 support;
1722 - *: GHASH and GF(2^m) multiplication implementations;
1726 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1727 (removal of unnecessary code)
1728 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1730 *) Add TLS key material exporter from RFC 5705.
1733 *) Add DTLS-SRTP negotiation from RFC 5764.
1736 *) Add Next Protocol Negotiation,
1737 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1738 disabled with a no-npn flag to config or Configure. Code donated
1740 [Adam Langley <agl@google.com> and Ben Laurie]
1742 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1743 NIST-P256, NIST-P521, with constant-time single point multiplication on
1744 typical inputs. Compiler support for the nonstandard type __uint128_t is
1745 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1746 Code made available under Apache License version 2.0.
1748 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1749 line to include this in your build of OpenSSL, and run "make depend" (or
1750 "make update"). This enables the following EC_METHODs:
1752 EC_GFp_nistp224_method()
1753 EC_GFp_nistp256_method()
1754 EC_GFp_nistp521_method()
1756 EC_GROUP_new_by_curve_name() will automatically use these (while
1757 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1759 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1761 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1762 all platforms. Move ssize_t definition from e_os.h to the public
1763 header file e_os2.h as it now appears in public header file cms.h
1766 *) New -sigopt option to the ca, req and x509 utilities. Additional
1767 signature parameters can be passed using this option and in
1771 *) Add RSA PSS signing function. This will generate and set the
1772 appropriate AlgorithmIdentifiers for PSS based on those in the
1773 corresponding EVP_MD_CTX structure. No application support yet.
1776 *) Support for companion algorithm specific ASN1 signing routines.
1777 New function ASN1_item_sign_ctx() signs a pre-initialised
1778 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1779 the appropriate parameters.
1782 *) Add new algorithm specific ASN1 verification initialisation function
1783 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1784 handling will be the same no matter what EVP_PKEY_METHOD is used.
1785 Add a PSS handler to support verification of PSS signatures: checked
1786 against a number of sample certificates.
1789 *) Add signature printing for PSS. Add PSS OIDs.
1790 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1792 *) Add algorithm specific signature printing. An individual ASN1 method
1793 can now print out signatures instead of the standard hex dump.
1795 More complex signatures (e.g. PSS) can print out more meaningful
1796 information. Include DSA version that prints out the signature
1800 *) Password based recipient info support for CMS library: implementing
1804 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1805 neatly separates the code into cipher and PBE sections and is required
1806 for some algorithms that split PBES2 into separate pieces (such as
1807 password based CMS).
1810 *) Session-handling fixes:
1811 - Fix handling of connections that are resuming with a session ID,
1812 but also support Session Tickets.
1813 - Fix a bug that suppressed issuing of a new ticket if the client
1814 presented a ticket with an expired session.
1815 - Try to set the ticket lifetime hint to something reasonable.
1816 - Make tickets shorter by excluding irrelevant information.
1817 - On the client side, don't ignore renewed tickets.
1818 [Adam Langley, Bodo Moeller (Google)]
1820 *) Fix PSK session representation.
1823 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1825 This work was sponsored by Intel.
1828 *) Add GCM support to TLS library. Some custom code is needed to split
1829 the IV between the fixed (from PRF) and explicit (from TLS record)
1830 portions. This adds all GCM ciphersuites supported by RFC5288 and
1831 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
1832 add a special AESGCM string for GCM only.
1835 *) Expand range of ctrls for AES GCM. Permit setting invocation
1836 field on decrypt and retrieval of invocation field only on encrypt.
1839 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1840 As required by RFC5289 these ciphersuites cannot be used if for
1841 versions of TLS earlier than 1.2.
1844 *) For FIPS capable OpenSSL interpret a NULL default public key method
1845 as unset and return the appopriate default but do *not* set the default.
1846 This means we can return the appopriate method in applications that
1847 swicth between FIPS and non-FIPS modes.
1850 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1851 ENGINE is used then we cannot handle that in the FIPS module so we
1852 keep original code iff non-FIPS operations are allowed.
1855 *) Add -attime option to openssl utilities.
1856 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1858 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1861 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1862 FIPS EC methods unconditionally for now.
1865 *) New build option no-ec2m to disable characteristic 2 code.
1868 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1869 all cases can be covered as some introduce binary incompatibilities.
1872 *) Redirect RSA operations to FIPS module including keygen,
1873 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1876 *) Add similar low level API blocking to ciphers.
1879 *) Low level digest APIs are not approved in FIPS mode: any attempt
1880 to use these will cause a fatal error. Applications that *really* want
1881 to use them can use the private_* version instead.
1884 *) Redirect cipher operations to FIPS module for FIPS builds.
1887 *) Redirect digest operations to FIPS module for FIPS builds.
1890 *) Update build system to add "fips" flag which will link in fipscanister.o
1891 for static and shared library builds embedding a signature if needed.
1894 *) Output TLS supported curves in preference order instead of numerical
1895 order. This is currently hardcoded for the highest order curves first.
1896 This should be configurable so applications can judge speed vs strength.
1899 *) Add TLS v1.2 server support for client authentication.
1902 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1906 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1907 FIPS modules versions.
1910 *) Add TLS v1.2 client side support for client authentication. Keep cache
1911 of handshake records longer as we don't know the hash algorithm to use
1912 until after the certificate request message is received.
1915 *) Initial TLS v1.2 client support. Add a default signature algorithms
1916 extension including all the algorithms we support. Parse new signature
1917 format in client key exchange. Relax some ECC signing restrictions for
1918 TLS v1.2 as indicated in RFC5246.
1921 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1922 to new signature format when needed using client digest preference.
1923 All server ciphersuites should now work correctly in TLS v1.2. No client
1924 support yet and no support for client certificates.
1927 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1928 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1929 ciphersuites. At present only RSA key exchange ciphersuites work with
1930 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1931 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1932 and version checking.
1935 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1936 with this defined it will not be affected by any changes to ssl internal
1937 structures. Add several utility functions to allow openssl application
1938 to work with OPENSSL_NO_SSL_INTERN defined.
1942 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1944 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1947 *) Permit abbreviated handshakes when renegotiating using the function
1948 SSL_renegotiate_abbreviated().
1949 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1951 *) Add call to ENGINE_register_all_complete() to
1952 ENGINE_load_builtin_engines(), so some implementations get used
1953 automatically instead of needing explicit application support.
1956 *) Add support for TLS key exporter as described in RFC5705.
1957 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1959 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1960 a few changes are required:
1962 Add SSL_OP_NO_TLSv1_1 flag.
1963 Add TLSv1_1 methods.
1964 Update version checking logic to handle version 1.1.
1965 Add explicit IV handling (ported from DTLS code).
1966 Add command line options to s_client/s_server.
1969 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1971 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1972 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1973 content decryption and always return the same error. Note: this attack
1974 needs on average 2^20 messages so it only affects automated senders. The
1975 old behaviour can be reenabled in the CMS code by setting the
1976 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1977 an MMA defence is not necessary.
1978 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1979 this issue. (CVE-2012-0884)
1982 *) Fix CVE-2011-4619: make sure we really are receiving a
1983 client hello before rejecting multiple SGC restarts. Thanks to
1984 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1987 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1989 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1990 Thanks to Antonio Martin, Enterprise Secure Access Research and
1991 Development, Cisco Systems, Inc. for discovering this bug and
1992 preparing a fix. (CVE-2012-0050)
1995 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1997 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1998 of the Vaudenay padding oracle attack on CBC mode encryption
1999 which enables an efficient plaintext recovery attack against
2000 the OpenSSL implementation of DTLS. Their attack exploits timing
2001 differences arising during decryption processing. A research
2002 paper describing this attack can be found at:
2003 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2004 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2005 Security Group at Royal Holloway, University of London
2006 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2007 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2008 for preparing the fix. (CVE-2011-4108)
2009 [Robin Seggelmann, Michael Tuexen]
2011 *) Clear bytes used for block padding of SSL 3.0 records.
2013 [Adam Langley (Google)]
2015 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2016 Kadianakis <desnacked@gmail.com> for discovering this issue and
2017 Adam Langley for preparing the fix. (CVE-2011-4619)
2018 [Adam Langley (Google)]
2020 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2021 [Andrey Kulikov <amdeich@gmail.com>]
2023 *) Prevent malformed RFC3779 data triggering an assertion failure.
2024 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2025 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2026 [Rob Austein <sra@hactrn.net>]
2028 *) Improved PRNG seeding for VOS.
2029 [Paul Green <Paul.Green@stratus.com>]
2031 *) Fix ssl_ciph.c set-up race.
2032 [Adam Langley (Google)]
2034 *) Fix spurious failures in ecdsatest.c.
2035 [Emilia Käsper (Google)]
2037 *) Fix the BIO_f_buffer() implementation (which was mixing different
2038 interpretations of the '..._len' fields).
2039 [Adam Langley (Google)]
2041 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2042 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2043 threads won't reuse the same blinding coefficients.
2045 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2046 lock to call BN_BLINDING_invert_ex, and avoids one use of
2047 BN_BLINDING_update for each BN_BLINDING structure (previously,
2048 the last update always remained unused).
2049 [Emilia Käsper (Google)]
2051 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2052 [Bob Buckholz (Google)]
2054 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2056 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2057 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2058 [Kaspar Brand <ossl@velox.ch>]
2060 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2061 for multi-threaded use of ECDH. (CVE-2011-3210)
2062 [Adam Langley (Google)]
2064 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2067 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2068 signature public key algorithm by using OID xref utilities instead.
2069 Before this you could only use some ECC ciphersuites with SHA1 only.
2072 *) Add protection against ECDSA timing attacks as mentioned in the paper
2073 by Billy Bob Brumley and Nicola Tuveri, see:
2075 http://eprint.iacr.org/2011/232.pdf
2077 [Billy Bob Brumley and Nicola Tuveri]
2079 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2081 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2082 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2084 *) Fix bug in string printing code: if *any* escaping is enabled we must
2085 escape the escape character (backslash) or the resulting string is
2089 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2091 *) Disable code workaround for ancient and obsolete Netscape browsers
2092 and servers: an attacker can use it in a ciphersuite downgrade attack.
2093 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2096 *) Fixed J-PAKE implementation error, originally discovered by
2097 Sebastien Martini, further info and confirmation from Stefan
2098 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2101 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2103 *) Fix extension code to avoid race conditions which can result in a buffer
2104 overrun vulnerability: resumed sessions must not be modified as they can
2105 be shared by multiple threads. CVE-2010-3864
2108 *) Fix WIN32 build system to correctly link an ENGINE directory into
2112 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2114 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2116 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2118 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2120 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2121 context. The operation can be customised via the ctrl mechanism in
2122 case ENGINEs want to include additional functionality.
2125 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2128 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2129 output hashes compatible with older versions of OpenSSL.
2130 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2132 *) Fix compression algorithm handling: if resuming a session use the
2133 compression algorithm of the resumed session instead of determining
2134 it from client hello again. Don't allow server to change algorithm.
2137 *) Add load_crls() function to apps tidying load_certs() too. Add option
2138 to verify utility to allow additional CRLs to be included.
2141 *) Update OCSP request code to permit adding custom headers to the request:
2142 some responders need this.
2145 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2147 [Julia Lawall <julia@diku.dk>]
2149 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2150 needlessly dereferenced structures, used obsolete functions and
2151 didn't handle all updated verify codes correctly.
2154 *) Disable MD2 in the default configuration.
2157 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2158 indicate the initial BIO being pushed or popped. This makes it possible
2159 to determine whether the BIO is the one explicitly called or as a result
2160 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2161 it handles reference counts correctly and doesn't zero out the I/O bio
2162 when it is not being explicitly popped. WARNING: applications which
2163 included workarounds for the old buggy behaviour will need to be modified
2164 or they could free up already freed BIOs.
2167 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2168 renaming to all platforms (within the 0.9.8 branch, this was
2169 done conditionally on Netware platforms to avoid a name clash).
2170 [Guenter <lists@gknw.net>]
2172 *) Add ECDHE and PSK support to DTLS.
2173 [Michael Tuexen <tuexen@fh-muenster.de>]
2175 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2179 *) Add "missing" function EVP_MD_flags() (without this the only way to
2180 retrieve a digest flags is by accessing the structure directly. Update
2181 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2182 or cipher is registered as in the "from" argument. Print out all
2183 registered digests in the dgst usage message instead of manually
2184 attempting to work them out.
2187 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2188 this allows the use of compression and extensions. Change default cipher
2189 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2190 by default unless an application cipher string requests it.
2193 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2194 key ids to find matching certificates and keys but some PKCS#12 files
2195 don't follow the (somewhat unwritten) rules and this strategy fails.
2196 Now just gather all certificates together and the first private key
2197 then look for the first certificate that matches the key.
2200 *) Support use of registered digest and cipher names for dgst and cipher
2201 commands instead of having to add each one as a special case. So now
2208 openssl dgst -sha256 foo
2210 and this works for ENGINE based algorithms too.
2214 *) Update Gost ENGINE to support parameter files.
2215 [Victor B. Wagner <vitus@cryptocom.ru>]
2217 *) Support GeneralizedTime in ca utility.
2218 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2220 *) Enhance the hash format used for certificate directory links. The new
2221 form uses the canonical encoding (meaning equivalent names will work
2222 even if they aren't identical) and uses SHA1 instead of MD5. This form
2223 is incompatible with the older format and as a result c_rehash should
2224 be used to rebuild symbolic links.
2227 *) Make PKCS#8 the default write format for private keys, replacing the
2228 traditional format. This form is standardised, more secure and doesn't
2229 include an implicit MD5 dependency.
2232 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2233 committed to OpenSSL should pass this lot as a minimum.
2236 *) Add session ticket override functionality for use by EAP-FAST.
2237 [Jouni Malinen <j@w1.fi>]
2239 *) Modify HMAC functions to return a value. Since these can be implemented
2240 in an ENGINE errors can occur.
2243 *) Type-checked OBJ_bsearch_ex.
2246 *) Type-checked OBJ_bsearch. Also some constification necessitated
2247 by type-checking. Still to come: TXT_DB, bsearch(?),
2248 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2252 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2253 seconds to a tm structure directly, instead of going through OS
2254 specific date routines. This avoids any issues with OS routines such
2255 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2256 and X509_time_adj_ex() to cover the extended range. The existing
2257 X509_time_adj() is still usable and will no longer have any date issues.
2260 *) Delta CRL support. New use deltas option which will attempt to locate
2261 and search any appropriate delta CRLs available.
2263 This work was sponsored by Google.
2266 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2267 code and add additional score elements. Validate alternate CRL paths
2268 as part of the CRL checking and indicate a new error "CRL path validation
2269 error" in this case. Applications wanting additional details can use
2270 the verify callback and check the new "parent" field. If this is not
2271 NULL CRL path validation is taking place. Existing applications wont
2272 see this because it requires extended CRL support which is off by
2275 This work was sponsored by Google.
2278 *) Support for freshest CRL extension.
2280 This work was sponsored by Google.
2283 *) Initial indirect CRL support. Currently only supported in the CRLs
2284 passed directly and not via lookup. Process certificate issuer
2285 CRL entry extension and lookup CRL entries by bother issuer name
2286 and serial number. Check and process CRL issuer entry in IDP extension.
2288 This work was sponsored by Google.
2291 *) Add support for distinct certificate and CRL paths. The CRL issuer
2292 certificate is validated separately in this case. Only enabled if
2293 an extended CRL support flag is set: this flag will enable additional
2294 CRL functionality in future.
2296 This work was sponsored by Google.
2299 *) Add support for policy mappings extension.
2301 This work was sponsored by Google.
2304 *) Fixes to pathlength constraint, self issued certificate handling,
2305 policy processing to align with RFC3280 and PKITS tests.
2307 This work was sponsored by Google.
2310 *) Support for name constraints certificate extension. DN, email, DNS
2311 and URI types are currently supported.
2313 This work was sponsored by Google.
2316 *) To cater for systems that provide a pointer-based thread ID rather
2317 than numeric, deprecate the current numeric thread ID mechanism and
2318 replace it with a structure and associated callback type. This
2319 mechanism allows a numeric "hash" to be extracted from a thread ID in
2320 either case, and on platforms where pointers are larger than 'long',
2321 mixing is done to help ensure the numeric 'hash' is usable even if it
2322 can't be guaranteed unique. The default mechanism is to use "&errno"
2323 as a pointer-based thread ID to distinguish between threads.
2325 Applications that want to provide their own thread IDs should now use
2326 CRYPTO_THREADID_set_callback() to register a callback that will call
2327 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2329 Note that ERR_remove_state() is now deprecated, because it is tied
2330 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2331 to free the current thread's error state should be replaced by
2332 ERR_remove_thread_state(NULL).
2334 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2335 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2336 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2337 application was previously providing a numeric thread callback that
2338 was inappropriate for distinguishing threads, then uniqueness might
2339 have been obtained with &errno that happened immediately in the
2340 intermediate development versions of OpenSSL; this is no longer the
2341 case, the numeric thread callback will now override the automatic use
2343 [Geoff Thorpe, with help from Bodo Moeller]
2345 *) Initial support for different CRL issuing certificates. This covers a
2346 simple case where the self issued certificates in the chain exist and
2347 the real CRL issuer is higher in the existing chain.
2349 This work was sponsored by Google.
2352 *) Removed effectively defunct crypto/store from the build.
2355 *) Revamp of STACK to provide stronger type-checking. Still to come:
2356 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2357 ASN1_STRING, CONF_VALUE.
2360 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2361 RAM on SSL connections. This option can save about 34k per idle SSL.
2364 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2365 STACK, TXT_DB, bsearch, qsort.
2368 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2369 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2370 support for data, signedData, compressedData, digestedData and
2371 encryptedData, envelopedData types included. Scripts to check against
2372 RFC4134 examples draft and interop and consistency checks of many
2373 content types and variants.
2376 *) Add options to enc utility to support use of zlib compression BIO.
2379 *) Extend mk1mf to support importing of options and assembly language
2380 files from Configure script, currently only included in VC-WIN32.
2381 The assembly language rules can now optionally generate the source
2382 files from the associated perl scripts.
2385 *) Implement remaining functionality needed to support GOST ciphersuites.
2386 Interop testing has been performed using CryptoPro implementations.
2387 [Victor B. Wagner <vitus@cryptocom.ru>]
2389 *) s390x assembler pack.
2392 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2396 *) Implement Opaque PRF Input TLS extension as specified in
2397 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2398 official specification yet and no extension type assignment by
2399 IANA exists, this extension (for now) will have to be explicitly
2400 enabled when building OpenSSL by providing the extension number
2401 to use. For example, specify an option
2403 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2405 to the "config" or "Configure" script to enable the extension,
2406 assuming extension number 0x9527 (which is a completely arbitrary
2407 and unofficial assignment based on the MD5 hash of the Internet
2408 Draft). Note that by doing so, you potentially lose
2409 interoperability with other TLS implementations since these might
2410 be using the same extension number for other purposes.
2412 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2413 opaque PRF input value to use in the handshake. This will create
2414 an interal copy of the length-'len' string at 'src', and will
2415 return non-zero for success.
2417 To get more control and flexibility, provide a callback function
2420 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2421 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2425 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2428 Callback function 'cb' will be called in handshakes, and is
2429 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2430 Argument 'arg' is for application purposes (the value as given to
2431 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2432 be provided to the callback function). The callback function
2433 has to return non-zero to report success: usually 1 to use opaque
2434 PRF input just if possible, or 2 to enforce use of the opaque PRF
2435 input. In the latter case, the library will abort the handshake
2436 if opaque PRF input is not successfully negotiated.
2438 Arguments 'peerinput' and 'len' given to the callback function
2439 will always be NULL and 0 in the case of a client. A server will
2440 see the client's opaque PRF input through these variables if
2441 available (NULL and 0 otherwise). Note that if the server
2442 provides an opaque PRF input, the length must be the same as the
2443 length of the client's opaque PRF input.
2445 Note that the callback function will only be called when creating
2446 a new session (session resumption can resume whatever was
2447 previously negotiated), and will not be called in SSL 2.0
2448 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2449 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2450 for applications that need to enforce opaque PRF input.
2454 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2457 [Victor B. Wagner <vitus@cryptocom.ru>]
2459 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2460 RFC4507bis. The encrypted ticket format is an encrypted encoded
2461 SSL_SESSION structure, that way new session features are automatically
2464 If a client application caches session in an SSL_SESSION structure
2465 support is transparent because tickets are now stored in the encoded
2468 The SSL_CTX structure automatically generates keys for ticket
2469 protection in servers so again support should be possible
2470 with no application modification.
2472 If a client or server wishes to disable RFC4507 support then the option
2473 SSL_OP_NO_TICKET can be set.
2475 Add a TLS extension debugging callback to allow the contents of any client
2476 or server extensions to be examined.
2478 This work was sponsored by Google.
2481 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2482 OpenSSL should now compile cleanly on gcc 4.2
2483 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2485 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2486 support including streaming MAC support: this is required for GOST
2487 ciphersuite support.
2488 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2490 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2491 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2492 to output in BER and PEM format.
2495 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2496 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2497 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2498 ENGINE support for HMAC keys which are unextractable. New -mac and
2499 -macopt options to dgst utility.
2502 *) New option -sigopt to dgst utility. Update dgst to use
2503 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2504 alternative signing paramaters such as X9.31 or PSS in the dgst
2508 *) Change ssl_cipher_apply_rule(), the internal function that does
2509 the work each time a ciphersuite string requests enabling
2510 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2511 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2512 the order of disabled ciphersuites such that those ciphersuites
2513 that most recently went from enabled to disabled not only stay
2514 in order with respect to each other, but also have higher priority
2515 than other disabled ciphersuites the next time ciphersuites are
2518 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2519 the same ciphersuites as with "HIGH" alone, but in a specific
2520 order where the PSK ciphersuites come first (since they are the
2521 most recently disabled ciphersuites when "HIGH" is parsed).
2523 Also, change ssl_create_cipher_list() (using this new
2524 funcionality) such that between otherwise identical
2525 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2529 *) Change ssl_create_cipher_list() so that it automatically
2530 arranges the ciphersuites in reasonable order before starting
2531 to process the rule string. Thus, the definition for "DEFAULT"
2532 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2533 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2534 This makes it much easier to arrive at a reasonable default order
2535 in applications for which anonymous ciphers are OK (meaning
2536 that you can't actually use DEFAULT).
2537 [Bodo Moeller; suggested by Victor Duchovni]
2539 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2540 processing) into multiple integers instead of setting
2541 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2542 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2543 (These masks as well as the individual bit definitions are hidden
2544 away into the non-exported interface ssl/ssl_locl.h, so this
2545 change to the definition of the SSL_CIPHER structure shouldn't
2546 affect applications.) This give us more bits for each of these
2547 categories, so there is no longer a need to coagulate AES128 and
2548 AES256 into a single algorithm bit, and to coagulate Camellia128
2549 and Camellia256 into a single algorithm bit, which has led to all
2552 Thus, among other things, the kludge introduced in 0.9.7m and
2553 0.9.8e for masking out AES256 independently of AES128 or masking
2554 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2556 With the change, we also introduce new ciphersuite aliases that
2557 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2561 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2562 Use the leftmost N bytes of the signature input if the input is
2563 larger than the prime q (with N being the size in bytes of q).
2566 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2567 it yet and it is largely untested.
2570 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2573 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2574 some compilers (gcc 4.2 and later) reject their use. Safestack is
2575 reimplemented. Update ASN1 to avoid use of legacy functions.
2578 *) Win32/64 targets are linked with Winsock2.
2581 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2582 to external functions. This can be used to increase CRL handling
2583 efficiency especially when CRLs are very large by (for example) storing
2584 the CRL revoked certificates in a database.
2587 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2588 new CRLs added to a directory can be used. New command line option
2589 -verify_return_error to s_client and s_server. This causes real errors
2590 to be returned by the verify callback instead of carrying on no matter
2591 what. This reflects the way a "real world" verify callback would behave.
2594 *) GOST engine, supporting several GOST algorithms and public key formats.
2595 Kindly donated by Cryptocom.
2598 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2599 partitioned by DP are handled but no indirect CRL or reason partitioning
2600 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2601 selected via a scoring technique which handles IDP and AKID in CRLs.
2604 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2605 will ultimately be used for all verify operations: this will remove the
2606 X509_STORE dependency on certificate verification and allow alternative
2607 lookup methods. X509_STORE based implementations of these two callbacks.
2610 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2611 Modify get_crl() to find a valid (unexpired) CRL if possible.
2614 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2615 this would be called X509_CRL_cmp() but that name is already used by
2616 a function that just compares CRL issuer names. Cache several CRL
2617 extensions in X509_CRL structure and cache CRLDP in X509.
2620 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2621 this maps equivalent X509_NAME structures into a consistent structure.
2622 Name comparison can then be performed rapidly using memcmp().
2625 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2629 *) Allow digests to supply their own micalg string for S/MIME type using
2630 the ctrl EVP_MD_CTRL_MICALG.
2633 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2634 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2635 ctrl. It can then customise the structure before and/or after signing
2639 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2640 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2641 to free up any added signature OIDs.
2644 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2645 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2646 digest and cipher tables. New options added to openssl utility:
2647 list-message-digest-algorithms and list-cipher-algorithms.
2650 *) Change the array representation of binary polynomials: the list
2651 of degrees of non-zero coefficients is now terminated with -1.
2652 Previously it was terminated with 0, which was also part of the
2653 value; thus, the array representation was not applicable to
2654 polynomials where t^0 has coefficient zero. This change makes
2655 the array representation useful in a more general context.
2658 *) Various modifications and fixes to SSL/TLS cipher string
2659 handling. For ECC, the code now distinguishes between fixed ECDH
2660 with RSA certificates on the one hand and with ECDSA certificates
2661 on the other hand, since these are separate ciphersuites. The
2662 unused code for Fortezza ciphersuites has been removed.
2664 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2665 (not "ECDHE"). For consistency with the code for DH
2666 certificates, use of ECDH certificates is now considered ECDH
2667 authentication, not RSA or ECDSA authentication (the latter is
2668 merely the CA's signing algorithm and not actively used in the
2671 The temporary ciphersuite alias "ECCdraft" is no longer
2672 available, and ECC ciphersuites are no longer excluded from "ALL"
2673 and "DEFAULT". The following aliases now exist for RFC 4492
2674 ciphersuites, most of these by analogy with the DH case:
2676 kECDHr - ECDH cert, signed with RSA
2677 kECDHe - ECDH cert, signed with ECDSA
2678 kECDH - ECDH cert (signed with either RSA or ECDSA)
2679 kEECDH - ephemeral ECDH
2680 ECDH - ECDH cert or ephemeral ECDH
2686 AECDH - anonymous ECDH
2687 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2691 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2692 Use correct micalg parameters depending on digest(s) in signed message.
2695 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2696 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2699 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2700 an engine to register a method. Add ENGINE lookups for methods and
2701 functional reference processing.
2704 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2705 EVP_{Sign,Verify}* which allow an application to customise the signature
2709 *) New -resign option to smime utility. This adds one or more signers
2710 to an existing PKCS#7 signedData structure. Also -md option to use an
2711 alternative message digest algorithm for signing.
2714 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2715 create PKCS7 structures containing multiple signers. Update smime
2716 application to support multiple signers.
2719 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2723 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2724 Reorganize PBE internals to lookup from a static table using NIDs,
2725 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2726 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2727 PRF which will be automatically used with PBES2.
2730 *) Replace the algorithm specific calls to generate keys in "req" with the
2734 *) Update PKCS#7 enveloped data routines to use new API. This is now
2735 supported by any public key method supporting the encrypt operation. A
2736 ctrl is added to allow the public key algorithm to examine or modify
2737 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2741 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2742 a default digest type to use. In most cases this will be SHA1 but some
2743 algorithms (such as GOST) need to specify an alternative digest. The
2744 return value indicates how strong the prefernce is 1 means optional and
2745 2 is mandatory (that is it is the only supported type). Modify
2746 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2747 use the default md. Update openssl utilities to use the default digest
2748 type for signing if it is not explicitly indicated.
2751 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2752 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2753 signing method from the key type. This effectively removes the link
2754 between digests and public key types.
2757 *) Add an OID cross reference table and utility functions. Its purpose is to
2758 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2759 rsaEncryption. This will allow some of the algorithm specific hackery
2760 needed to use the correct OID to be removed.
2763 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2764 structures for PKCS7_sign(). They are now set up by the relevant public
2768 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2771 *) Add support for key derivation (agreement) in the API, DH method and
2775 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2776 public and private key formats. As a side effect these add additional
2777 command line functionality not previously available: DSA signatures can be
2778 generated and verified using pkeyutl and DH key support and generation in
2783 [Oliver Tappe <zooey@hirschkaefer.de>]
2785 *) New make target "install_html_docs" installs HTML renditions of the
2787 [Oliver Tappe <zooey@hirschkaefer.de>]
2789 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
2790 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2791 support key and parameter generation and add initial key generation
2792 functionality for RSA.
2795 *) Add functions for main EVP_PKEY_method operations. The undocumented
2796 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2797 EVP_PKEY_{encrypt,decrypt}_old.
2800 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2801 key API, doesn't do much yet.
2804 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2805 public key algorithms. New option to openssl utility:
2806 "list-public-key-algorithms" to print out info.
2809 *) Implement the Supported Elliptic Curves Extension for
2810 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2813 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2814 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2817 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2818 utilities such as rsa, dsa, dsaparam etc except they process any key
2822 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2823 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2824 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2828 *) Initial support for pluggable public key ASN1.
2829 De-spaghettify the public key ASN1 handling. Move public and private
2830 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2831 algorithm specific handling to a single module within the relevant
2832 algorithm directory. Add functions to allow (near) opaque processing
2833 of public and private key structures.
2836 *) Implement the Supported Point Formats Extension for
2837 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2840 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2841 for the psk identity [hint] and the psk callback functions to the
2842 SSL_SESSION, SSL and SSL_CTX structure.
2845 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2849 SSL_CTX_use_psk_identity_hint
2850 SSL_get_psk_identity_hint
2851 SSL_get_psk_identity
2852 SSL_use_psk_identity_hint
2854 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2856 *) Add RFC 3161 compliant time stamp request creation, response generation
2857 and response verification functionality.
2858 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2860 *) Add initial support for TLS extensions, specifically for the server_name
2861 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2862 have new members for a host name. The SSL data structure has an
2863 additional member SSL_CTX *initial_ctx so that new sessions can be
2864 stored in that context to allow for session resumption, even after the
2865 SSL has been switched to a new SSL_CTX in reaction to a client's
2866 server_name extension.
2868 New functions (subject to change):
2870 SSL_get_servername()
2871 SSL_get_servername_type()
2874 New CTRL codes and macros (subject to change):
2876 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2877 - SSL_CTX_set_tlsext_servername_callback()
2878 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2879 - SSL_CTX_set_tlsext_servername_arg()
2880 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2882 openssl s_client has a new '-servername ...' option.
2884 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2885 '-key2 ...', '-servername_fatal' (subject to change). This allows
2886 testing the HostName extension for a specific single host name ('-cert'
2887 and '-key' remain fallbacks for handshakes without HostName
2888 negotiation). If the unrecogninzed_name alert has to be sent, this by
2889 default is a warning; it becomes fatal with the '-servername_fatal'
2892 [Peter Sylvester, Remy Allais, Christophe Renou]
2894 *) Whirlpool hash implementation is added.
2897 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2898 bn(64,32). Because of instruction set limitations it doesn't have
2899 any negative impact on performance. This was done mostly in order
2900 to make it possible to share assembler modules, such as bn_mul_mont
2901 implementations, between 32- and 64-bit builds without hassle.
2904 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2905 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2909 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2910 dedicated Montgomery multiplication procedure, is introduced.
2911 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2912 "64-bit" performance on certain 32-bit targets.
2915 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2916 in SSL structures. New SSL ctrl to set maximum send fragment size.
2917 Save memory by seeting the I/O buffer sizes dynamically instead of
2918 using the maximum available value.
2921 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2922 in addition to the text details.
2925 *) Very, very preliminary EXPERIMENTAL support for printing of general
2926 ASN1 structures. This currently produces rather ugly output and doesn't
2927 handle several customised structures at all.
2930 *) Integrated support for PVK file format and some related formats such
2931 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2932 these in the 'rsa' and 'dsa' utilities.
2935 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2938 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2939 place for the (very old) "NETSCAPE" format certificates which are now
2940 handled using new ASN1 code equivalents.
2943 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2944 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2945 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2948 *) Modify CRL distribution points extension code to print out previously
2949 unsupported fields. Enhance extension setting code to allow setting of
2953 *) Add print and set support for Issuing Distribution Point CRL extension.
2956 *) Change 'Configure' script to enable Camellia by default.
2959 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2961 *) When rejecting SSL/TLS records due to an incorrect version number, never
2962 update s->server with a new major version number. As of
2963 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2964 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2965 the previous behavior could result in a read attempt at NULL when
2966 receiving specific incorrect SSL/TLS records once record payload
2967 protection is active. (CVE-2010-0740)
2968 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2970 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2971 could be crashed if the relevant tables were not present (e.g. chrooted).
2972 [Tomas Hoger <thoger@redhat.com>]
2974 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2976 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2977 [Martin Olsson, Neel Mehta]
2979 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2980 accommodate for stack sorting, always a write lock!).
2983 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2984 excessive delays in the RAND_poll(): over a minute. As a workaround
2985 include a time check in the inner Heap32Next loop too.
2988 *) The code that handled flushing of data in SSL/TLS originally used the
2989 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2990 the problem outlined in PR#1949. The fix suggested there however can
2991 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2992 of Apache). So instead simplify the code to flush unconditionally.
2993 This should be fine since flushing with no data to flush is a no op.
2996 *) Handle TLS versions 2.0 and later properly and correctly use the
2997 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2998 off ancient servers have a habit of sticking around for a while...
3001 *) Modify compression code so it frees up structures without using the
3002 ex_data callbacks. This works around a problem where some applications
3003 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3004 restarting) then use compression (e.g. SSL with compression) later.
3005 This results in significant per-connection memory leaks and
3006 has caused some security issues including CVE-2008-1678 and
3010 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3011 change when encrypting or decrypting.
3014 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3015 connect and renegotiate with servers which do not support RI.
3016 Until RI is more widely deployed this option is enabled by default.
3019 *) Add "missing" ssl ctrls to clear options and mode.
3022 *) If client attempts to renegotiate and doesn't support RI respond with
3023 a no_renegotiation alert as required by RFC5746. Some renegotiating
3024 TLS clients will continue a connection gracefully when they receive
3025 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3026 waiting for a server hello which it will never receive. Now we treat a
3027 received no_renegotiation alert as a fatal error. This is because
3028 applications requesting a renegotiation might well expect it to succeed
3029 and would have no code in place to handle the server denying it so the
3030 only safe thing to do is to terminate the connection.
3033 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3034 peer supports secure renegotiation and 0 otherwise. Print out peer
3035 renegotiation support in s_client/s_server.
3038 *) Replace the highly broken and deprecated SPKAC certification method with
3039 the updated NID creation version. This should correctly handle UTF8.
3042 *) Implement RFC5746. Re-enable renegotiation but require the extension
3043 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3044 turns out to be a bad idea. It has been replaced by
3045 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3046 SSL_CTX_set_options(). This is really not recommended unless you
3047 know what you are doing.
3048 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3050 *) Fixes to stateless session resumption handling. Use initial_ctx when
3051 issuing and attempting to decrypt tickets in case it has changed during
3052 servername handling. Use a non-zero length session ID when attempting
3053 stateless session resumption: this makes it possible to determine if
3054 a resumption has occurred immediately after receiving server hello
3055 (several places in OpenSSL subtly assume this) instead of later in
3059 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3060 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3061 fixes for a few places where the return code is not checked
3063 [Julia Lawall <julia@diku.dk>]
3065 *) Add --strict-warnings option to Configure script to include devteam
3066 warnings in other configurations.
3069 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3070 makes it possible to install openssl libraries in locations which
3071 have names other than "lib", for example "/usr/lib64" which some
3073 [Steve Henson, based on patch from Jeremy Utley]
3075 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3076 X690 8.9.12 and can produce some misleading textual output of OIDs.
3077 [Steve Henson, reported by Dan Kaminsky]
3079 *) Delete MD2 from algorithm tables. This follows the recommendation in
3080 several standards that it is not used in new applications due to
3081 several cryptographic weaknesses. For binary compatibility reasons
3082 the MD2 API is still compiled in by default.
3085 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3089 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3090 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3092 [Guenter <lists@gknw.net>]
3094 *) Fix the server certificate chain building code to use X509_verify_cert(),
3095 it used to have an ad-hoc builder which was unable to cope with anything
3096 other than a simple chain.
3097 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3099 *) Don't check self signed certificate signatures in X509_verify_cert()
3100 by default (a flag can override this): it just wastes time without
3101 adding any security. As a useful side effect self signed root CAs
3102 with non-FIPS digests are now usable in FIPS mode.
3105 *) In dtls1_process_out_of_seq_message() the check if the current message
3106 is already buffered was missing. For every new message was memory
3107 allocated, allowing an attacker to perform an denial of service attack
3108 with sending out of seq handshake messages until there is no memory
3109 left. Additionally every future messege was buffered, even if the
3110 sequence number made no sense and would be part of another handshake.
3111 So only messages with sequence numbers less than 10 in advance will be
3112 buffered. (CVE-2009-1378)
3113 [Robin Seggelmann, discovered by Daniel Mentz]
3115 *) Records are buffered if they arrive with a future epoch to be
3116 processed after finishing the corresponding handshake. There is
3117 currently no limitation to this buffer allowing an attacker to perform
3118 a DOS attack with sending records with future epochs until there is no
3119 memory left. This patch adds the pqueue_size() function to detemine
3120 the size of a buffer and limits the record buffer to 100 entries.
3122 [Robin Seggelmann, discovered by Daniel Mentz]
3124 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3125 parent structure is freed. (CVE-2009-1379)
3128 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3129 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3132 [Ilya O. <vrghost@gmail.com>]
3134 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3136 *) Disable renegotiation completely - this fixes a severe security
3137 problem (CVE-2009-3555) at the cost of breaking all
3138 renegotiation. Renegotiation can be re-enabled by setting
3139 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3140 run-time. This is really not recommended unless you know what
3144 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3146 *) Don't set val to NULL when freeing up structures, it is freed up by
3147 underlying code. If sizeof(void *) > sizeof(long) this can result in
3148 zeroing past the valid field. (CVE-2009-0789)
3149 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3151 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3152 checked correctly. This would allow some invalid signed attributes to
3153 appear to verify correctly. (CVE-2009-0591)
3154 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3156 *) Reject UniversalString and BMPString types with invalid lengths. This
3157 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3158 a legal length. (CVE-2009-0590)
3161 *) Set S/MIME signing as the default purpose rather than setting it
3162 unconditionally. This allows applications to override it at the store
3166 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3167 to handle some structures.
3170 *) Improve efficiency of mem_gets: don't search whole buffer each time
3172 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3174 *) New -hex option for openssl rand.
3177 *) Print out UTF8String and NumericString when parsing ASN1.
3180 *) Support NumericString type for name components.
3183 *) Allow CC in the environment to override the automatically chosen
3184 compiler. Note that nothing is done to ensure flags work with the
3188 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3190 *) Properly check EVP_VerifyFinal() and similar return values
3192 [Ben Laurie, Bodo Moeller, Google Security Team]
3194 *) Enable TLS extensions by default.
3197 *) Allow the CHIL engine to be loaded, whether the application is
3198 multithreaded or not. (This does not release the developer from the
3199 obligation to set up the dynamic locking callbacks.)
3200 [Sander Temme <sander@temme.net>]
3202 *) Use correct exit code if there is an error in dgst command.
3203 [Steve Henson; problem pointed out by Roland Dirlewanger]
3205 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3206 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3209 *) Add experimental JPAKE support, including demo authentication in
3210 s_client and s_server.
3213 *) Set the comparison function in v3_addr_canonize().
3214 [Rob Austein <sra@hactrn.net>]
3216 *) Add support for XMPP STARTTLS in s_client.
3217 [Philip Paeps <philip@freebsd.org>]
3219 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3220 to ensure that even with this option, only ciphersuites in the
3221 server's preference list will be accepted. (Note that the option
3222 applies only when resuming a session, so the earlier behavior was
3223 just about the algorithm choice for symmetric cryptography.)
3226 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3228 *) Fix NULL pointer dereference if a DTLS server received
3229 ChangeCipherSpec as first record (CVE-2009-1386).
3232 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
3233 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3236 *) The fix in 0.9.8c that supposedly got rid of unsafe
3237 double-checked locking was incomplete for RSA blinding,
3238 addressing just one layer of what turns out to have been
3239 doubly unsafe triple-checked locking.
3241 So now fix this for real by retiring the MONT_HELPER macro
3242 in crypto/rsa/rsa_eay.c.
3244 [Bodo Moeller; problem pointed out by Marius Schilder]
3246 *) Various precautionary measures:
3248 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3250 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3251 (NB: This would require knowledge of the secret session ticket key
3252 to exploit, in which case you'd be SOL either way.)
3254 - Change bn_nist.c so that it will properly handle input BIGNUMs
3255 outside the expected range.
3257 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3260 [Neel Mehta, Bodo Moeller]
3262 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3263 the load fails. Useful for distros.
3264 [Ben Laurie and the FreeBSD team]
3266 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3269 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3272 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3274 This work was sponsored by Logica.
3277 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3278 keystores. Support for SSL/TLS client authentication too.
3279 Not compiled unless enable-capieng specified to Configure.
3281 This work was sponsored by Logica.
3284 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3285 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3286 attribute creation routines such as certifcate requests and PKCS#12
3290 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3292 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3293 handshake which could lead to a cilent crash as found using the
3294 Codenomicon TLS test suite (CVE-2008-1672)
3295 [Steve Henson, Mark Cox]
3297 *) Fix double free in TLS server name extensions which could lead to
3298 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3301 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3303 Clear the error queue to ensure that error entries left from
3304 older function calls do not interfere with the correct operation.
3305 [Lutz Jaenicke, Erik de Castro Lopo]
3307 *) Remove root CA certificates of commercial CAs:
3309 The OpenSSL project does not recommend any specific CA and does not
3310 have any policy with respect to including or excluding any CA.
3311 Therefore it does not make any sense to ship an arbitrary selection
3312 of root CA certificates with the OpenSSL software.
3315 *) RSA OAEP patches to fix two separate invalid memory reads.
3316 The first one involves inputs when 'lzero' is greater than
3317 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3318 before the beginning of from). The second one involves inputs where
3319 the 'db' section contains nothing but zeroes (there is a one-byte
3320 invalid read after the end of 'db').
3321 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3323 *) Partial backport from 0.9.9-dev:
3325 Introduce bn_mul_mont (dedicated Montgomery multiplication
3326 procedure) as a candidate for BIGNUM assembler implementation.
3327 While 0.9.9-dev uses assembler for various architectures, only
3328 x86_64 is available by default here in the 0.9.8 branch, and
3329 32-bit x86 is available through a compile-time setting.
3331 To try the 32-bit x86 assembler implementation, use Configure
3332 option "enable-montasm" (which exists only for this backport).
3334 As "enable-montasm" for 32-bit x86 disclaims code stability
3335 anyway, in this constellation we activate additional code
3336 backported from 0.9.9-dev for further performance improvements,
3337 namely BN_from_montgomery_word. (To enable this otherwise,
3338 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3340 [Andy Polyakov (backport partially by Bodo Moeller)]
3342 *) Add TLS session ticket callback. This allows an application to set
3343 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3344 values. This is useful for key rollover for example where several key
3345 sets may exist with different names.
3348 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3349 This was broken until now in 0.9.8 releases, such that the only way
3350 a registered ENGINE could be used (assuming it initialises
3351 successfully on the host) was to explicitly set it as the default
3352 for the relevant algorithms. This is in contradiction with 0.9.7
3353 behaviour and the documentation. With this fix, when an ENGINE is
3354 registered into a given algorithm's table of implementations, the
3355 'uptodate' flag is reset so that auto-discovery will be used next
3356 time a new context for that algorithm attempts to select an
3358 [Ian Lister (tweaked by Geoff Thorpe)]
3360 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3361 implemention in the following ways:
3363 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3366 Lack of BER streaming support means one pass streaming processing is
3367 only supported if data is detached: setting the streaming flag is
3368 ignored for embedded content.
3370 CMS support is disabled by default and must be explicitly enabled
3371 with the enable-cms configuration option.
3374 *) Update the GMP engine glue to do direct copies between BIGNUM and
3375 mpz_t when openssl and GMP use the same limb size. Otherwise the
3376 existing "conversion via a text string export" trick is still used.
3377 [Paul Sheer <paulsheer@gmail.com>]
3379 *) Zlib compression BIO. This is a filter BIO which compressed and
3380 uncompresses any data passed through it.
3383 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3384 RFC3394 compatible AES key wrapping.
3387 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3388 sets string data without copying. X509_ALGOR_set0() and
3389 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3390 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3391 from an X509_ATTRIBUTE structure optionally checking it occurs only
3392 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3396 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3397 to get the expected BN_FLG_CONSTTIME behavior.
3398 [Bodo Moeller (Google)]
3402 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3403 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3404 - added some more tests to do_tests.pl
3405 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3406 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3407 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3408 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3409 - various changes to netware.pl to enable gcc-cross builds on Win32
3411 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3412 - various changes to fix missing prototype warnings
3413 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3414 - added AES, WHIRLPOOL and CPUID assembler code to build files
3415 - added missing AES assembler make rules to mk1mf.pl
3416 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3417 [Guenter Knauf <eflash@gmx.net>]
3419 *) Implement certificate status request TLS extension defined in RFC3546.
3420 A client can set the appropriate parameters and receive the encoded
3421 OCSP response via a callback. A server can query the supplied parameters
3422 and set the encoded OCSP response in the callback. Add simplified examples
3423 to s_client and s_server.
3426 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3428 *) Fix various bugs:
3429 + Binary incompatibility of ssl_ctx_st structure
3430 + DTLS interoperation with non-compliant servers
3431 + Don't call get_session_cb() without proposed session
3432 + Fix ia64 assembler code
3433 [Andy Polyakov, Steve Henson]
3435 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3437 *) DTLS Handshake overhaul. There were longstanding issues with
3438 OpenSSL DTLS implementation, which were making it impossible for
3439 RFC 4347 compliant client to communicate with OpenSSL server.
3440 Unfortunately just fixing these incompatibilities would "cut off"
3441 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3442 server keeps tolerating non RFC compliant syntax. The opposite is
3443 not true, 0.9.8f client can not communicate with earlier server.
3444 This update even addresses CVE-2007-4995.
3447 *) Changes to avoid need for function casts in OpenSSL: some compilers
3448 (gcc 4.2 and later) reject their use.
3449 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3452 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3453 RFC4507bis. The encrypted ticket format is an encrypted encoded
3454 SSL_SESSION structure, that way new session features are automatically
3457 If a client application caches session in an SSL_SESSION structure
3458 support is transparent because tickets are now stored in the encoded
3461 The SSL_CTX structure automatically generates keys for ticket
3462 protection in servers so again support should be possible
3463 with no application modification.
3465 If a client or server wishes to disable RFC4507 support then the option
3466 SSL_OP_NO_TICKET can be set.
3468 Add a TLS extension debugging callback to allow the contents of any client
3469 or server extensions to be examined.
3471 This work was sponsored by Google.
3474 *) Add initial support for TLS extensions, specifically for the server_name
3475 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3476 have new members for a host name. The SSL data structure has an
3477 additional member SSL_CTX *initial_ctx so that new sessions can be
3478 stored in that context to allow for session resumption, even after the
3479 SSL has been switched to a new SSL_CTX in reaction to a client's
3480 server_name extension.
3482 New functions (subject to change):
3484 SSL_get_servername()
3485 SSL_get_servername_type()
3488 New CTRL codes and macros (subject to change):
3490 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3491 - SSL_CTX_set_tlsext_servername_callback()
3492 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3493 - SSL_CTX_set_tlsext_servername_arg()
3494 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3496 openssl s_client has a new '-servername ...' option.
3498 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3499 '-key2 ...', '-servername_fatal' (subject to change). This allows
3500 testing the HostName extension for a specific single host name ('-cert'
3501 and '-key' remain fallbacks for handshakes without HostName
3502 negotiation). If the unrecogninzed_name alert has to be sent, this by
3503 default is a warning; it becomes fatal with the '-servername_fatal'
3506 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3508 *) Add AES and SSE2 assembly language support to VC++ build.
3511 *) Mitigate attack on final subtraction in Montgomery reduction.
3514 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3515 (which previously caused an internal error).
3518 *) Squeeze another 10% out of IGE mode when in != out.
3521 *) AES IGE mode speedup.
3522 [Dean Gaudet (Google)]
3524 *) Add the Korean symmetric 128-bit cipher SEED (see
3525 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3526 add SEED ciphersuites from RFC 4162:
3528 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3529 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3530 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3531 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3533 To minimize changes between patchlevels in the OpenSSL 0.9.8
3534 series, SEED remains excluded from compilation unless OpenSSL
3535 is configured with 'enable-seed'.
3536 [KISA, Bodo Moeller]
3538 *) Mitigate branch prediction attacks, which can be practical if a
3539 single processor is shared, allowing a spy process to extract
3540 information. For detailed background information, see
3541 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3542 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3543 and Necessary Software Countermeasures"). The core of the change
3544 are new versions BN_div_no_branch() and
3545 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3546 respectively, which are slower, but avoid the security-relevant
3547 conditional branches. These are automatically called by BN_div()
3548 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3549 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3550 remove a conditional branch.
3552 BN_FLG_CONSTTIME is the new name for the previous
3553 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3554 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3555 in the exponent causes BN_mod_exp_mont() to use the alternative
3556 implementation in BN_mod_exp_mont_consttime().) The old name
3557 remains as a deprecated alias.
3559 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3560 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3561 constant-time implementations for more than just exponentiation.
3562 Here too the old name is kept as a deprecated alias.
3564 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3565 the BN_BLINDING structure gets an independent copy of the
3566 modulus. This means that the previous "BIGNUM *m" argument to
3567 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3568 essentially becomes "const BIGNUM *m", although we can't actually
3569 change this in the header file before 0.9.9. It allows
3570 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3571 enable BN_FLG_CONSTTIME.
3573 [Matthew D Wood (Intel Corp)]
3575 *) In the SSL/TLS server implementation, be strict about session ID
3576 context matching (which matters if an application uses a single
3577 external cache for different purposes). Previously,
3578 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3579 set. This did ensure strict client verification, but meant that,
3580 with applications using a single external cache for quite
3581 different requirements, clients could circumvent ciphersuite
3582 restrictions for a given session ID context by starting a session
3583 in a different context.
3586 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3587 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3588 authentication-only ciphersuites.
3591 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3592 not complete and could lead to a possible single byte overflow
3593 (CVE-2007-5135) [Ben Laurie]
3595 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3597 *) Since AES128 and AES256 (and similarly Camellia128 and
3598 Camellia256) share a single mask bit in the logic of
3599 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3600 kludge to work properly if AES128 is available and AES256 isn't
3601 (or if Camellia128 is available and Camellia256 isn't).
3604 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3605 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3606 When a point or a seed is encoded in a BIT STRING, we need to
3607 prevent the removal of trailing zero bits to get the proper DER
3608 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3609 of a NamedBitList, for which trailing 0 bits need to be removed.)
3612 *) Have SSL/TLS server implementation tolerate "mismatched" record
3613 protocol version while receiving ClientHello even if the
3614 ClientHello is fragmented. (The server can't insist on the
3615 particular protocol version it has chosen before the ServerHello
3616 message has informed the client about his choice.)
3619 *) Add RFC 3779 support.
3620 [Rob Austein for ARIN, Ben Laurie]
3622 *) Load error codes if they are not already present instead of using a
3623 static variable. This allows them to be cleanly unloaded and reloaded.
3624 Improve header file function name parsing.
3627 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3628 or CAPABILITY handshake as required by RFCs.
3631 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3633 *) Introduce limits to prevent malicious keys being able to
3634 cause a denial of service. (CVE-2006-2940)
3635 [Steve Henson, Bodo Moeller]
3637 *) Fix ASN.1 parsing of certain invalid structures that can result
3638 in a denial of service. (CVE-2006-2937) [Steve Henson]
3640 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3641 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3643 *) Fix SSL client code which could crash if connecting to a
3644 malicious SSLv2 server. (CVE-2006-4343)
3645 [Tavis Ormandy and Will Drewry, Google Security Team]
3647 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3648 match only those. Before that, "AES256-SHA" would be interpreted
3649 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3650 the same strength classification in 0.9.7h) as we currently only
3651 have a single AES bit in the ciphersuite description bitmap.
3652 That change, however, also applied to ciphersuite strings such as
3653 "RC4-MD5" that intentionally matched multiple ciphersuites --
3654 namely, SSL 2.0 ciphersuites in addition to the more common ones
3655 from SSL 3.0/TLS 1.0.
3657 So we change the selection algorithm again: Naming an explicit
3658 ciphersuite selects this one ciphersuite, and any other similar
3659 ciphersuite (same bitmap) from *other* protocol versions.
3660 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3661 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3663 Since SSL 2.0 does not have any ciphersuites for which the
3664 128/256 bit distinction would be relevant, this works for now.
3665 The proper fix will be to use different bits for AES128 and
3666 AES256, which would have avoided the problems from the beginning;
3667 however, bits are scarce, so we can only do this in a new release
3668 (not just a patchlevel) when we can change the SSL_CIPHER
3669 definition to split the single 'unsigned long mask' bitmap into
3670 multiple values to extend the available space.
3674 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3676 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3677 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3679 *) Add AES IGE and biIGE modes.
3682 *) Change the Unix randomness entropy gathering to use poll() when
3683 possible instead of select(), since the latter has some
3684 undesirable limitations.
3685 [Darryl Miles via Richard Levitte and Bodo Moeller]
3687 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3688 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3689 cannot be implicitly activated as part of, e.g., the "AES" alias.
3690 However, please upgrade to OpenSSL 0.9.9[-dev] for
3691 non-experimental use of the ECC ciphersuites to get TLS extension
3692 support, which is required for curve and point format negotiation
3693 to avoid potential handshake problems.
3696 *) Disable rogue ciphersuites:
3698 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3699 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3700 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3702 The latter two were purportedly from
3703 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3706 Also deactivate the remaining ciphersuites from
3707 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3708 unofficial, and the ID has long expired.
3711 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3712 dual-core machines) and other potential thread-safety issues.
3715 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3716 versions), which is now available for royalty-free use
3717 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3718 Also, add Camellia TLS ciphersuites from RFC 4132.
3720 To minimize changes between patchlevels in the OpenSSL 0.9.8
3721 series, Camellia remains excluded from compilation unless OpenSSL
3722 is configured with 'enable-camellia'.
3725 *) Disable the padding bug check when compression is in use. The padding
3726 bug check assumes the first packet is of even length, this is not
3727 necessarily true if compresssion is enabled and can result in false
3728 positives causing handshake failure. The actual bug test is ancient
3729 code so it is hoped that implementations will either have fixed it by
3730 now or any which still have the bug do not support compression.
3733 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3735 *) When applying a cipher rule check to see if string match is an explicit
3736 cipher suite and only match that one cipher suite if it is.
3739 *) Link in manifests for VC++ if needed.
3740 [Austin Ziegler <halostatue@gmail.com>]
3742 *) Update support for ECC-based TLS ciphersuites according to
3743 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3744 TLS extensions, which are supported starting with the 0.9.9
3745 branch, not in the OpenSSL 0.9.8 branch).
3748 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3749 opaque EVP_CIPHER_CTX handling.
3752 *) Fixes and enhancements to zlib compression code. We now only use
3753 "zlib1.dll" and use the default __cdecl calling convention on Win32
3754 to conform with the standards mentioned here:
3755 http://www.zlib.net/DLL_FAQ.txt
3756 Static zlib linking now works on Windows and the new --with-zlib-include
3757 --with-zlib-lib options to Configure can be used to supply the location
3758 of the headers and library. Gracefully handle case where zlib library
3762 *) Several fixes and enhancements to the OID generation code. The old code
3763 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3764 handle numbers larger than ULONG_MAX, truncated printing and had a
3765 non standard OBJ_obj2txt() behaviour.
3768 *) Add support for building of engines under engine/ as shared libraries
3769 under VC++ build system.
3772 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3773 Hopefully, we will not see any false combination of paths any more.
3776 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3778 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3779 (part of SSL_OP_ALL). This option used to disable the
3780 countermeasure against man-in-the-middle protocol-version
3781 rollback in the SSL 2.0 server implementation, which is a bad
3782 idea. (CVE-2005-2969)
3784 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3785 for Information Security, National Institute of Advanced Industrial
3786 Science and Technology [AIST], Japan)]
3788 *) Add two function to clear and return the verify parameter flags.
3791 *) Keep cipherlists sorted in the source instead of sorting them at
3792 runtime, thus removing the need for a lock.
3795 *) Avoid some small subgroup attacks in Diffie-Hellman.
3796 [Nick Mathewson and Ben Laurie]
3798 *) Add functions for well-known primes.
3801 *) Extended Windows CE support.
3802 [Satoshi Nakamura and Andy Polyakov]
3804 *) Initialize SSL_METHOD structures at compile time instead of during
3805 runtime, thus removing the need for a lock.
3808 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3809 attempting to decrypt each encrypted key in turn. Add support to
3813 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3815 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3818 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3821 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3822 key into the same file any more.
3825 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3828 *) Add -utf8 command line and config file option to 'ca'.
3829 [Stefan <stf@udoma.org]
3831 *) Removed the macro des_crypt(), as it seems to conflict with some
3832 libraries. Use DES_crypt().
3835 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3836 involves renaming the source and generated shared-libs for
3837 both. The engines will accept the corrected or legacy ids
3838 ('ncipher' and '4758_cca' respectively) when binding. NB,
3839 this only applies when building 'shared'.
3840 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3842 *) Add attribute functions to EVP_PKEY structure. Modify
3843 PKCS12_create() to recognize a CSP name attribute and
3844 use it. Make -CSP option work again in pkcs12 utility.
3847 *) Add new functionality to the bn blinding code:
3848 - automatic re-creation of the BN_BLINDING parameters after
3849 a fixed number of uses (currently 32)
3850 - add new function for parameter creation
3851 - introduce flags to control the update behaviour of the
3852 BN_BLINDING parameters
3853 - hide BN_BLINDING structure
3854 Add a second BN_BLINDING slot to the RSA structure to improve
3855 performance when a single RSA object is shared among several
3859 *) Add support for DTLS.
3860 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3862 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3863 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3866 *) Remove buggy and incompletet DH cert support from
3867 ssl/ssl_rsa.c and ssl/s3_both.c
3870 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3871 the apps/openssl applications.
3874 *) Compile clean with "-Wall -Wmissing-prototypes
3875 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3876 DEBUG_SAFESTACK must also be set.
3879 *) Change ./Configure so that certain algorithms can be disabled by default.
3880 The new counterpiece to "no-xxx" is "enable-xxx".
3882 The patented RC5 and MDC2 algorithms will now be disabled unless
3883 "enable-rc5" and "enable-mdc2", respectively, are specified.
3885 (IDEA remains enabled despite being patented. This is because IDEA
3886 is frequently required for interoperability, and there is no license
3887 fee for non-commercial use. As before, "no-idea" can be used to
3888 avoid this algorithm.)
3892 *) Add processing of proxy certificates (see RFC 3820). This work was
3893 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3894 EGEE (Enabling Grids for E-science in Europe).
3897 *) RC4 performance overhaul on modern architectures/implementations, such
3898 as Intel P4, IA-64 and AMD64.
3901 *) New utility extract-section.pl. This can be used specify an alternative
3902 section number in a pod file instead of having to treat each file as
3903 a separate case in Makefile. This can be done by adding two lines to the
3906 =for comment openssl_section:XXX
3908 The blank line is mandatory.
3912 *) New arguments -certform, -keyform and -pass for s_client and s_server
3913 to allow alternative format key and certificate files and passphrase
3917 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3918 update associated structures and add various utility functions.
3920 Add new policy related verify parameters, include policy checking in
3921 standard verify code. Enhance 'smime' application with extra parameters
3922 to support policy checking and print out.
3925 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3926 Nehemiah processors. These extensions support AES encryption in hardware
3927 as well as RNG (though RNG support is currently disabled).
3928 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3930 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3933 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3934 [Andy Polyakov and a number of other people]
3936 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3937 implementation contributed by IBM.
3938 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3940 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3941 exponent rather than 'unsigned long'. There is a corresponding change to
3942 the new 'rsa_keygen' element of the RSA_METHOD structure.
3943 [Jelte Jansen, Geoff Thorpe]
3945 *) Functionality for creating the initial serial number file is now
3946 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3948 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3949 number file to 1, which is bound to cause problems. To avoid
3950 the problems while respecting compatibility between different 0.9.7
3951 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3952 CA.pl for serial number initialization. With the new release 0.9.8,
3953 we can fix the problem directly in the 'ca' utility.)
3956 *) Reduced header interdepencies by declaring more opaque objects in
3957 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3958 give fewer recursive includes, which could break lazy source code - so
3959 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3960 developers should define this symbol when building and using openssl to
3961 ensure they track the recommended behaviour, interfaces, [etc], but
3962 backwards-compatible behaviour prevails when this isn't defined.
3965 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3968 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3969 This will generate a random key of the appropriate length based on the
3970 cipher context. The EVP_CIPHER can provide its own random key generation
3971 routine to support keys of a specific form. This is used in the des and
3972 3des routines to generate a key of the correct parity. Update S/MIME
3973 code to use new functions and hence generate correct parity DES keys.
3974 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3975 valid (weak or incorrect parity).
3978 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3979 as looking them up. This is useful when the verified structure may contain
3980 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3981 present unless the new PKCS7_NO_CRL flag is asserted.
3984 *) Extend ASN1 oid configuration module. It now additionally accepts the
3987 shortName = some long name, 1.2.3.4
3990 *) Reimplemented the BN_CTX implementation. There is now no more static
3991 limitation on the number of variables it can handle nor the depth of the
3992 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3993 information can now expand as required, and rather than having a single
3994 static array of bignums, BN_CTX now uses a linked-list of such arrays
3995 allowing it to expand on demand whilst maintaining the usefulness of
3996 BN_CTX's "bundling".
3999 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4000 to allow all RSA operations to function using a single BN_CTX.
4003 *) Preliminary support for certificate policy evaluation and checking. This
4004 is initially intended to pass the tests outlined in "Conformance Testing
4005 of Relying Party Client Certificate Path Processing Logic" v1.07.
4008 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4009 remained unused and not that useful. A variety of other little bignum
4010 tweaks and fixes have also been made continuing on from the audit (see
4014 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4015 associated ASN1, EVP and SSL functions and old ASN1 macros.
4018 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4019 and this should never fail. So the return value from the use of
4020 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4021 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4024 *) BN_CTX_get() should return zero-valued bignums, providing the same
4025 initialised value as BN_new().
4026 [Geoff Thorpe, suggested by Ulf Möller]
4028 *) Support for inhibitAnyPolicy certificate extension.
4031 *) An audit of the BIGNUM code is underway, for which debugging code is
4032 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4033 is considered valid when processing BIGNUMs, and causes execution to
4034 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4035 further steps are taken to deliberately pollute unused data in BIGNUM
4036 structures to try and expose faulty code further on. For now, openssl will
4037 (in its default mode of operation) continue to tolerate the inconsistent
4038 forms that it has tolerated in the past, but authors and packagers should
4039 consider trying openssl and their own applications when compiled with
4040 these debugging symbols defined. It will help highlight potential bugs in
4041 their own code, and will improve the test coverage for OpenSSL itself. At
4042 some point, these tighter rules will become openssl's default to improve
4043 maintainability, though the assert()s and other overheads will remain only
4044 in debugging configurations. See bn.h for more details.
4045 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4047 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4048 that can only be obtained through BN_CTX_new() (which implicitly
4049 initialises it). The presence of this function only made it possible
4050 to overwrite an existing structure (and cause memory leaks).
4053 *) Because of the callback-based approach for implementing LHASH as a
4054 template type, lh_insert() adds opaque objects to hash-tables and
4055 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4056 to clean up those corresponding objects before destroying the hash table
4057 (and losing the object pointers). So some over-zealous constifications in
4058 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4059 objects as "const" and the lh_doall[_arg] callback wrappers are not
4060 prototyped to have "const" restrictions on the object pointers they are
4061 given (and so aren't required to cast them away any more).
4064 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4065 (speed) prefers to use its own implementation. The two implementations
4066 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4067 its object type properly exposed (MS_TM) instead of casting to/from "char
4068 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4069 aren't necessarily the greatest nomenclatures - but this is what was used
4070 internally to the implementation so I've used that for now.
4073 *) Ensure that deprecated functions do not get compiled when
4074 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4075 the self-tests were still using deprecated key-generation functions so
4076 these have been updated also.
4079 *) Reorganise PKCS#7 code to separate the digest location functionality
4080 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
4081 New function PKCS7_set_digest() to set the digest type for PKCS#7
4082 digestedData type. Add additional code to correctly generate the
4083 digestedData type and add support for this type in PKCS7 initialization
4087 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4088 structure of type "other".
4091 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4092 sure the loop does correctly stop and breaking ("division by zero")
4093 modulus operations are not performed. The (pre-generated) prime
4094 table crypto/bn/bn_prime.h was already correct, but it could not be
4095 re-generated on some platforms because of the "division by zero"
4096 situation in the script.
4097 [Ralf S. Engelschall]
4099 *) Update support for ECC-based TLS ciphersuites according to
4100 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4101 SHA-1 now is only used for "small" curves (where the
4102 representation of a field element takes up to 24 bytes); for
4103 larger curves, the field element resulting from ECDH is directly
4104 used as premaster secret.
4105 [Douglas Stebila (Sun Microsystems Laboratories)]
4107 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4108 curve secp160r1 to the tests.
4109 [Douglas Stebila (Sun Microsystems Laboratories)]
4111 *) Add the possibility to load symbols globally with DSO.
4112 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4114 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4115 control of the error stack.
4118 *) Add support for STORE in ENGINE.
4121 *) Add the STORE type. The intention is to provide a common interface
4122 to certificate and key stores, be they simple file-based stores, or
4123 HSM-type store, or LDAP stores, or...
4124 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4127 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4128 pass a list of arguments to any function as well as provide a way
4129 for a function to pass data back to the caller.
4132 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4133 works like BUF_strdup() but can be used to duplicate a portion of
4134 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4138 *) Add the function sk_find_ex() which works like sk_find(), but will
4139 return an index to an element even if an exact match couldn't be
4140 found. The index is guaranteed to point at the element where the
4141 searched-for key would be inserted to preserve sorting order.
4144 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4145 takes an extra flags argument for optional functionality. Currently,
4146 the following flags are defined:
4148 OBJ_BSEARCH_VALUE_ON_NOMATCH
4149 This one gets OBJ_bsearch_ex() to return a pointer to the first
4150 element where the comparing function returns a negative or zero
4153 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4154 This one gets OBJ_bsearch_ex() to return a pointer to the first
4155 element where the comparing function returns zero. This is useful
4156 if there are more than one element where the comparing function
4160 *) Make it possible to create self-signed certificates with 'openssl ca'
4161 in such a way that the self-signed certificate becomes part of the
4162 CA database and uses the same mechanisms for serial number generation
4163 as all other certificate signing. The new flag '-selfsign' enables
4164 this functionality. Adapt CA.sh and CA.pl.in.
4167 *) Add functionality to check the public key of a certificate request
4168 against a given private. This is useful to check that a certificate
4169 request can be signed by that key (self-signing).
4172 *) Make it possible to have multiple active certificates with the same
4173 subject in the CA index file. This is done only if the keyword
4174 'unique_subject' is set to 'no' in the main CA section (default
4175 if 'CA_default') of the configuration file. The value is saved
4176 with the database itself in a separate index attribute file,
4177 named like the index file with '.attr' appended to the name.
4180 *) Generate muti valued AVAs using '+' notation in config files for
4184 *) Support for nameConstraints certificate extension.
4187 *) Support for policyConstraints certificate extension.
4190 *) Support for policyMappings certificate extension.
4193 *) Make sure the default DSA_METHOD implementation only uses its
4194 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4195 and change its own handlers to be NULL so as to remove unnecessary
4196 indirection. This lets alternative implementations fallback to the
4197 default implementation more easily.
4200 *) Support for directoryName in GeneralName related extensions
4204 *) Make it possible to link applications using Makefile.shared.
4205 Make that possible even when linking against static libraries!
4208 *) Support for single pass processing for S/MIME signing. This now
4209 means that S/MIME signing can be done from a pipe, in addition
4210 cleartext signing (multipart/signed type) is effectively streaming
4211 and the signed data does not need to be all held in memory.
4213 This is done with a new flag PKCS7_STREAM. When this flag is set
4214 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4215 is done after the data is output (and digests calculated) in
4216 SMIME_write_PKCS7().
4219 *) Add full support for -rpath/-R, both in shared libraries and
4220 applications, at least on the platforms where it's known how
4224 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4225 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4226 will now compute a table of multiples of the generator that
4227 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4228 faster (notably in the case of a single point multiplication,
4229 scalar * generator).
4230 [Nils Larsch, Bodo Moeller]
4232 *) IPv6 support for certificate extensions. The various extensions
4233 which use the IP:a.b.c.d can now take IPv6 addresses using the
4234 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4238 *) Added an ENGINE that implements RSA by performing private key
4239 exponentiations with the GMP library. The conversions to and from
4240 GMP's mpz_t format aren't optimised nor are any montgomery forms
4241 cached, and on x86 it appears OpenSSL's own performance has caught up.
4242 However there are likely to be other architectures where GMP could
4243 provide a boost. This ENGINE is not built in by default, but it can be
4244 specified at Configure time and should be accompanied by the necessary
4245 linker additions, eg;
4246 ./config -DOPENSSL_USE_GMP -lgmp
4249 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4250 testing availability of engines with "-t" - the old behaviour is
4251 produced by increasing the feature's verbosity with "-tt".
4254 *) ECDSA routines: under certain error conditions uninitialized BN objects
4255 could be freed. Solution: make sure initialization is performed early
4256 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4260 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4261 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4262 software implementations. For DSA and DH, parameter generation can
4263 also be overriden by providing the appropriate method callbacks.
4266 *) Change the "progress" mechanism used in key-generation and
4267 primality testing to functions that take a new BN_GENCB pointer in
4268 place of callback/argument pairs. The new API functions have "_ex"
4269 postfixes and the older functions are reimplemented as wrappers for
4270 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4271 declarations of the old functions to help (graceful) attempts to
4272 migrate to the new functions. Also, the new key-generation API
4273 functions operate on a caller-supplied key-structure and return
4274 success/failure rather than returning a key or NULL - this is to
4275 help make "keygen" another member function of RSA_METHOD etc.
4277 Example for using the new callback interface:
4279 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4283 BN_GENCB_set(&my_cb, my_callback, my_arg);
4285 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4286 /* For the meaning of a, b in calls to my_callback(), see the
4287 * documentation of the function that calls the callback.
4288 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4289 * my_callback should return 1 if it wants BN_is_prime_ex()
4290 * to continue, or 0 to stop.
4295 *) Change the ZLIB compression method to be stateful, and make it
4296 available to TLS with the number defined in
4297 draft-ietf-tls-compression-04.txt.
4300 *) Add the ASN.1 structures and functions for CertificatePair, which
4301 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4303 CertificatePair ::= SEQUENCE {
4304 forward [0] Certificate OPTIONAL,
4305 reverse [1] Certificate OPTIONAL,
4306 -- at least one of the pair shall be present -- }
4308 Also implement the PEM functions to read and write certificate
4309 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4311 This needed to be defined, mostly for the sake of the LDAP
4312 attribute crossCertificatePair, but may prove useful elsewhere as
4316 *) Make it possible to inhibit symlinking of shared libraries in
4317 Makefile.shared, for Cygwin's sake.
4320 *) Extend the BIGNUM API by creating a function
4321 void BN_set_negative(BIGNUM *a, int neg);
4322 and a macro that behave like
4323 int BN_is_negative(const BIGNUM *a);
4325 to avoid the need to access 'a->neg' directly in applications.
4328 *) Implement fast modular reduction for pseudo-Mersenne primes
4329 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4330 EC_GROUP_new_curve_GFp() will now automatically use this
4332 [Nils Larsch <nla@trustcenter.de>]
4334 *) Add new lock type (CRYPTO_LOCK_BN).
4337 *) Change the ENGINE framework to automatically load engines
4338 dynamically from specific directories unless they could be
4339 found to already be built in or loaded. Move all the
4340 current engines except for the cryptodev one to a new
4342 The engines in engines/ are built as shared libraries if
4343 the "shared" options was given to ./Configure or ./config.
4344 Otherwise, they are inserted in libcrypto.a.
4345 /usr/local/ssl/engines is the default directory for dynamic
4346 engines, but that can be overriden at configure time through
4347 the usual use of --prefix and/or --openssldir, and at run
4348 time with the environment variable OPENSSL_ENGINES.
4349 [Geoff Thorpe and Richard Levitte]
4351 *) Add Makefile.shared, a helper makefile to build shared
4352 libraries. Addapt Makefile.org.
4355 *) Add version info to Win32 DLLs.
4356 [Peter 'Luna' Runestig" <peter@runestig.com>]
4358 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4359 can be added using this API to created arbitrary PKCS#12
4360 files while avoiding the low level API.
4362 New options to PKCS12_create(), key or cert can be NULL and
4363 will then be omitted from the output file. The encryption
4364 algorithm NIDs can be set to -1 for no encryption, the mac
4365 iteration count can be set to 0 to omit the mac.
4367 Enhance pkcs12 utility by making the -nokeys and -nocerts
4368 options work when creating a PKCS#12 file. New option -nomac
4369 to omit the mac, NONE can be set for an encryption algorithm.
4370 New code is modified to use the enhanced PKCS12_create()
4371 instead of the low level API.
4374 *) Extend ASN1 encoder to support indefinite length constructed
4375 encoding. This can output sequences tags and octet strings in
4376 this form. Modify pk7_asn1.c to support indefinite length
4377 encoding. This is experimental and needs additional code to
4378 be useful, such as an ASN1 bio and some enhanced streaming
4381 Extend template encode functionality so that tagging is passed
4382 down to the template encoder.
4385 *) Let 'openssl req' fail if an argument to '-newkey' is not
4386 recognized instead of using RSA as a default.
4389 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4390 As these are not official, they are not included in "ALL";
4391 the "ECCdraft" ciphersuite group alias can be used to select them.
4392 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4394 *) Add ECDH engine support.
4395 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4397 *) Add ECDH in new directory crypto/ecdh/.
4398 [Douglas Stebila (Sun Microsystems Laboratories)]
4400 *) Let BN_rand_range() abort with an error after 100 iterations
4401 without success (which indicates a broken PRNG).
4404 *) Change BN_mod_sqrt() so that it verifies that the input value
4405 is really the square of the return value. (Previously,
4406 BN_mod_sqrt would show GIGO behaviour.)
4409 *) Add named elliptic curves over binary fields from X9.62, SECG,
4410 and WAP/WTLS; add OIDs that were still missing.
4412 [Sheueling Chang Shantz and Douglas Stebila
4413 (Sun Microsystems Laboratories)]
4415 *) Extend the EC library for elliptic curves over binary fields
4416 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4419 EC_GF2m_simple_method
4423 EC_GROUP_new_curve_GF2m
4424 EC_GROUP_set_curve_GF2m
4425 EC_GROUP_get_curve_GF2m
4426 EC_POINT_set_affine_coordinates_GF2m
4427 EC_POINT_get_affine_coordinates_GF2m
4428 EC_POINT_set_compressed_coordinates_GF2m
4430 Point compression for binary fields is disabled by default for
4431 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4434 As binary polynomials are represented as BIGNUMs, various members
4435 of the EC_GROUP and EC_POINT data structures can be shared
4436 between the implementations for prime fields and binary fields;
4437 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4438 are essentially identical to their ..._GFp counterparts.
4439 (For simplicity, the '..._GFp' prefix has been dropped from
4440 various internal method names.)
4442 An internal 'field_div' method (similar to 'field_mul' and
4443 'field_sqr') has been added; this is used only for binary fields.
4445 [Sheueling Chang Shantz and Douglas Stebila
4446 (Sun Microsystems Laboratories)]
4448 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4449 through methods ('mul', 'precompute_mult').
4451 The generic implementations (now internally called 'ec_wNAF_mul'
4452 and 'ec_wNAF_precomputed_mult') remain the default if these
4453 methods are undefined.
4455 [Sheueling Chang Shantz and Douglas Stebila
4456 (Sun Microsystems Laboratories)]
4458 *) New function EC_GROUP_get_degree, which is defined through
4459 EC_METHOD. For curves over prime fields, this returns the bit
4460 length of the modulus.
4462 [Sheueling Chang Shantz and Douglas Stebila
4463 (Sun Microsystems Laboratories)]
4465 *) New functions EC_GROUP_dup, EC_POINT_dup.
4466 (These simply call ..._new and ..._copy).
4468 [Sheueling Chang Shantz and Douglas Stebila
4469 (Sun Microsystems Laboratories)]
4471 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4472 Polynomials are represented as BIGNUMs (where the sign bit is not
4473 used) in the following functions [macros]:
4476 BN_GF2m_sub [= BN_GF2m_add]
4477 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4478 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4479 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4481 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4482 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4483 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4484 BN_GF2m_cmp [= BN_ucmp]
4486 (Note that only the 'mod' functions are actually for fields GF(2^m).
4487 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4489 For some functions, an the irreducible polynomial defining a
4490 field can be given as an 'unsigned int[]' with strictly
4491 decreasing elements giving the indices of those bits that are set;
4492 i.e., p[] represents the polynomial
4493 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4495 p[0] > p[1] > ... > p[k] = 0.
4496 This applies to the following functions:
4501 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4502 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4504 BN_GF2m_mod_sqrt_arr
4505 BN_GF2m_mod_solve_quad_arr
4509 Conversion can be performed by the following functions:
4514 bntest.c has additional tests for binary polynomial arithmetic.
4516 Two implementations for BN_GF2m_mod_div() are available.
4517 The default algorithm simply uses BN_GF2m_mod_inv() and
4518 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4519 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4520 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4522 [Sheueling Chang Shantz and Douglas Stebila
4523 (Sun Microsystems Laboratories)]
4525 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4526 functionality is disabled at compile-time.
4527 [Douglas Stebila <douglas.stebila@sun.com>]
4529 *) Change default behaviour of 'openssl asn1parse' so that more
4530 information is visible when viewing, e.g., a certificate:
4532 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4533 mode the content of non-printable OCTET STRINGs is output in a
4534 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4535 avoid the appearance of a printable string.
4536 [Nils Larsch <nla@trustcenter.de>]
4538 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4540 EC_GROUP_set_asn1_flag()
4541 EC_GROUP_get_asn1_flag()
4542 EC_GROUP_set_point_conversion_form()
4543 EC_GROUP_get_point_conversion_form()
4544 These control ASN1 encoding details:
4545 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4546 has been set to OPENSSL_EC_NAMED_CURVE.
4547 - Points are encoded in uncompressed form by default; options for
4548 asn1_for are as for point2oct, namely
4549 POINT_CONVERSION_COMPRESSED
4550 POINT_CONVERSION_UNCOMPRESSED
4551 POINT_CONVERSION_HYBRID
4553 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4556 EC_GROUP_get0_seed()
4557 EC_GROUP_get_seed_len()
4558 This is used only for ASN1 purposes (so far).
4559 [Nils Larsch <nla@trustcenter.de>]
4561 *) Add 'field_type' member to EC_METHOD, which holds the NID
4562 of the appropriate field type OID. The new function
4563 EC_METHOD_get_field_type() returns this value.
4564 [Nils Larsch <nla@trustcenter.de>]
4569 EC_POINT_point2hex()
4570 EC_POINT_hex2point()
4571 providing useful interfaces to EC_POINT_point2oct() and
4572 EC_POINT_oct2point().
4573 [Nils Larsch <nla@trustcenter.de>]
4575 *) Change internals of the EC library so that the functions
4576 EC_GROUP_set_generator()
4577 EC_GROUP_get_generator()
4578 EC_GROUP_get_order()
4579 EC_GROUP_get_cofactor()
4580 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4581 to methods, which would lead to unnecessary code duplication when
4582 adding different types of curves.
4583 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4585 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4586 arithmetic, and such that modified wNAFs are generated
4587 (which avoid length expansion in many cases).
4590 *) Add a function EC_GROUP_check_discriminant() (defined via
4591 EC_METHOD) that verifies that the curve discriminant is non-zero.
4593 Add a function EC_GROUP_check() that makes some sanity tests
4594 on a EC_GROUP, its generator and order. This includes
4595 EC_GROUP_check_discriminant().
4596 [Nils Larsch <nla@trustcenter.de>]
4598 *) Add ECDSA in new directory crypto/ecdsa/.
4600 Add applications 'openssl ecparam' and 'openssl ecdsa'
4601 (these are based on 'openssl dsaparam' and 'openssl dsa').
4603 ECDSA support is also included in various other files across the
4604 library. Most notably,
4605 - 'openssl req' now has a '-newkey ecdsa:file' option;
4606 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4607 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4608 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4609 them suitable for ECDSA where domain parameters must be
4610 extracted before the specific public key;
4611 - ECDSA engine support has been added.
4612 [Nils Larsch <nla@trustcenter.de>]
4614 *) Include some named elliptic curves, and add OIDs from X9.62,
4615 SECG, and WAP/WTLS. Each curve can be obtained from the new
4617 EC_GROUP_new_by_curve_name(),
4618 and the list of available named curves can be obtained with
4619 EC_get_builtin_curves().
4620 Also add a 'curve_name' member to EC_GROUP objects, which can be
4622 EC_GROUP_set_curve_name()
4623 EC_GROUP_get_curve_name()
4624 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4626 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4627 was actually never needed) and in BN_mul(). The removal in BN_mul()
4628 required a small change in bn_mul_part_recursive() and the addition
4629 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4630 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4631 bn_sub_words() and bn_add_words() except they take arrays with
4635 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4637 *) Cleanse PEM buffers before freeing them since they may contain
4639 [Benjamin Bennett <ben@psc.edu>]
4641 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4642 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4643 authentication-only ciphersuites.
4646 *) Since AES128 and AES256 share a single mask bit in the logic of
4647 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4648 kludge to work properly if AES128 is available and AES256 isn't.
4651 *) Expand security boundary to match 1.1.1 module.
4654 *) Remove redundant features: hash file source, editing of test vectors
4655 modify fipsld to use external fips_premain.c signature.
4658 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4659 run algorithm test programs.
4662 *) Make algorithm test programs more tolerant of whitespace.
4665 *) Have SSL/TLS server implementation tolerate "mismatched" record
4666 protocol version while receiving ClientHello even if the
4667 ClientHello is fragmented. (The server can't insist on the
4668 particular protocol version it has chosen before the ServerHello
4669 message has informed the client about his choice.)
4672 *) Load error codes if they are not already present instead of using a
4673 static variable. This allows them to be cleanly unloaded and reloaded.
4676 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4678 *) Introduce limits to prevent malicious keys being able to
4679 cause a denial of service. (CVE-2006-2940)
4680 [Steve Henson, Bodo Moeller]
4682 *) Fix ASN.1 parsing of certain invalid structures that can result
4683 in a denial of service. (CVE-2006-2937) [Steve Henson]
4685 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4686 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4688 *) Fix SSL client code which could crash if connecting to a
4689 malicious SSLv2 server. (CVE-2006-4343)
4690 [Tavis Ormandy and Will Drewry, Google Security Team]
4692 *) Change ciphersuite string processing so that an explicit
4693 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4694 will no longer include "AES128-SHA"), and any other similar
4695 ciphersuite (same bitmap) from *other* protocol versions (so that
4696 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4697 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4698 changes from 0.9.8b and 0.9.8d.
4701 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4703 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4704 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4706 *) Change the Unix randomness entropy gathering to use poll() when
4707 possible instead of select(), since the latter has some
4708 undesirable limitations.
4709 [Darryl Miles via Richard Levitte and Bodo Moeller]
4711 *) Disable rogue ciphersuites:
4713 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4714 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4715 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4717 The latter two were purportedly from
4718 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4721 Also deactive the remaining ciphersuites from
4722 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4723 unofficial, and the ID has long expired.
4726 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4727 dual-core machines) and other potential thread-safety issues.
4730 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4732 *) Adapt fipsld and the build system to link against the validated FIPS
4733 module in FIPS mode.
4736 *) Fixes for VC++ 2005 build under Windows.
4739 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4740 from a Windows bash shell such as MSYS. It is autodetected from the
4741 "config" script when run from a VC++ environment. Modify standard VC++
4742 build to use fipscanister.o from the GNU make build.
4745 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4747 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4748 The value now differs depending on if you build for FIPS or not.
4749 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4750 safely run with a non-FIPSed libcrypto, as it may crash because of
4751 the difference induced by this change.
4754 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4756 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4757 (part of SSL_OP_ALL). This option used to disable the
4758 countermeasure against man-in-the-middle protocol-version
4759 rollback in the SSL 2.0 server implementation, which is a bad
4760 idea. (CVE-2005-2969)
4762 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4763 for Information Security, National Institute of Advanced Industrial
4764 Science and Technology [AIST], Japan)]
4766 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4767 mainly for FIPS compliance and not fully integrated at this stage.
4770 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4771 the exponentiation using a fixed-length exponent. (Otherwise,
4772 the information leaked through timing could expose the secret key
4773 after many signatures; cf. Bleichenbacher's attack on DSA with
4777 *) Make a new fixed-window mod_exp implementation the default for
4778 RSA, DSA, and DH private-key operations so that the sequence of
4779 squares and multiplies and the memory access pattern are
4780 independent of the particular secret key. This will mitigate
4781 cache-timing and potential related attacks.
4783 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4784 and this is automatically used by BN_mod_exp_mont() if the new flag
4785 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4786 will use this BN flag for private exponents unless the flag
4787 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4788 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4790 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4792 *) Change the client implementation for SSLv23_method() and
4793 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4794 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4795 (Previously, the SSL 2.0 backwards compatible Client Hello
4796 message format would be used even with SSL_OP_NO_SSLv2.)
4799 *) Add support for smime-type MIME parameter in S/MIME messages which some
4803 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4804 a threadsafe manner. Modify rsa code to use new function and add calls
4805 to dsa and dh code (which had race conditions before).
4808 *) Include the fixed error library code in the C error file definitions
4809 instead of fixing them up at runtime. This keeps the error code
4810 structures constant.
4813 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4815 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4818 *) Fixes for newer kerberos headers. NB: the casts are needed because
4819 the 'length' field is signed on one version and unsigned on another
4820 with no (?) obvious way to tell the difference, without these VC++
4821 complains. Also the "definition" of FAR (blank) is no longer included
4822 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4823 some needed definitions.
4826 *) Undo Cygwin change.
4829 *) Added support for proxy certificates according to RFC 3820.
4830 Because they may be a security thread to unaware applications,
4831 they must be explicitely allowed in run-time. See
4832 docs/HOWTO/proxy_certificates.txt for further information.
4835 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4837 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4838 server and client random values. Previously
4839 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4840 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4842 This change has negligible security impact because:
4844 1. Server and client random values still have 24 bytes of pseudo random
4847 2. Server and client random values are sent in the clear in the initial
4850 3. The master secret is derived using the premaster secret (48 bytes in
4851 size for static RSA ciphersuites) as well as client server and random
4854 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4857 [Stephen Henson, reported by UK NISCC]
4859 *) Use Windows randomness collection on Cygwin.
4862 *) Fix hang in EGD/PRNGD query when communication socket is closed
4863 prematurely by EGD/PRNGD.
4864 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4866 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4869 *) Back-port of selected performance improvements from development
4870 branch, as well as improved support for PowerPC platforms.
4873 *) Add lots of checks for memory allocation failure, error codes to indicate
4874 failure and freeing up memory if a failure occurs.
4875 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4877 *) Add new -passin argument to dgst.
4880 *) Perform some character comparisons of different types in X509_NAME_cmp:
4881 this is needed for some certificates that reencode DNs into UTF8Strings
4882 (in violation of RFC3280) and can't or wont issue name rollover
4886 *) Make an explicit check during certificate validation to see that
4887 the CA setting in each certificate on the chain is correct. As a
4888 side effect always do the following basic checks on extensions,
4889 not just when there's an associated purpose to the check:
4891 - if there is an unhandled critical extension (unless the user
4892 has chosen to ignore this fault)
4893 - if the path length has been exceeded (if one is set at all)
4894 - that certain extensions fit the associated purpose (if one has
4898 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4900 *) Avoid a race condition when CRLs are checked in a multi threaded
4901 environment. This would happen due to the reordering of the revoked
4902 entries during signature checking and serial number lookup. Now the
4903 encoding is cached and the serial number sort performed under a lock.
4904 Add new STACK function sk_is_sorted().
4907 *) Add Delta CRL to the extension code.
4910 *) Various fixes to s3_pkt.c so alerts are sent properly.
4911 [David Holmes <d.holmes@f5.com>]
4913 *) Reduce the chances of duplicate issuer name and serial numbers (in
4914 violation of RFC3280) using the OpenSSL certificate creation utilities.
4915 This is done by creating a random 64 bit value for the initial serial
4916 number when a serial number file is created or when a self signed
4917 certificate is created using 'openssl req -x509'. The initial serial
4918 number file is created using 'openssl x509 -next_serial' in CA.pl
4919 rather than being initialized to 1.
4922 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4924 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4925 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4926 [Joe Orton, Steve Henson]
4928 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4930 [Joe Orton, Steve Henson]
4932 *) Make it possible to have multiple active certificates with the same
4933 subject in the CA index file. This is done only if the keyword
4934 'unique_subject' is set to 'no' in the main CA section (default
4935 if 'CA_default') of the configuration file. The value is saved
4936 with the database itself in a separate index attribute file,
4937 named like the index file with '.attr' appended to the name.
4940 *) X509 verify fixes. Disable broken certificate workarounds when
4941 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4942 keyUsage extension present. Don't accept CRLs with unhandled critical
4943 extensions: since verify currently doesn't process CRL extensions this
4944 rejects a CRL with *any* critical extensions. Add new verify error codes
4948 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4949 A clarification of RFC2560 will require the use of OCTET STRINGs and
4950 some implementations cannot handle the current raw format. Since OpenSSL
4951 copies and compares OCSP nonces as opaque blobs without any attempt at
4952 parsing them this should not create any compatibility issues.
4955 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4956 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4957 this HMAC (and other) operations are several times slower than OpenSSL
4961 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4962 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4964 *) Use the correct content when signing type "other".
4967 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4969 *) Fix various bugs revealed by running the NISCC test suite:
4971 Stop out of bounds reads in the ASN1 code when presented with
4972 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4974 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4976 If verify callback ignores invalid public key errors don't try to check
4977 certificate signature with the NULL public key.
4981 *) New -ignore_err option in ocsp application to stop the server
4982 exiting on the first error in a request.
4985 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4986 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4990 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4991 extra data after the compression methods not only for TLS 1.0
4992 but also for SSL 3.0 (as required by the specification).
4993 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4995 *) Change X509_certificate_type() to mark the key as exported/exportable
4996 when it's 512 *bits* long, not 512 bytes.
4999 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5000 blocks during encryption.
5003 *) Various fixes to base64 BIO and non blocking I/O. On write
5004 flushes were not handled properly if the BIO retried. On read
5005 data was not being buffered properly and had various logic bugs.
5006 This also affects blocking I/O when the data being decoded is a
5010 *) Various S/MIME bugfixes and compatibility changes:
5011 output correct application/pkcs7 MIME type if
5012 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5013 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5014 of files as .eml work). Correctly handle very long lines in MIME
5018 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5020 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5021 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5022 a protocol version number mismatch like a decryption error
5023 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5026 *) Turn on RSA blinding by default in the default implementation
5027 to avoid a timing attack. Applications that don't want it can call
5028 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5029 They would be ill-advised to do so in most cases.
5030 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5032 *) Change RSA blinding code so that it works when the PRNG is not
5033 seeded (in this case, the secret RSA exponent is abused as
5034 an unpredictable seed -- if it is not unpredictable, there
5035 is no point in blinding anyway). Make RSA blinding thread-safe
5036 by remembering the creator's thread ID in rsa->blinding and
5037 having all other threads use local one-time blinding factors
5038 (this requires more computation than sharing rsa->blinding, but
5039 avoids excessive locking; and if an RSA object is not shared
5040 between threads, blinding will still be very fast).
5043 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5044 ENGINE as defaults for all supported algorithms irrespective of
5045 the 'flags' parameter. 'flags' is now honoured, so applications
5046 should make sure they are passing it correctly.
5049 *) Target "mingw" now allows native Windows code to be generated in
5050 the Cygwin environment as well as with the MinGW compiler.
5053 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5055 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5056 via timing by performing a MAC computation even if incorrrect
5057 block cipher padding has been found. This is a countermeasure
5058 against active attacks where the attacker has to distinguish
5059 between bad padding and a MAC verification error. (CVE-2003-0078)
5061 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5062 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5063 Martin Vuagnoux (EPFL, Ilion)]
5065 *) Make the no-err option work as intended. The intention with no-err
5066 is not to have the whole error stack handling routines removed from
5067 libcrypto, it's only intended to remove all the function name and
5068 reason texts, thereby removing some of the footprint that may not
5069 be interesting if those errors aren't displayed anyway.
5071 NOTE: it's still possible for any application or module to have it's
5072 own set of error texts inserted. The routines are there, just not
5073 used by default when no-err is given.
5076 *) Add support for FreeBSD on IA64.
5077 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5079 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5080 Kerberos function mit_des_cbc_cksum(). Before this change,
5081 the value returned by DES_cbc_cksum() was like the one from
5082 mit_des_cbc_cksum(), except the bytes were swapped.
5083 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5085 *) Allow an application to disable the automatic SSL chain building.
5086 Before this a rather primitive chain build was always performed in
5087 ssl3_output_cert_chain(): an application had no way to send the
5088 correct chain if the automatic operation produced an incorrect result.
5090 Now the chain builder is disabled if either:
5092 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5094 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5096 The reasoning behind this is that an application would not want the
5097 auto chain building to take place if extra chain certificates are
5098 present and it might also want a means of sending no additional
5099 certificates (for example the chain has two certificates and the
5103 *) Add the possibility to build without the ENGINE framework.
5104 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5106 *) Under Win32 gmtime() can return NULL: check return value in
5107 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5110 *) DSA routines: under certain error conditions uninitialized BN objects
5111 could be freed. Solution: make sure initialization is performed early
5112 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5113 Nils Larsch <nla@trustcenter.de> via PR#459)
5116 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5117 checked on reconnect on the client side, therefore session resumption
5118 could still fail with a "ssl session id is different" error. This
5119 behaviour is masked when SSL_OP_ALL is used due to
5120 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5121 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5122 followup to PR #377.
5125 *) IA-32 assembler support enhancements: unified ELF targets, support
5126 for SCO/Caldera platforms, fix for Cygwin shared build.
5129 *) Add support for FreeBSD on sparc64. As a consequence, support for
5130 FreeBSD on non-x86 processors is separate from x86 processors on
5131 the config script, much like the NetBSD support.
5132 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5134 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5136 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5139 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5140 code (06) was taken as the first octet of the session ID and the last
5141 octet was ignored consequently. As a result SSLv2 client side session
5142 caching could not have worked due to the session ID mismatch between
5144 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5148 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5149 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5153 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5154 seems that in spite of existing for more than a year, many application
5155 author have done nothing to provide the necessary callbacks, which
5156 means that this particular engine will not work properly anywhere.
5157 This is a very unfortunate situation which forces us, in the name
5158 of usability, to give the hw_ncipher.c a static lock, which is part
5160 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5161 appear in 0.9.8 or later. We EXPECT application authors to have
5162 dealt properly with this when 0.9.8 is released (unless we actually
5163 make such changes in the libcrypto locking code that changes will
5164 have to be made anyway).
5167 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5168 octets have been read, EOF or an error occurs. Without this change
5169 some truncated ASN1 structures will not produce an error.
5172 *) Disable Heimdal support, since it hasn't been fully implemented.
5173 Still give the possibility to force the use of Heimdal, but with
5174 warnings and a request that patches get sent to openssl-dev.
5177 *) Add the VC-CE target, introduce the WINCE sysname, and add
5178 INSTALL.WCE and appropriate conditionals to make it build.
5179 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5181 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5182 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5183 edit numbers of the version.
5184 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5186 *) Introduce safe string copy and catenation functions
5187 (BUF_strlcpy() and BUF_strlcat()).
5188 [Ben Laurie (CHATS) and Richard Levitte]
5190 *) Avoid using fixed-size buffers for one-line DNs.
5191 [Ben Laurie (CHATS)]
5193 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5194 resizing buffers containing secrets, and use where appropriate.
5195 [Ben Laurie (CHATS)]
5197 *) Avoid using fixed size buffers for configuration file location.
5198 [Ben Laurie (CHATS)]
5200 *) Avoid filename truncation for various CA files.
5201 [Ben Laurie (CHATS)]
5203 *) Use sizeof in preference to magic numbers.
5204 [Ben Laurie (CHATS)]
5206 *) Avoid filename truncation in cert requests.
5207 [Ben Laurie (CHATS)]
5209 *) Add assertions to check for (supposedly impossible) buffer
5211 [Ben Laurie (CHATS)]
5213 *) Don't cache truncated DNS entries in the local cache (this could
5214 potentially lead to a spoofing attack).
5215 [Ben Laurie (CHATS)]
5217 *) Fix various buffers to be large enough for hex/decimal
5218 representations in a platform independent manner.
5219 [Ben Laurie (CHATS)]
5221 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5222 resizing buffers containing secrets, and use where appropriate.
5223 [Ben Laurie (CHATS)]
5225 *) Add BIO_indent() to avoid much slightly worrying code to do
5227 [Ben Laurie (CHATS)]
5229 *) Convert sprintf()/BIO_puts() to BIO_printf().
5230 [Ben Laurie (CHATS)]
5232 *) buffer_gets() could terminate with the buffer only half
5234 [Ben Laurie (CHATS)]
5236 *) Add assertions to prevent user-supplied crypto functions from
5237 overflowing internal buffers by having large block sizes, etc.
5238 [Ben Laurie (CHATS)]
5240 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5242 [Ben Laurie (CHATS)]
5244 *) Eliminate unused copy of key in RC4.
5245 [Ben Laurie (CHATS)]
5247 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5248 [Ben Laurie (CHATS)]
5250 *) Fix off-by-one error in EGD path.
5251 [Ben Laurie (CHATS)]
5253 *) If RANDFILE path is too long, ignore instead of truncating.
5254 [Ben Laurie (CHATS)]
5256 *) Eliminate unused and incorrectly sized X.509 structure
5258 [Ben Laurie (CHATS)]
5260 *) Eliminate unused and dangerous function knumber().
5261 [Ben Laurie (CHATS)]
5263 *) Eliminate unused and dangerous structure, KSSL_ERR.
5264 [Ben Laurie (CHATS)]
5266 *) Protect against overlong session ID context length in an encoded
5267 session object. Since these are local, this does not appear to be
5269 [Ben Laurie (CHATS)]
5271 *) Change from security patch (see 0.9.6e below) that did not affect
5272 the 0.9.6 release series:
5274 Remote buffer overflow in SSL3 protocol - an attacker could
5275 supply an oversized master key in Kerberos-enabled versions.
5277 [Ben Laurie (CHATS)]
5279 *) Change the SSL kerb5 codes to match RFC 2712.
5282 *) Make -nameopt work fully for req and add -reqopt switch.
5283 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5285 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5286 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5288 *) Make sure tests can be performed even if the corresponding algorithms
5289 have been removed entirely. This was also the last step to make
5290 OpenSSL compilable with DJGPP under all reasonable conditions.
5291 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5293 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5294 to allow version independent disabling of normally unselected ciphers,
5295 which may be activated as a side-effect of selecting a single cipher.
5297 (E.g., cipher list string "RSA" enables ciphersuites that are left
5298 out of "ALL" because they do not provide symmetric encryption.
5299 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5300 [Lutz Jaenicke, Bodo Moeller]
5302 *) Add appropriate support for separate platform-dependent build
5303 directories. The recommended way to make a platform-dependent
5304 build directory is the following (tested on Linux), maybe with
5307 # Place yourself outside of the OpenSSL source tree. In
5308 # this example, the environment variable OPENSSL_SOURCE
5309 # is assumed to contain the absolute OpenSSL source directory.
5310 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5311 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5312 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5313 mkdir -p `dirname $F`
5314 ln -s $OPENSSL_SOURCE/$F $F
5317 To be absolutely sure not to disturb the source tree, a "make clean"
5318 is a good thing. If it isn't successfull, don't worry about it,
5319 it probably means the source directory is very clean.
5322 *) Make sure any ENGINE control commands make local copies of string
5323 pointers passed to them whenever necessary. Otherwise it is possible
5324 the caller may have overwritten (or deallocated) the original string
5325 data when a later ENGINE operation tries to use the stored values.
5326 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5328 *) Improve diagnostics in file reading and command-line digests.
5329 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5331 *) Add AES modes CFB and OFB to the object database. Correct an
5332 error in AES-CFB decryption.
5335 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5336 allows existing EVP_CIPHER_CTX structures to be reused after
5337 calling EVP_*Final(). This behaviour is used by encryption
5338 BIOs and some applications. This has the side effect that
5339 applications must explicitly clean up cipher contexts with
5340 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5343 *) Check the values of dna and dnb in bn_mul_recursive before calling
5344 bn_mul_comba (a non zero value means the a or b arrays do not contain
5345 n2 elements) and fallback to bn_mul_normal if either is not zero.
5348 *) Fix escaping of non-ASCII characters when using the -subj option
5349 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5352 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5353 form for "surname", serialNumber has no short form.
5354 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5355 therefore remove "mail" short name for "internet 7".
5356 The OID for unique identifiers in X509 certificates is
5357 x500UniqueIdentifier, not uniqueIdentifier.
5358 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5361 *) Add an "init" command to the ENGINE config module and auto initialize
5362 ENGINEs. Without any "init" command the ENGINE will be initialized
5363 after all ctrl commands have been executed on it. If init=1 the
5364 ENGINE is initailized at that point (ctrls before that point are run
5365 on the uninitialized ENGINE and after on the initialized one). If
5366 init=0 then the ENGINE will not be iniatialized at all.
5369 *) Fix the 'app_verify_callback' interface so that the user-defined
5370 argument is actually passed to the callback: In the
5371 SSL_CTX_set_cert_verify_callback() prototype, the callback
5372 declaration has been changed from
5375 int (*cb)(X509_STORE_CTX *,void *);
5376 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5377 i=s->ctx->app_verify_callback(&ctx)
5378 has been changed into
5379 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5381 To update applications using SSL_CTX_set_cert_verify_callback(),
5382 a dummy argument can be added to their callback functions.
5383 [D. K. Smetters <smetters@parc.xerox.com>]
5385 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5386 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5388 *) Add and OPENSSL_LOAD_CONF define which will cause
5389 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5390 This allows older applications to transparently support certain
5391 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5392 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5393 load the config file and OPENSSL_add_all_algorithms_conf() which will
5394 always load it have also been added.
5397 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5398 Adjust NIDs and EVP layer.
5399 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5401 *) Config modules support in openssl utility.
5403 Most commands now load modules from the config file,
5404 though in a few (such as version) this isn't done
5405 because it couldn't be used for anything.
5407 In the case of ca and req the config file used is
5408 the same as the utility itself: that is the -config
5409 command line option can be used to specify an
5413 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5414 use "openssl_conf" if filename is NULL use default openssl config file.
5417 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5418 config section name. Add a new flag to tolerate a missing config file
5419 and move code to CONF_modules_load_file().
5422 *) Support for crypto accelerator cards from Accelerated Encryption
5423 Processing, www.aep.ie. (Use engine 'aep')
5424 The support was copied from 0.9.6c [engine] and adapted/corrected
5425 to work with the new engine framework.
5426 [AEP Inc. and Richard Levitte]
5428 *) Support for SureWare crypto accelerator cards from Baltimore
5429 Technologies. (Use engine 'sureware')
5430 The support was copied from 0.9.6c [engine] and adapted
5431 to work with the new engine framework.
5434 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5435 make the newer ENGINE framework commands for the CHIL engine work.
5436 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5438 *) Make it possible to produce shared libraries on ReliantUNIX.
5439 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5441 *) Add the configuration target debug-linux-ppro.
5442 Make 'openssl rsa' use the general key loading routines
5443 implemented in apps.c, and make those routines able to
5444 handle the key format FORMAT_NETSCAPE and the variant
5446 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5448 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5449 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5451 *) Add -keyform to rsautl, and document -engine.
5452 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5454 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5455 BIO_R_NO_SUCH_FILE error code rather than the generic
5456 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5459 *) Add new functions
5461 ERR_peek_last_error_line
5462 ERR_peek_last_error_line_data.
5463 These are similar to
5466 ERR_peek_error_line_data,
5467 but report on the latest error recorded rather than the first one
5468 still in the error queue.
5469 [Ben Laurie, Bodo Moeller]
5471 *) default_algorithms option in ENGINE config module. This allows things
5473 default_algorithms = ALL
5474 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5477 *) Prelminary ENGINE config module.
5480 *) New experimental application configuration code.
5483 *) Change the AES code to follow the same name structure as all other
5484 symmetric ciphers, and behave the same way. Move everything to
5485 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5486 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5488 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5489 [Ben Laurie and Theo de Raadt]
5491 *) Add option to output public keys in req command.
5492 [Massimiliano Pala madwolf@openca.org]
5494 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5495 (up to about 10% better than before for P-192 and P-224).
5498 *) New functions/macros
5500 SSL_CTX_set_msg_callback(ctx, cb)
5501 SSL_CTX_set_msg_callback_arg(ctx, arg)
5502 SSL_set_msg_callback(ssl, cb)
5503 SSL_set_msg_callback_arg(ssl, arg)
5505 to request calling a callback function
5507 void cb(int write_p, int version, int content_type,
5508 const void *buf, size_t len, SSL *ssl, void *arg)
5510 whenever a protocol message has been completely received
5511 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5512 protocol version according to which the SSL library interprets
5513 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5514 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5515 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5516 specification (change_cipher_spec(20), alert(21), handshake(22)).
5517 'buf' and 'len' point to the actual message, 'ssl' to the
5518 SSL object, and 'arg' is the application-defined value set by
5519 SSL[_CTX]_set_msg_callback_arg().
5521 'openssl s_client' and 'openssl s_server' have new '-msg' options
5522 to enable a callback that displays all protocol messages.
5525 *) Change the shared library support so shared libraries are built as
5526 soon as the corresponding static library is finished, and thereby get
5527 openssl and the test programs linked against the shared library.
5528 This still only happens when the keyword "shard" has been given to
5529 the configuration scripts.
5531 NOTE: shared library support is still an experimental thing, and
5532 backward binary compatibility is still not guaranteed.
5533 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5535 *) Add support for Subject Information Access extension.
5536 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5538 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5539 additional bytes when new memory had to be allocated, not just
5540 when reusing an existing buffer.
5543 *) New command line and configuration option 'utf8' for the req command.
5544 This allows field values to be specified as UTF8 strings.
5547 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5548 runs for the former and machine-readable output for the latter.
5551 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5552 of the e-mail address in the DN (i.e., it will go into a certificate
5553 extension only). The new configuration file option 'email_in_dn = no'
5554 has the same effect.
5555 [Massimiliano Pala madwolf@openca.org]
5557 *) Change all functions with names starting with des_ to be starting
5558 with DES_ instead. Add wrappers that are compatible with libdes,
5559 but are named _ossl_old_des_*. Finally, add macros that map the
5560 des_* symbols to the corresponding _ossl_old_des_* if libdes
5561 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5562 desired, the des_* symbols will be mapped to DES_*, with one
5565 Since we provide two compatibility mappings, the user needs to
5566 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5567 compatibility is desired. The default (i.e., when that macro
5568 isn't defined) is OpenSSL 0.9.6c compatibility.
5570 There are also macros that enable and disable the support of old
5571 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5572 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5573 are defined, the default will apply: to support the old des routines.
5575 In either case, one must include openssl/des.h to get the correct
5576 definitions. Do not try to just include openssl/des_old.h, that
5579 NOTE: This is a major break of an old API into a new one. Software
5580 authors are encouraged to switch to the DES_ style functions. Some
5581 time in the future, des_old.h and the libdes compatibility functions
5582 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5583 default), and then completely removed.
5586 *) Test for certificates which contain unsupported critical extensions.
5587 If such a certificate is found during a verify operation it is
5588 rejected by default: this behaviour can be overridden by either
5589 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5590 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5591 X509_supported_extension() has also been added which returns 1 if a
5592 particular extension is supported.
5595 *) Modify the behaviour of EVP cipher functions in similar way to digests
5596 to retain compatibility with existing code.
5599 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5600 compatibility with existing code. In particular the 'ctx' parameter does
5601 not have to be to be initialized before the call to EVP_DigestInit() and
5602 it is tidied up after a call to EVP_DigestFinal(). New function
5603 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5604 EVP_MD_CTX_copy() changed to not require the destination to be
5605 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5606 requires the destination to be valid.
5608 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5609 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5612 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5613 so that complete 'Handshake' protocol structures are kept in memory
5614 instead of overwriting 'msg_type' and 'length' with 'body' data.
5617 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5618 [Massimo Santin via Richard Levitte]
5620 *) Major restructuring to the underlying ENGINE code. This includes
5621 reduction of linker bloat, separation of pure "ENGINE" manipulation
5622 (initialisation, etc) from functionality dealing with implementations
5623 of specific crypto iterfaces. This change also introduces integrated
5624 support for symmetric ciphers and digest implementations - so ENGINEs
5625 can now accelerate these by providing EVP_CIPHER and EVP_MD
5626 implementations of their own. This is detailed in crypto/engine/README
5627 as it couldn't be adequately described here. However, there are a few
5628 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5629 were changed in the original introduction of ENGINE code have now
5630 reverted back - the hooking from this code to ENGINE is now a good
5631 deal more passive and at run-time, operations deal directly with
5632 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5633 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5634 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5635 they were not being used by the framework as there is no concept of a
5636 BIGNUM_METHOD and they could not be generalised to the new
5637 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5638 ENGINE_cpy() has been removed as it cannot be consistently defined in
5642 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5645 *) Change mkdef.pl to sort symbols that get the same entry number,
5646 and make sure the automatically generated functions ERR_load_*
5647 become part of libeay.num as well.
5650 *) New function SSL_renegotiate_pending(). This returns true once
5651 renegotiation has been requested (either SSL_renegotiate() call
5652 or HelloRequest/ClientHello receveived from the peer) and becomes
5653 false once a handshake has been completed.
5654 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5655 sends a HelloRequest, but does not ensure that a handshake takes
5656 place. SSL_renegotiate_pending() is useful for checking if the
5657 client has followed the request.)
5660 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5661 By default, clients may request session resumption even during
5662 renegotiation (if session ID contexts permit); with this option,
5663 session resumption is possible only in the first handshake.
5665 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5666 more bits available for options that should not be part of
5667 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5670 *) Add some demos for certificate and certificate request creation.
5673 *) Make maximum certificate chain size accepted from the peer application
5674 settable (SSL*_get/set_max_cert_list()), as proposed by
5675 "Douglas E. Engert" <deengert@anl.gov>.
5678 *) Add support for shared libraries for Unixware-7
5679 (Boyd Lynn Gerber <gerberb@zenez.com>).
5682 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5683 be done prior to destruction. Use this to unload error strings from
5684 ENGINEs that load their own error strings. NB: This adds two new API
5685 functions to "get" and "set" this destroy handler in an ENGINE.
5688 *) Alter all existing ENGINE implementations (except "openssl" and
5689 "openbsd") to dynamically instantiate their own error strings. This
5690 makes them more flexible to be built both as statically-linked ENGINEs
5691 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5692 Also, add stub code to each that makes building them as self-contained
5693 shared-libraries easier (see README.ENGINE).
5696 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5697 implementations into applications that are completely implemented in
5698 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5699 commands that can be used to configure what shared-library to load and
5700 to control aspects of the way it is handled. Also, made an update to
5701 the README.ENGINE file that brings its information up-to-date and
5702 provides some information and instructions on the "dynamic" ENGINE
5703 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5706 *) Make it possible to unload ranges of ERR strings with a new
5707 "ERR_unload_strings" function.
5710 *) Add a copy() function to EVP_MD.
5713 *) Make EVP_MD routines take a context pointer instead of just the
5714 md_data void pointer.
5717 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5718 that the digest can only process a single chunk of data
5719 (typically because it is provided by a piece of
5720 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5721 is only going to provide a single chunk of data, and hence the
5722 framework needn't accumulate the data for oneshot drivers.
5725 *) As with "ERR", make it possible to replace the underlying "ex_data"
5726 functions. This change also alters the storage and management of global
5727 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5728 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5729 index counters. The API functions that use this state have been changed
5730 to take a "class_index" rather than pointers to the class's local STACK
5731 and counter, and there is now an API function to dynamically create new
5732 classes. This centralisation allows us to (a) plug a lot of the
5733 thread-safety problems that existed, and (b) makes it possible to clean
5734 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5735 such data would previously have always leaked in application code and
5736 workarounds were in place to make the memory debugging turn a blind eye
5737 to it. Application code that doesn't use this new function will still
5738 leak as before, but their memory debugging output will announce it now
5739 rather than letting it slide.
5741 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5742 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5743 has a return value to indicate success or failure.
5746 *) Make it possible to replace the underlying "ERR" functions such that the
5747 global state (2 LHASH tables and 2 locks) is only used by the "default"
5748 implementation. This change also adds two functions to "get" and "set"
5749 the implementation prior to it being automatically set the first time
5750 any other ERR function takes place. Ie. an application can call "get",
5751 pass the return value to a module it has just loaded, and that module
5752 can call its own "set" function using that value. This means the
5753 module's "ERR" operations will use (and modify) the error state in the
5754 application and not in its own statically linked copy of OpenSSL code.
5757 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5758 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5759 the operation, and provides a more encapsulated way for external code
5760 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5761 to use these functions rather than manually incrementing the counts.
5763 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5766 *) Add EVP test program.
5769 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5772 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5773 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5774 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5775 These allow a CRL to be built without having to access X509_CRL fields
5776 directly. Modify 'ca' application to use new functions.
5779 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5780 bug workarounds. Rollback attack detection is a security feature.
5781 The problem will only arise on OpenSSL servers when TLSv1 is not
5782 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5783 Software authors not wanting to support TLSv1 will have special reasons
5784 for their choice and can explicitly enable this option.
5785 [Bodo Moeller, Lutz Jaenicke]
5787 *) Rationalise EVP so it can be extended: don't include a union of
5788 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5789 (similar to those existing for EVP_CIPHER_CTX).
5794 EVP_MD_CTX_init(&md); /* new function call */
5795 EVP_DigestInit(&md, EVP_sha1());
5796 EVP_DigestUpdate(&md, in, len);
5797 EVP_DigestFinal(&md, out, NULL);
5798 EVP_MD_CTX_cleanup(&md); /* new function call */
5802 *) Make DES key schedule conform to the usual scheme, as well as
5803 correcting its structure. This means that calls to DES functions
5804 now have to pass a pointer to a des_key_schedule instead of a
5805 plain des_key_schedule (which was actually always a pointer
5808 des_key_schedule ks;
5810 des_set_key_checked(..., &ks);
5811 des_ncbc_encrypt(..., &ks, ...);
5813 (Note that a later change renames 'des_...' into 'DES_...'.)
5816 *) Initial reduction of linker bloat: the use of some functions, such as
5817 PEM causes large amounts of unused functions to be linked in due to
5818 poor organisation. For example pem_all.c contains every PEM function
5819 which has a knock on effect of linking in large amounts of (unused)
5820 ASN1 code. Grouping together similar functions and splitting unrelated
5821 functions prevents this.
5824 *) Cleanup of EVP macros.
5827 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5828 correct _ecb suffix.
5831 *) Add initial OCSP responder support to ocsp application. The
5832 revocation information is handled using the text based index
5833 use by the ca application. The responder can either handle
5834 requests generated internally, supplied in files (for example
5835 via a CGI script) or using an internal minimal server.
5838 *) Add configuration choices to get zlib compression for TLS.
5841 *) Changes to Kerberos SSL for RFC 2712 compliance:
5842 1. Implemented real KerberosWrapper, instead of just using
5843 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5844 2. Implemented optional authenticator field of KerberosWrapper.
5846 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5847 and authenticator structs; see crypto/krb5/.
5849 Generalized Kerberos calls to support multiple Kerberos libraries.
5850 [Vern Staats <staatsvr@asc.hpc.mil>,
5851 Jeffrey Altman <jaltman@columbia.edu>
5852 via Richard Levitte]
5854 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5855 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5856 values for each of the key sizes rather than having just
5857 parameters (and 'speed' generating keys each time).
5860 *) Speed up EVP routines.
5863 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5864 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5865 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5866 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5868 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5869 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5870 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5873 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5875 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5878 *) Added the OS2-EMX target.
5879 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5881 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5882 to support NCONF routines in extension code. New function CONF_set_nconf()
5883 to allow functions which take an NCONF to also handle the old LHASH
5884 structure: this means that the old CONF compatible routines can be
5885 retained (in particular wrt extensions) without having to duplicate the
5886 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5889 *) Enhance the general user interface with mechanisms for inner control
5890 and with possibilities to have yes/no kind of prompts.
5893 *) Change all calls to low level digest routines in the library and
5894 applications to use EVP. Add missing calls to HMAC_cleanup() and
5895 don't assume HMAC_CTX can be copied using memcpy().
5896 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5898 *) Add the possibility to control engines through control names but with
5899 arbitrary arguments instead of just a string.
5900 Change the key loaders to take a UI_METHOD instead of a callback
5901 function pointer. NOTE: this breaks binary compatibility with earlier
5902 versions of OpenSSL [engine].
5903 Adapt the nCipher code for these new conditions and add a card insertion
5907 *) Enhance the general user interface with mechanisms to better support
5908 dialog box interfaces, application-defined prompts, the possibility
5909 to use defaults (for example default passwords from somewhere else)
5910 and interrupts/cancellations.
5913 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5914 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5917 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5918 tidy up some unnecessarily weird code in 'sk_new()').
5919 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5921 *) Change the key loading routines for ENGINEs to use the same kind
5922 callback (pem_password_cb) as all other routines that need this
5926 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5927 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5928 than this minimum value is recommended.
5931 *) New random seeder for OpenVMS, using the system process statistics
5932 that are easily reachable.
5935 *) Windows apparently can't transparently handle global
5936 variables defined in DLLs. Initialisations such as:
5938 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5940 wont compile. This is used by the any applications that need to
5941 declare their own ASN1 modules. This was fixed by adding the option
5942 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5943 needed for static libraries under Win32.
5946 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5947 setting of purpose and trust fields. New X509_STORE trust and
5948 purpose functions and tidy up setting in other SSL functions.
5951 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5952 structure. These are inherited by X509_STORE_CTX when it is
5953 initialised. This allows various defaults to be set in the
5954 X509_STORE structure (such as flags for CRL checking and custom
5955 purpose or trust settings) for functions which only use X509_STORE_CTX
5956 internally such as S/MIME.
5958 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5959 trust settings if they are not set in X509_STORE. This allows X509_STORE
5960 purposes and trust (in S/MIME for example) to override any set by default.
5962 Add command line options for CRL checking to smime, s_client and s_server
5966 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5967 are set then the CRL is looked up in the X509_STORE structure and
5968 its validity and signature checked, then if the certificate is found
5969 in the CRL the verify fails with a revoked error.
5971 Various new CRL related callbacks added to X509_STORE_CTX structure.
5973 Command line options added to 'verify' application to support this.
5975 This needs some additional work, such as being able to handle multiple
5976 CRLs with different times, extension based lookup (rather than just
5977 by subject name) and ultimately more complete V2 CRL extension
5981 *) Add a general user interface API (crypto/ui/). This is designed
5982 to replace things like des_read_password and friends (backward
5983 compatibility functions using this new API are provided).
5984 The purpose is to remove prompting functions from the DES code
5985 section as well as provide for prompting through dialog boxes in
5986 a window system and the like.
5989 *) Add "ex_data" support to ENGINE so implementations can add state at a
5990 per-structure level rather than having to store it globally.
5993 *) Make it possible for ENGINE structures to be copied when retrieved by
5994 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5995 This causes the "original" ENGINE structure to act like a template,
5996 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5997 operational state can be localised to each ENGINE structure, despite the
5998 fact they all share the same "methods". New ENGINE structures returned in
5999 this case have no functional references and the return value is the single
6000 structural reference. This matches the single structural reference returned
6001 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6005 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6006 needs to match any other type at all we need to manually clear the
6010 *) Changes to the "openssl engine" utility to include;
6011 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6012 about an ENGINE's available control commands.
6013 - executing control commands from command line arguments using the
6014 '-pre' and '-post' switches. '-post' is only used if '-t' is
6015 specified and the ENGINE is successfully initialised. The syntax for
6016 the individual commands are colon-separated, for example;
6017 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6020 *) New dynamic control command support for ENGINEs. ENGINEs can now
6021 declare their own commands (numbers), names (strings), descriptions,
6022 and input types for run-time discovery by calling applications. A
6023 subset of these commands are implicitly classed as "executable"
6024 depending on their input type, and only these can be invoked through
6025 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6026 can be based on user input, config files, etc). The distinction is
6027 that "executable" commands cannot return anything other than a boolean
6028 result and can only support numeric or string input, whereas some
6029 discoverable commands may only be for direct use through
6030 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6031 pointers, or other custom uses. The "executable" commands are to
6032 support parameterisations of ENGINE behaviour that can be
6033 unambiguously defined by ENGINEs and used consistently across any
6034 OpenSSL-based application. Commands have been added to all the
6035 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6036 control over shared-library paths without source code alterations.
6039 *) Changed all ENGINE implementations to dynamically allocate their
6040 ENGINEs rather than declaring them statically. Apart from this being
6041 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6042 this also allows the implementations to compile without using the
6043 internal engine_int.h header.
6046 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6047 'const' value. Any code that should be able to modify a RAND_METHOD
6048 should already have non-const pointers to it (ie. they should only
6049 modify their own ones).
6052 *) Made a variety of little tweaks to the ENGINE code.
6053 - "atalla" and "ubsec" string definitions were moved from header files
6054 to C code. "nuron" string definitions were placed in variables
6055 rather than hard-coded - allowing parameterisation of these values
6056 later on via ctrl() commands.
6057 - Removed unused "#if 0"'d code.
6058 - Fixed engine list iteration code so it uses ENGINE_free() to release
6059 structural references.
6060 - Constified the RAND_METHOD element of ENGINE structures.
6061 - Constified various get/set functions as appropriate and added
6062 missing functions (including a catch-all ENGINE_cpy that duplicates
6063 all ENGINE values onto a new ENGINE except reference counts/state).
6064 - Removed NULL parameter checks in get/set functions. Setting a method
6065 or function to NULL is a way of cancelling out a previously set
6066 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6067 and doesn't justify the extra error symbols and code.
6068 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6069 flags from engine_int.h to engine.h.
6070 - Changed prototypes for ENGINE handler functions (init(), finish(),
6071 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6074 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6075 to the algorithm using long division. The binary algorithm can be
6076 used only if the modulus is odd. On 32-bit systems, it is faster
6077 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6078 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6079 up to 450 bits. In 64-bit environments, the binary algorithm
6080 appears to be advantageous for much longer moduli; here we use it
6081 for moduli up to 2048 bits.
6084 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6085 could not support the combine flag in choice fields.
6088 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6089 extensions from a certificate request to the certificate.
6092 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6093 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6094 file: this allows the display of the certificate about to be
6095 signed to be customised, to allow certain fields to be included
6096 or excluded and extension details. The old system didn't display
6097 multicharacter strings properly, omitted fields not in the policy
6098 and couldn't display additional details such as extensions.
6101 *) Function EC_POINTs_mul for multiple scalar multiplication
6102 of an arbitrary number of elliptic curve points
6103 \sum scalars[i]*points[i],
6104 optionally including the generator defined for the EC_GROUP:
6105 scalar*generator + \sum scalars[i]*points[i].
6107 EC_POINT_mul is a simple wrapper function for the typical case
6108 that the point list has just one item (besides the optional
6112 *) First EC_METHODs for curves over GF(p):
6114 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6115 operations and provides various method functions that can also
6116 operate with faster implementations of modular arithmetic.
6118 EC_GFp_mont_method() reuses most functions that are part of
6119 EC_GFp_simple_method, but uses Montgomery arithmetic.
6121 [Bodo Moeller; point addition and point doubling
6122 implementation directly derived from source code provided by
6123 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6125 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6126 crypto/ec/ec_lib.c):
6128 Curves are EC_GROUP objects (with an optional group generator)
6129 based on EC_METHODs that are built into the library.
6131 Points are EC_POINT objects based on EC_GROUP objects.
6133 Most of the framework would be able to handle curves over arbitrary
6134 finite fields, but as there are no obvious types for fields other
6135 than GF(p), some functions are limited to that for now.
6138 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6139 that the file contains a complete HTTP response.
6142 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6143 change the def and num file printf format specifier from "%-40sXXX"
6144 to "%-39s XXX". The latter will always guarantee a space after the
6145 field while the former will cause them to run together if the field
6146 is 40 of more characters long.
6149 *) Constify the cipher and digest 'method' functions and structures
6150 and modify related functions to take constant EVP_MD and EVP_CIPHER
6154 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6155 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6158 *) Modify EVP_Digest*() routines so they now return values. Although the
6159 internal software routines can never fail additional hardware versions
6163 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6165 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6166 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6169 ERR_R_NESTED_ASN1_ERROR
6171 ERR_R_MISSING_ASN1_EOS
6172 were 4 .. 9, conflicting with
6173 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6175 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6176 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6178 Add new error code 'ERR_R_INTERNAL_ERROR'.
6181 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6185 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6186 sets the subject name for a new request or supersedes the
6187 subject name in a given request. Formats that can be parsed are
6188 'CN=Some Name, OU=myOU, C=IT'
6190 'CN=Some Name/OU=myOU/C=IT'.
6192 Add options '-batch' and '-verbose' to 'openssl req'.
6193 [Massimiliano Pala <madwolf@hackmasters.net>]
6195 *) Introduce the possibility to access global variables through
6196 functions on platform were that's the best way to handle exporting
6197 global variables in shared libraries. To enable this functionality,
6198 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6199 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6200 is normally done by Configure or something similar).
6202 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6203 in the source file (foo.c) like this:
6205 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6206 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6208 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6209 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6211 OPENSSL_DECLARE_GLOBAL(int,foo);
6212 #define foo OPENSSL_GLOBAL_REF(foo)
6213 OPENSSL_DECLARE_GLOBAL(double,bar);
6214 #define bar OPENSSL_GLOBAL_REF(bar)
6216 The #defines are very important, and therefore so is including the
6217 header file everywhere where the defined globals are used.
6219 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6220 of ASN.1 items, but that structure is a bit different.
6222 The largest change is in util/mkdef.pl which has been enhanced with
6223 better and easier to understand logic to choose which symbols should
6224 go into the Windows .def files as well as a number of fixes and code
6225 cleanup (among others, algorithm keywords are now sorted
6226 lexicographically to avoid constant rewrites).
6229 *) In BN_div() keep a copy of the sign of 'num' before writing the
6230 result to 'rm' because if rm==num the value will be overwritten
6231 and produce the wrong result if 'num' is negative: this caused
6232 problems with BN_mod() and BN_nnmod().
6235 *) Function OCSP_request_verify(). This checks the signature on an
6236 OCSP request and verifies the signer certificate. The signer
6237 certificate is just checked for a generic purpose and OCSP request
6241 *) Add OCSP_check_validity() function to check the validity of OCSP
6242 responses. OCSP responses are prepared in real time and may only
6243 be a few seconds old. Simply checking that the current time lies
6244 between thisUpdate and nextUpdate max reject otherwise valid responses
6245 caused by either OCSP responder or client clock inaccuracy. Instead
6246 we allow thisUpdate and nextUpdate to fall within a certain period of
6247 the current time. The age of the response can also optionally be
6248 checked. Two new options -validity_period and -status_age added to
6252 *) If signature or public key algorithm is unrecognized print out its
6253 OID rather that just UNKNOWN.
6256 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6257 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6258 ID to be generated from the issuer certificate alone which can then be
6259 passed to OCSP_id_issuer_cmp().
6262 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6263 ASN1 modules to export functions returning ASN1_ITEM pointers
6264 instead of the ASN1_ITEM structures themselves. This adds several
6265 new macros which allow the underlying ASN1 function/structure to
6266 be accessed transparently. As a result code should not use ASN1_ITEM
6267 references directly (such as &X509_it) but instead use the relevant
6268 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6269 use of the new ASN1 code on platforms where exporting structures
6270 is problematical (for example in shared libraries) but exporting
6271 functions returning pointers to structures is not.
6274 *) Add support for overriding the generation of SSL/TLS session IDs.
6275 These callbacks can be registered either in an SSL_CTX or per SSL.
6276 The purpose of this is to allow applications to control, if they wish,
6277 the arbitrary values chosen for use as session IDs, particularly as it
6278 can be useful for session caching in multiple-server environments. A
6279 command-line switch for testing this (and any client code that wishes
6280 to use such a feature) has been added to "s_server".
6281 [Geoff Thorpe, Lutz Jaenicke]
6283 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6284 of the form '#if defined(...) || defined(...) || ...' and
6285 '#if !defined(...) && !defined(...) && ...'. This also avoids
6286 the growing number of special cases it was previously handling.
6289 *) Make all configuration macros available for application by making
6290 sure they are available in opensslconf.h, by giving them names starting
6291 with "OPENSSL_" to avoid conflicts with other packages and by making
6292 sure e_os2.h will cover all platform-specific cases together with
6294 Additionally, it is now possible to define configuration/platform-
6295 specific names (called "system identities"). In the C code, these
6296 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6297 macro with the name beginning with "OPENSSL_SYS_", which is determined
6298 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6302 *) New option -set_serial to 'req' and 'x509' this allows the serial
6303 number to use to be specified on the command line. Previously self
6304 signed certificates were hard coded with serial number 0 and the
6305 CA options of 'x509' had to use a serial number in a file which was
6309 *) New options to 'ca' utility to support V2 CRL entry extensions.
6310 Currently CRL reason, invalidity date and hold instruction are
6311 supported. Add new CRL extensions to V3 code and some new objects.
6314 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6315 disable standard block padding (aka PKCS#5 padding) in the EVP
6316 API, which was previously mandatory. This means that the data is
6317 not padded in any way and so the total length much be a multiple
6318 of the block size, otherwise an error occurs.
6321 *) Initial (incomplete) OCSP SSL support.
6324 *) New function OCSP_parse_url(). This splits up a URL into its host,
6325 port and path components: primarily to parse OCSP URLs. New -url
6326 option to ocsp utility.
6329 *) New nonce behavior. The return value of OCSP_check_nonce() now
6330 reflects the various checks performed. Applications can decide
6331 whether to tolerate certain situations such as an absent nonce
6332 in a response when one was present in a request: the ocsp application
6333 just prints out a warning. New function OCSP_add1_basic_nonce()
6334 this is to allow responders to include a nonce in a response even if
6335 the request is nonce-less.
6338 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6339 skipped when using openssl x509 multiple times on a single input file,
6340 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6343 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6344 set string type: to handle setting ASN1_TIME structures. Fix ca
6345 utility to correctly initialize revocation date of CRLs.
6348 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6349 the clients preferred ciphersuites and rather use its own preferences.
6350 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6351 Internet Explorer by ensuring unchanged hash method during stepup.
6352 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6355 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6356 to aes and add a new 'exist' option to print out symbols that don't
6360 *) Additional options to ocsp utility to allow flags to be set and
6361 additional certificates supplied.
6364 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6365 OCSP client a number of certificate to only verify the response
6369 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6370 handle the new API. Currently only ECB, CBC modes supported. Add new
6373 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6374 Encryption Standard (AES) Ciphersuites for Transport Layer
6375 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6376 not enabled by default and were not part of the "ALL" ciphersuite
6377 alias because they were not yet official; they could be
6378 explicitly requested by specifying the "AESdraft" ciphersuite
6379 group alias. In the final release of OpenSSL 0.9.7, the group
6380 alias is called "AES" and is part of "ALL".)
6381 [Ben Laurie, Steve Henson, Bodo Moeller]
6383 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6384 request to response.
6387 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6388 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6389 extract information from a certificate request. OCSP_response_create()
6390 creates a response and optionally adds a basic response structure.
6391 OCSP_basic_add1_status() adds a complete single response to a basic
6392 response and returns the OCSP_SINGLERESP structure just added (to allow
6393 extensions to be included for example). OCSP_basic_add1_cert() adds a
6394 certificate to a basic response and OCSP_basic_sign() signs a basic
6395 response with various flags. New helper functions ASN1_TIME_check()
6396 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6397 (converts ASN1_TIME to GeneralizedTime).
6400 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6401 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6402 structure from a certificate. X509_pubkey_digest() digests the public_key
6403 contents: this is used in various key identifiers.
6406 *) Make sk_sort() tolerate a NULL argument.
6407 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6409 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6410 passed by the function are trusted implicitly. If any of them signed the
6411 response then it is assumed to be valid and is not verified.
6414 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6415 to data. This was previously part of the PKCS7 ASN1 code. This
6416 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6417 [Steve Henson, reported by Kenneth R. Robinette
6418 <support@securenetterm.com>]
6420 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6421 routines: without these tracing memory leaks is very painful.
6422 Fix leaks in PKCS12 and PKCS7 routines.
6425 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6426 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6427 effectively meant GeneralizedTime would never be used. Now it
6428 is initialised to -1 but X509_time_adj() now has to check the value
6429 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6430 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6431 [Steve Henson, reported by Kenneth R. Robinette
6432 <support@securenetterm.com>]
6434 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6435 result in a zero length in the ASN1_INTEGER structure which was
6436 not consistent with the structure when d2i_ASN1_INTEGER() was used
6437 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6438 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6439 where it did not print out a minus for negative ASN1_INTEGER.
6442 *) Add summary printout to ocsp utility. The various functions which
6443 convert status values to strings have been renamed to:
6444 OCSP_response_status_str(), OCSP_cert_status_str() and
6445 OCSP_crl_reason_str() and are no longer static. New options
6446 to verify nonce values and to disable verification. OCSP response
6447 printout format cleaned up.
6450 *) Add additional OCSP certificate checks. These are those specified
6451 in RFC2560. This consists of two separate checks: the CA of the
6452 certificate being checked must either be the OCSP signer certificate
6453 or the issuer of the OCSP signer certificate. In the latter case the
6454 OCSP signer certificate must contain the OCSP signing extended key
6455 usage. This check is performed by attempting to match the OCSP
6456 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6457 in the OCSP_CERTID structures of the response.
6460 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6461 and related routines. This uses the standard OpenSSL certificate
6462 verify routines to perform initial checks (just CA validity) and
6463 to obtain the certificate chain. Then additional checks will be
6464 performed on the chain. Currently the root CA is checked to see
6465 if it is explicitly trusted for OCSP signing. This is used to set
6466 a root CA as a global signing root: that is any certificate that
6467 chains to that CA is an acceptable OCSP signing certificate.
6470 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6471 extensions from a separate configuration file.
6472 As when reading extensions from the main configuration file,
6473 the '-extensions ...' option may be used for specifying the
6475 [Massimiliano Pala <madwolf@comune.modena.it>]
6477 *) New OCSP utility. Allows OCSP requests to be generated or
6478 read. The request can be sent to a responder and the output
6479 parsed, outputed or printed in text form. Not complete yet:
6480 still needs to check the OCSP response validity.
6483 *) New subcommands for 'openssl ca':
6484 'openssl ca -status <serial>' prints the status of the cert with
6485 the given serial number (according to the index file).
6486 'openssl ca -updatedb' updates the expiry status of certificates
6488 [Massimiliano Pala <madwolf@comune.modena.it>]
6490 *) New '-newreq-nodes' command option to CA.pl. This is like
6491 '-newreq', but calls 'openssl req' with the '-nodes' option
6492 so that the resulting key is not encrypted.
6493 [Damien Miller <djm@mindrot.org>]
6495 *) New configuration for the GNU Hurd.
6496 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6498 *) Initial code to implement OCSP basic response verify. This
6499 is currently incomplete. Currently just finds the signer's
6500 certificate and verifies the signature on the response.
6503 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6504 value of OPENSSLDIR. This is available via the new '-d' option
6505 to 'openssl version', and is also included in 'openssl version -a'.
6508 *) Allowing defining memory allocation callbacks that will be given
6509 file name and line number information in additional arguments
6510 (a const char* and an int). The basic functionality remains, as
6511 well as the original possibility to just replace malloc(),
6512 realloc() and free() by functions that do not know about these
6513 additional arguments. To register and find out the current
6514 settings for extended allocation functions, the following
6515 functions are provided:
6517 CRYPTO_set_mem_ex_functions
6518 CRYPTO_set_locked_mem_ex_functions
6519 CRYPTO_get_mem_ex_functions
6520 CRYPTO_get_locked_mem_ex_functions
6522 These work the same way as CRYPTO_set_mem_functions and friends.
6523 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6524 extended allocation function is enabled.
6525 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6526 a conventional allocation function is enabled.
6527 [Richard Levitte, Bodo Moeller]
6529 *) Finish off removing the remaining LHASH function pointer casts.
6530 There should no longer be any prototype-casting required when using
6531 the LHASH abstraction, and any casts that remain are "bugs". See
6532 the callback types and macros at the head of lhash.h for details
6533 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6536 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6537 If /dev/[u]random devices are not available or do not return enough
6538 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6540 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6541 /etc/entropy will be queried once each in this sequence, quering stops
6542 when enough entropy was collected without querying more sockets.
6545 *) Change the Unix RAND_poll() variant to be able to poll several
6546 random devices, as specified by DEVRANDOM, until a sufficient amount
6547 of data has been collected. We spend at most 10 ms on each file
6548 (select timeout) and read in non-blocking mode. DEVRANDOM now
6549 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6550 (previously it was just the string "/dev/urandom"), so on typical
6551 platforms the 10 ms delay will never occur.
6552 Also separate out the Unix variant to its own file, rand_unix.c.
6553 For VMS, there's a currently-empty rand_vms.c.
6556 *) Move OCSP client related routines to ocsp_cl.c. These
6557 provide utility functions which an application needing
6558 to issue a request to an OCSP responder and analyse the
6559 response will typically need: as opposed to those which an
6560 OCSP responder itself would need which will be added later.
6562 OCSP_request_sign() signs an OCSP request with an API similar
6563 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6564 response. OCSP_response_get1_basic() extracts basic response
6565 from response. OCSP_resp_find_status(): finds and extracts status
6566 information from an OCSP_CERTID structure (which will be created
6567 when the request structure is built). These are built from lower
6568 level functions which work on OCSP_SINGLERESP structures but
6569 wont normally be used unless the application wishes to examine
6570 extensions in the OCSP response for example.
6572 Replace nonce routines with a pair of functions.
6573 OCSP_request_add1_nonce() adds a nonce value and optionally
6574 generates a random value. OCSP_check_nonce() checks the
6575 validity of the nonce in an OCSP response.
6578 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6579 This doesn't copy the supplied OCSP_CERTID and avoids the
6580 need to free up the newly created id. Change return type
6581 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6582 This can then be used to add extensions to the request.
6583 Deleted OCSP_request_new(), since most of its functionality
6584 is now in OCSP_REQUEST_new() (and the case insensitive name
6585 clash) apart from the ability to set the request name which
6586 will be added elsewhere.
6589 *) Update OCSP API. Remove obsolete extensions argument from
6590 various functions. Extensions are now handled using the new
6591 OCSP extension code. New simple OCSP HTTP function which
6592 can be used to send requests and parse the response.
6595 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6596 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6597 uses the special reorder version of SET OF to sort the attributes
6598 and reorder them to match the encoded order. This resolves a long
6599 standing problem: a verify on a PKCS7 structure just after signing
6600 it used to fail because the attribute order did not match the
6601 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6602 it uses the received order. This is necessary to tolerate some broken
6603 software that does not order SET OF. This is handled by encoding
6604 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6605 to produce the required SET OF.
6608 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6609 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6610 files to get correct declarations of the ASN.1 item variables.
6613 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6614 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6615 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6616 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6617 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6618 ASN1_ITEM and no wrapper functions.
6621 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6622 replace the old function pointer based I/O routines. Change most of
6623 the *_d2i_bio() and *_d2i_fp() functions to use these.
6626 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6627 lines, recognice more "algorithms" that can be deselected, and make
6628 it complain about algorithm deselection that isn't recognised.
6631 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6632 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6633 to use new functions. Add NO_ASN1_OLD which can be set to remove
6634 some old style ASN1 functions: this can be used to determine if old
6635 code will still work when these eventually go away.
6638 *) New extension functions for OCSP structures, these follow the
6639 same conventions as certificates and CRLs.
6642 *) New function X509V3_add1_i2d(). This automatically encodes and
6643 adds an extension. Its behaviour can be customised with various
6644 flags to append, replace or delete. Various wrappers added for
6645 certifcates and CRLs.
6648 *) Fix to avoid calling the underlying ASN1 print routine when
6649 an extension cannot be parsed. Correct a typo in the
6650 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6653 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
6654 entries for variables.
6657 *) Add functionality to apps/openssl.c for detecting locking
6658 problems: As the program is single-threaded, all we have
6659 to do is register a locking callback using an array for
6660 storing which locks are currently held by the program.
6663 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6664 SSL_get_ex_data_X509_STORE_idx(), which is used in
6665 ssl_verify_cert_chain() and thus can be called at any time
6666 during TLS/SSL handshakes so that thread-safety is essential.
6667 Unfortunately, the ex_data design is not at all suited
6668 for multi-threaded use, so it probably should be abolished.
6671 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6672 [Broadcom, tweaked and integrated by Geoff Thorpe]
6674 *) Move common extension printing code to new function
6675 X509V3_print_extensions(). Reorganise OCSP print routines and
6676 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6679 *) New function X509_signature_print() to remove duplication in some
6683 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6684 set (this was treated exactly the same as SET OF previously). This
6685 is used to reorder the STACK representing the structure to match the
6686 encoding. This will be used to get round a problem where a PKCS7
6687 structure which was signed could not be verified because the STACK
6688 order did not reflect the encoded order.
6691 *) Reimplement the OCSP ASN1 module using the new code.
6694 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6695 for its ASN1 operations. The old style function pointers still exist
6696 for now but they will eventually go away.
6699 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6700 completely replaces the old ASN1 functionality with a table driven
6701 encoder and decoder which interprets an ASN1_ITEM structure describing
6702 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6703 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6704 has also been converted to the new form.
6707 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6708 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6709 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6710 for negative moduli.
6713 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6714 of not touching the result's sign bit.
6717 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6721 *) Changed the LHASH code to use prototypes for callbacks, and created
6722 macros to declare and implement thin (optionally static) functions
6723 that provide type-safety and avoid function pointer casting for the
6724 type-specific callbacks.
6727 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6729 [Veers Staats <staatsvr@asc.hpc.mil>,
6730 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6732 *) Reformat the FAQ so the different questions and answers can be divided
6733 in sections depending on the subject.
6736 *) Have the zlib compression code load ZLIB.DLL dynamically under
6740 *) New function BN_mod_sqrt for computing square roots modulo a prime
6741 (using the probabilistic Tonelli-Shanks algorithm unless
6742 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6743 be handled deterministically).
6744 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6746 *) Make BN_mod_inverse faster by explicitly handling small quotients
6747 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6748 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6751 *) New function BN_kronecker.
6754 *) Fix BN_gcd so that it works on negative inputs; the result is
6755 positive unless both parameters are zero.
6756 Previously something reasonably close to an infinite loop was
6757 possible because numbers could be growing instead of shrinking
6758 in the implementation of Euclid's algorithm.
6761 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6762 sign of the number in question.
6764 Fix BN_is_word(a,w) to work correctly for w == 0.
6766 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6767 because its test if the absolute value of 'a' equals 'w'.
6768 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6769 it exists mostly for use in the implementations of BN_is_zero(),
6770 BN_is_one(), and BN_is_word().
6773 *) New function BN_swap.
6776 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6777 the exponentiation functions are more likely to produce reasonable
6778 results on negative inputs.
6781 *) Change BN_mod_mul so that the result is always non-negative.
6782 Previously, it could be negative if one of the factors was negative;
6783 I don't think anyone really wanted that behaviour.
6786 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6787 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6788 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6789 and add new functions:
6798 BN_mod_lshift1_quick
6802 These functions always generate non-negative results.
6804 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6805 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6807 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6808 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6809 be reduced modulo m.
6810 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6813 The following entry accidentily appeared in the CHANGES file
6814 distributed with OpenSSL 0.9.7. The modifications described in
6815 it do *not* apply to OpenSSL 0.9.7.
6817 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6818 was actually never needed) and in BN_mul(). The removal in BN_mul()
6819 required a small change in bn_mul_part_recursive() and the addition
6820 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6821 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6822 bn_sub_words() and bn_add_words() except they take arrays with
6827 *) In 'openssl passwd', verify passwords read from the terminal
6828 unless the '-salt' option is used (which usually means that
6829 verification would just waste user's time since the resulting
6830 hash is going to be compared with some given password hash)
6831 or the new '-noverify' option is used.
6833 This is an incompatible change, but it does not affect
6834 non-interactive use of 'openssl passwd' (passwords on the command
6835 line, '-stdin' option, '-in ...' option) and thus should not
6839 *) Remove all references to RSAref, since there's no more need for it.
6842 *) Make DSO load along a path given through an environment variable
6843 (SHLIB_PATH) with shl_load().
6846 *) Constify the ENGINE code as a result of BIGNUM constification.
6847 Also constify the RSA code and most things related to it. In a
6848 few places, most notable in the depth of the ASN.1 code, ugly
6849 casts back to non-const were required (to be solved at a later
6853 *) Make it so the openssl application has all engines loaded by default.
6856 *) Constify the BIGNUM routines a little more.
6859 *) Add the following functions:
6861 ENGINE_load_cswift()
6863 ENGINE_load_atalla()
6865 ENGINE_load_builtin_engines()
6867 That way, an application can itself choose if external engines that
6868 are built-in in OpenSSL shall ever be used or not. The benefit is
6869 that applications won't have to be linked with libdl or other dso
6870 libraries unless it's really needed.
6872 Changed 'openssl engine' to load all engines on demand.
6873 Changed the engine header files to avoid the duplication of some
6874 declarations (they differed!).
6877 *) 'openssl engine' can now list capabilities.
6880 *) Better error reporting in 'openssl engine'.
6883 *) Never call load_dh_param(NULL) in s_server.
6886 *) Add engine application. It can currently list engines by name and
6887 identity, and test if they are actually available.
6890 *) Improve RPM specification file by forcing symbolic linking and making
6891 sure the installed documentation is also owned by root.root.
6892 [Damien Miller <djm@mindrot.org>]
6894 *) Give the OpenSSL applications more possibilities to make use of
6895 keys (public as well as private) handled by engines.
6898 *) Add OCSP code that comes from CertCo.
6901 *) Add VMS support for the Rijndael code.
6904 *) Added untested support for Nuron crypto accelerator.
6907 *) Add support for external cryptographic devices. This code was
6908 previously distributed separately as the "engine" branch.
6909 [Geoff Thorpe, Richard Levitte]
6911 *) Rework the filename-translation in the DSO code. It is now possible to
6912 have far greater control over how a "name" is turned into a filename
6913 depending on the operating environment and any oddities about the
6914 different shared library filenames on each system.
6917 *) Support threads on FreeBSD-elf in Configure.
6920 *) Fix for SHA1 assembly problem with MASM: it produces
6921 warnings about corrupt line number information when assembling
6922 with debugging information. This is caused by the overlapping
6924 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6927 NCONF_get_number() has no error checking at all. As a replacement,
6928 NCONF_get_number_e() is defined (_e for "error checking") and is
6929 promoted strongly. The old NCONF_get_number is kept around for
6930 binary backward compatibility.
6931 Make it possible for methods to load from something other than a BIO,
6932 by providing a function pointer that is given a name instead of a BIO.
6933 For example, this could be used to load configuration data from an
6937 *) Fix for non blocking accept BIOs. Added new I/O special reason
6938 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6939 with non blocking I/O was not possible because no retry code was
6940 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6944 *) Added the beginnings of Rijndael support.
6947 *) Fix for bug in DirectoryString mask setting. Add support for
6948 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6949 to allow certificate printing to more controllable, additional
6950 'certopt' option to 'x509' to allow new printing options to be
6954 *) Clean old EAY MD5 hack from e_os.h.
6957 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6959 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6960 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6961 [Joe Orton, Steve Henson]
6963 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6965 *) Fix additional bug revealed by the NISCC test suite:
6967 Stop bug triggering large recursion when presented with
6968 certain ASN.1 tags (CVE-2003-0851)
6971 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6973 *) Fix various bugs revealed by running the NISCC test suite:
6975 Stop out of bounds reads in the ASN1 code when presented with
6976 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6978 If verify callback ignores invalid public key errors don't try to check
6979 certificate signature with the NULL public key.
6983 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6984 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6988 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6989 extra data after the compression methods not only for TLS 1.0
6990 but also for SSL 3.0 (as required by the specification).
6991 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6993 *) Change X509_certificate_type() to mark the key as exported/exportable
6994 when it's 512 *bits* long, not 512 bytes.
6997 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6999 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7000 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7001 a protocol version number mismatch like a decryption error
7002 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7005 *) Turn on RSA blinding by default in the default implementation
7006 to avoid a timing attack. Applications that don't want it can call
7007 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7008 They would be ill-advised to do so in most cases.
7009 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7011 *) Change RSA blinding code so that it works when the PRNG is not
7012 seeded (in this case, the secret RSA exponent is abused as
7013 an unpredictable seed -- if it is not unpredictable, there
7014 is no point in blinding anyway). Make RSA blinding thread-safe
7015 by remembering the creator's thread ID in rsa->blinding and
7016 having all other threads use local one-time blinding factors
7017 (this requires more computation than sharing rsa->blinding, but
7018 avoids excessive locking; and if an RSA object is not shared
7019 between threads, blinding will still be very fast).
7022 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7024 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7025 via timing by performing a MAC computation even if incorrrect
7026 block cipher padding has been found. This is a countermeasure
7027 against active attacks where the attacker has to distinguish
7028 between bad padding and a MAC verification error. (CVE-2003-0078)
7030 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7031 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7032 Martin Vuagnoux (EPFL, Ilion)]
7034 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7036 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7037 memory from it's contents. This is done with a counter that will
7038 place alternating values in each byte. This can be used to solve
7039 two issues: 1) the removal of calls to memset() by highly optimizing
7040 compilers, and 2) cleansing with other values than 0, since those can
7041 be read through on certain media, for example a swap space on disk.
7044 *) Bugfix: client side session caching did not work with external caching,
7045 because the session->cipher setting was not restored when reloading
7046 from the external cache. This problem was masked, when
7047 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7048 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7051 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7052 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7053 [Zeev Lieber <zeev-l@yahoo.com>]
7055 *) Undo an undocumented change introduced in 0.9.6e which caused
7056 repeated calls to OpenSSL_add_all_ciphers() and
7057 OpenSSL_add_all_digests() to be ignored, even after calling
7061 *) Change the default configuration reader to deal with last line not
7062 being properly terminated.
7065 *) Change X509_NAME_cmp() so it applies the special rules on handling
7066 DN values that are of type PrintableString, as well as RDNs of type
7067 emailAddress where the value has the type ia5String.
7068 [stefank@valicert.com via Richard Levitte]
7070 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7071 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7072 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7073 the bitwise-OR of the two for use by the majority of applications
7074 wanting this behaviour, and update the docs. The documented
7075 behaviour and actual behaviour were inconsistent and had been
7076 changing anyway, so this is more a bug-fix than a behavioural
7078 [Geoff Thorpe, diagnosed by Nadav Har'El]
7080 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7081 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7084 *) Fix initialization code race conditions in
7085 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7086 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7087 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7088 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7089 ssl2_get_cipher_by_char(),
7090 ssl3_get_cipher_by_char().
7091 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7093 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7094 the cached sessions are flushed, as the remove_cb() might use ex_data
7095 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7096 (see [openssl.org #212]).
7097 [Geoff Thorpe, Lutz Jaenicke]
7099 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7100 length, instead of the encoding length to d2i_ASN1_OBJECT.
7103 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7105 *) [In 0.9.6g-engine release:]
7106 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7107 [Lynn Gazis <lgazis@rainbow.com>]
7109 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7111 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7112 and get fix the header length calculation.
7113 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7114 Alon Kantor <alonk@checkpoint.com> (and others),
7117 *) Use proper error handling instead of 'assertions' in buffer
7118 overflow checks added in 0.9.6e. This prevents DoS (the
7119 assertions could call abort()).
7120 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7122 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7124 *) Add various sanity checks to asn1_get_length() to reject
7125 the ASN1 length bytes if they exceed sizeof(long), will appear
7126 negative or the content length exceeds the length of the
7128 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7130 *) Fix cipher selection routines: ciphers without encryption had no flags
7131 for the cipher strength set and where therefore not handled correctly
7132 by the selection routines (PR #130).
7135 *) Fix EVP_dsa_sha macro.
7139 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7140 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7141 that was added in OpenSSL 0.9.6d.
7143 As the countermeasure turned out to be incompatible with some
7144 broken SSL implementations, the new option is part of SSL_OP_ALL.
7145 SSL_OP_ALL is usually employed when compatibility with weird SSL
7146 implementations is desired (e.g. '-bugs' option to 's_client' and
7147 's_server'), so the new option is automatically set in many
7151 *) Changes in security patch:
7153 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7154 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7155 Air Force Materiel Command, USAF, under agreement number
7158 *) Add various sanity checks to asn1_get_length() to reject
7159 the ASN1 length bytes if they exceed sizeof(long), will appear
7160 negative or the content length exceeds the length of the
7161 supplied buffer. (CVE-2002-0659)
7162 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7164 *) Assertions for various potential buffer overflows, not known to
7166 [Ben Laurie (CHATS)]
7168 *) Various temporary buffers to hold ASCII versions of integers were
7169 too small for 64 bit platforms. (CVE-2002-0655)
7170 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7172 *) Remote buffer overflow in SSL3 protocol - an attacker could
7173 supply an oversized session ID to a client. (CVE-2002-0656)
7174 [Ben Laurie (CHATS)]
7176 *) Remote buffer overflow in SSL2 protocol - an attacker could
7177 supply an oversized client master key. (CVE-2002-0656)
7178 [Ben Laurie (CHATS)]
7180 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7182 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7183 encoded as NULL) with id-dsa-with-sha1.
7184 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7186 *) Check various X509_...() return values in apps/req.c.
7187 [Nils Larsch <nla@trustcenter.de>]
7189 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7190 an end-of-file condition would erronously be flagged, when the CRLF
7191 was just at the end of a processed block. The bug was discovered when
7192 processing data through a buffering memory BIO handing the data to a
7193 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7194 <ptsekov@syntrex.com> and Nedelcho Stanev.
7197 *) Implement a countermeasure against a vulnerability recently found
7198 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7199 before application data chunks to avoid the use of known IVs
7200 with data potentially chosen by the attacker.
7203 *) Fix length checks in ssl3_get_client_hello().
7206 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7207 to prevent ssl3_read_internal() from incorrectly assuming that
7208 ssl3_read_bytes() found application data while handshake
7209 processing was enabled when in fact s->s3->in_read_app_data was
7210 merely automatically cleared during the initial handshake.
7211 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7213 *) Fix object definitions for Private and Enterprise: they were not
7214 recognized in their shortname (=lowercase) representation. Extend
7215 obj_dat.pl to issue an error when using undefined keywords instead
7216 of silently ignoring the problem (Svenning Sorensen
7217 <sss@sss.dnsalias.net>).
7220 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7221 generators, i.e. generators other than 2 and 5. (Previously, the
7222 code did not properly initialise the 'add' and 'rem' values to
7223 BN_generate_prime().)
7225 In the new general case, we do not insist that 'generator' is
7226 actually a primitive root: This requirement is rather pointless;
7227 a generator of the order-q subgroup is just as good, if not
7231 *) Map new X509 verification errors to alerts. Discovered and submitted by
7232 Tom Wu <tom@arcot.com>.
7235 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7236 returning non-zero before the data has been completely received
7237 when using non-blocking I/O.
7238 [Bodo Moeller; problem pointed out by John Hughes]
7240 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7241 [Ben Laurie, Lutz Jaenicke]
7243 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7244 Yoram Zahavi <YoramZ@gilian.com>).
7247 *) Add information about CygWin 1.3 and on, and preserve proper
7248 configuration for the versions before that.
7249 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7251 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7252 check whether we deal with a copy of a session and do not delete from
7253 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7254 <izhar@checkpoint.com>.
7257 *) Do not store session data into the internal session cache, if it
7258 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7259 flag is set). Proposed by Aslam <aslam@funk.com>.
7262 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7266 *) [In 0.9.6d-engine release:]
7267 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7268 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7270 *) Add the configuration target linux-s390x.
7271 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7273 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7274 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7275 variable as an indication that a ClientHello message has been
7276 received. As the flag value will be lost between multiple
7277 invocations of ssl3_accept when using non-blocking I/O, the
7278 function may not be aware that a handshake has actually taken
7279 place, thus preventing a new session from being added to the
7282 To avoid this problem, we now set s->new_session to 2 instead of
7283 using a local variable.
7284 [Lutz Jaenicke, Bodo Moeller]
7286 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7287 if the SSL_R_LENGTH_MISMATCH error is detected.
7288 [Geoff Thorpe, Bodo Moeller]
7290 *) New 'shared_ldflag' column in Configure platform table.
7293 *) Fix EVP_CIPHER_mode macro.
7294 ["Dan S. Camper" <dan@bti.net>]
7296 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7297 type, we must throw them away by setting rr->length to 0.
7298 [D P Chang <dpc@qualys.com>]
7300 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7302 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7303 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7304 worked incorrectly for those cases where range = 10..._2 and
7305 3*range is two bits longer than range.)
7308 *) Only add signing time to PKCS7 structures if it is not already
7312 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7313 OBJ_ld_ce should be OBJ_id_ce.
7314 Also some ip-pda OIDs in crypto/objects/objects.txt were
7315 incorrect (cf. RFC 3039).
7316 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7318 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7319 returns early because it has nothing to do.
7320 [Andy Schneider <andy.schneider@bjss.co.uk>]
7322 *) [In 0.9.6c-engine release:]
7323 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7324 [Andy Schneider <andy.schneider@bjss.co.uk>]
7326 *) [In 0.9.6c-engine release:]
7327 Add support for Cryptographic Appliance's keyserver technology.
7328 (Use engine 'keyclient')
7329 [Cryptographic Appliances and Geoff Thorpe]
7331 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7332 is called via tools/c89.sh because arguments have to be
7333 rearranged (all '-L' options must appear before the first object
7335 [Richard Shapiro <rshapiro@abinitio.com>]
7337 *) [In 0.9.6c-engine release:]
7338 Add support for Broadcom crypto accelerator cards, backported
7340 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7342 *) [In 0.9.6c-engine release:]
7343 Add support for SureWare crypto accelerator cards from
7344 Baltimore Technologies. (Use engine 'sureware')
7345 [Baltimore Technologies and Mark Cox]
7347 *) [In 0.9.6c-engine release:]
7348 Add support for crypto accelerator cards from Accelerated
7349 Encryption Processing, www.aep.ie. (Use engine 'aep')
7350 [AEP Inc. and Mark Cox]
7352 *) Add a configuration entry for gcc on UnixWare.
7353 [Gary Benson <gbenson@redhat.com>]
7355 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7356 messages are stored in a single piece (fixed-length part and
7357 variable-length part combined) and fix various bugs found on the way.
7360 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7361 instead. BIO_gethostbyname() does not know what timeouts are
7362 appropriate, so entries would stay in cache even when they have
7364 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7366 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7367 faced with a pathologically small ClientHello fragment that does
7368 not contain client_version: Instead of aborting with an error,
7369 simply choose the highest available protocol version (i.e.,
7370 TLS 1.0 unless it is disabled). In practice, ClientHello
7371 messages are never sent like this, but this change gives us
7372 strictly correct behaviour at least for TLS.
7375 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7376 never resets s->method to s->ctx->method when called from within
7377 one of the SSL handshake functions.
7378 [Bodo Moeller; problem pointed out by Niko Baric]
7380 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7381 (sent using the client's version number) if client_version is
7382 smaller than the protocol version in use. Also change
7383 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7384 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7385 the client will at least see that alert.
7388 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7392 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7393 client receives HelloRequest while in a handshake.
7394 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7396 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7397 should end in 'break', not 'goto end' which circuments various
7398 cleanups done in state SSL_ST_OK. But session related stuff
7399 must be disabled for SSL_ST_OK in the case that we just sent a
7402 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7403 before just sending a HelloRequest.
7404 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7406 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7407 reveal whether illegal block cipher padding was found or a MAC
7408 verification error occured. (Neither SSLerr() codes nor alerts
7409 are directly visible to potential attackers, but the information
7410 may leak via logfiles.)
7412 Similar changes are not required for the SSL 2.0 implementation
7413 because the number of padding bytes is sent in clear for SSL 2.0,
7414 and the extra bytes are just ignored. However ssl/s2_pkt.c
7415 failed to verify that the purported number of padding bytes is in
7419 *) Add OpenUNIX-8 support including shared libraries
7420 (Boyd Lynn Gerber <gerberb@zenez.com>).
7423 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7424 'wristwatch attack' using huge encoding parameters (cf.
7425 James H. Manger's CRYPTO 2001 paper). Note that the
7426 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7427 encoding parameters and hence was not vulnerable.
7430 *) BN_sqr() bug fix.
7431 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7433 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7434 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7435 followed by modular reduction.
7436 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7438 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7439 equivalent based on BN_pseudo_rand() instead of BN_rand().
7442 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7443 This function was broken, as the check for a new client hello message
7444 to handle SGC did not allow these large messages.
7445 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7448 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7451 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7452 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7455 *) Rework the configuration and shared library support for Tru64 Unix.
7456 The configuration part makes use of modern compiler features and
7457 still retains old compiler behavior for those that run older versions
7458 of the OS. The shared library support part includes a variant that
7459 uses the RPATH feature, and is available through the special
7460 configuration target "alpha-cc-rpath", which will never be selected
7462 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7464 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7465 with the same message size as in ssl3_get_certificate_request().
7466 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7467 messages might inadvertently be reject as too long.
7468 [Petr Lampa <lampa@fee.vutbr.cz>]
7470 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7473 *) Modified SSL library such that the verify_callback that has been set
7474 specificly for an SSL object with SSL_set_verify() is actually being
7475 used. Before the change, a verify_callback set with this function was
7476 ignored and the verify_callback() set in the SSL_CTX at the time of
7477 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7478 to allow the necessary settings.
7481 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7482 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7483 done automatically (in contradiction to the requirements of the C
7484 standard). This made problems when used from OpenSSH.
7487 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7488 dh->length and always used
7490 BN_rand_range(priv_key, dh->p).
7492 BN_rand_range() is not necessary for Diffie-Hellman, and this
7493 specific range makes Diffie-Hellman unnecessarily inefficient if
7494 dh->length (recommended exponent length) is much smaller than the
7495 length of dh->p. We could use BN_rand_range() if the order of
7496 the subgroup was stored in the DH structure, but we only have
7501 BN_rand(priv_key, l, ...)
7503 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7509 RSA_eay_public_encrypt
7510 RSA_eay_private_decrypt
7511 RSA_eay_private_encrypt (signing)
7512 RSA_eay_public_decrypt (signature verification)
7514 (default implementations for RSA_public_encrypt,
7515 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7516 always reject numbers >= n.
7519 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7520 to synchronize access to 'locking_thread'. This is necessary on
7521 systems where access to 'locking_thread' (an 'unsigned long'
7522 variable) is not atomic.
7525 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7526 *before* setting the 'crypto_lock_rand' flag. The previous code had
7527 a race condition if 0 is a valid thread ID.
7528 [Travis Vitek <vitek@roguewave.com>]
7530 *) Add support for shared libraries under Irix.
7531 [Albert Chin-A-Young <china@thewrittenword.com>]
7533 *) Add configuration option to build on Linux on both big-endian and
7535 [Ralf Baechle <ralf@uni-koblenz.de>]
7537 *) Add the possibility to create shared libraries on HP-UX.
7540 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7542 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7543 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7544 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7545 PRNG state recovery was possible based on the output of
7546 one PRNG request appropriately sized to gain knowledge on
7547 'md' followed by enough consecutive 1-byte PRNG requests
7548 to traverse all of 'state'.
7550 1. When updating 'md_local' (the current thread's copy of 'md')
7551 during PRNG output generation, hash all of the previous
7552 'md_local' value, not just the half used for PRNG output.
7554 2. Make the number of bytes from 'state' included into the hash
7555 independent from the number of PRNG bytes requested.
7557 The first measure alone would be sufficient to avoid
7558 Markku-Juhani's attack. (Actually it had never occurred
7559 to me that the half of 'md_local' used for chaining was the
7560 half from which PRNG output bytes were taken -- I had always
7561 assumed that the secret half would be used.) The second
7562 measure makes sure that additional data from 'state' is never
7563 mixed into 'md_local' in small portions; this heuristically
7564 further strengthens the PRNG.
7567 *) Fix crypto/bn/asm/mips3.s.
7570 *) When only the key is given to "enc", the IV is undefined. Print out
7571 an error message in this case.
7574 *) Handle special case when X509_NAME is empty in X509 printing routines.
7577 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7578 positive and less than q.
7581 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7582 used: it isn't thread safe and the add_lock_callback should handle
7584 [Paul Rose <Paul.Rose@bridge.com>]
7586 *) Verify that incoming data obeys the block size in
7587 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7591 [Ulf Möller, Bodo Möller]
7593 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7594 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7595 when fixing the server behaviour for backwards-compatible 'client
7596 hello' messages. (Note that the attack is impractical against
7597 SSL 3.0 and TLS 1.0 anyway because length and version checking
7598 means that the probability of guessing a valid ciphertext is
7599 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7602 Before 0.9.5, the countermeasure (hide the error by generating a
7603 random 'decryption result') did not work properly because
7604 ERR_clear_error() was missing, meaning that SSL_get_error() would
7605 detect the supposedly ignored error.
7607 Both problems are now fixed.
7610 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7611 (previously it was 1024).
7614 *) Fix for compatibility mode trust settings: ignore trust settings
7615 unless some valid trust or reject settings are present.
7618 *) Fix for blowfish EVP: its a variable length cipher.
7621 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7622 parameters in DSA public key structures and return an error in the
7623 DSA routines if parameters are absent.
7626 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7627 in the current directory if neither $RANDFILE nor $HOME was set.
7628 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7629 caused some confusion to Windows users who haven't defined $HOME.
7630 Thus RAND_file_name() is changed again: e_os.h can define a
7631 DEFAULT_HOME, which will be used if $HOME is not set.
7632 For Windows, we use "C:"; on other platforms, we still require
7633 environment variables.
7635 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7636 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7637 having multiple threads call RAND_poll() concurrently.
7640 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7641 combination of a flag and a thread ID variable.
7642 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7643 flag), *other* threads can enter ssleay_add_bytes without obeying
7644 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7645 that they do not hold after the first thread unsets add_do_not_lock).
7648 *) Change bctest again: '-x' expressions are not available in all
7652 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7654 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7655 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7657 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7658 the default extension for executables, if any. Also, make the perl
7659 scripts that use symlink() to test if it really exists and use "cp"
7660 if it doesn't. All this made OpenSSL compilable and installable in
7664 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7665 If SEQUENCE is length is indefinite just set c->slen to the total
7666 amount of data available.
7667 [Steve Henson, reported by shige@FreeBSD.org]
7668 [This change does not apply to 0.9.7.]
7670 *) Change bctest to avoid here-documents inside command substitution
7671 (workaround for FreeBSD /bin/sh bug).
7672 For compatibility with Ultrix, avoid shell functions (introduced
7673 in the bctest version that searches along $PATH).
7676 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7677 with des_encrypt() defined on some operating systems, like Solaris
7681 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7682 On the Importance of Eliminating Errors in Cryptographic
7683 Computations, J. Cryptology 14 (2001) 2, 101-119,
7684 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7687 *) MIPS assembler BIGNUM division bug fix.
7690 *) Disabled incorrect Alpha assembler code.
7693 *) Fix PKCS#7 decode routines so they correctly update the length
7694 after reading an EOC for the EXPLICIT tag.
7696 [This change does not apply to 0.9.7.]
7698 *) Fix bug in PKCS#12 key generation routines. This was triggered
7699 if a 3DES key was generated with a 0 initial byte. Include
7700 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7701 (but broken) behaviour.
7704 *) Enhance bctest to search for a working bc along $PATH and print
7706 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7708 *) Fix memory leaks in err.c: free err_data string if necessary;
7709 don't write to the wrong index in ERR_set_error_data.
7712 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7716 *) Replace rdtsc with _emit statements for VC++ version 5.
7717 [Jeremy Cooper <jeremy@baymoo.org>]
7719 *) Make it possible to reuse SSLv2 sessions.
7722 *) In copy_email() check for >= 0 as a return value for
7723 X509_NAME_get_index_by_NID() since 0 is a valid index.
7724 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7726 *) Avoid coredump with unsupported or invalid public keys by checking if
7727 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7728 PKCS7_verify() fails with non detached data.
7731 *) Don't use getenv in library functions when run as setuid/setgid.
7732 New function OPENSSL_issetugid().
7735 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7736 due to incorrect handling of multi-threading:
7738 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7740 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7742 3. Count how many times MemCheck_off() has been called so that
7743 nested use can be treated correctly. This also avoids
7744 inband-signalling in the previous code (which relied on the
7745 assumption that thread ID 0 is impossible).
7748 *) Add "-rand" option also to s_client and s_server.
7751 *) Fix CPU detection on Irix 6.x.
7752 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7753 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7755 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7758 [This change does not apply to 0.9.7.]
7760 *) Use the cached encoding of an X509_NAME structure rather than
7761 copying it. This is apparently the reason for the libsafe "errors"
7762 but the code is actually correct.
7765 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7766 Bleichenbacher's DSA attack.
7767 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7768 to be set and top=0 forces the highest bit to be set; top=-1 is new
7769 and leaves the highest bit random.
7770 [Ulf Moeller, Bodo Moeller]
7772 *) In the NCONF_...-based implementations for CONF_... queries
7773 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7774 a temporary CONF structure with the data component set to NULL
7775 (which gives segmentation faults in lh_retrieve).
7776 Instead, use NULL for the CONF pointer in CONF_get_string and
7777 CONF_get_number (which may use environment variables) and directly
7778 return NULL from CONF_get_section.
7781 *) Fix potential buffer overrun for EBCDIC.
7784 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7785 keyUsage if basicConstraints absent for a CA.
7788 *) Make SMIME_write_PKCS7() write mail header values with a format that
7789 is more generally accepted (no spaces before the semicolon), since
7790 some programs can't parse those values properly otherwise. Also make
7791 sure BIO's that break lines after each write do not create invalid
7795 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7796 macros previously used would not encode an empty SEQUENCE OF
7797 and break the signature.
7799 [This change does not apply to 0.9.7.]
7801 *) Zero the premaster secret after deriving the master secret in
7805 *) Add some EVP_add_digest_alias registrations (as found in
7806 OpenSSL_add_all_digests()) to SSL_library_init()
7807 aka OpenSSL_add_ssl_algorithms(). This provides improved
7808 compatibility with peers using X.509 certificates
7809 with unconventional AlgorithmIdentifier OIDs.
7812 *) Fix for Irix with NO_ASM.
7813 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7815 *) ./config script fixes.
7816 [Ulf Moeller, Richard Levitte]
7818 *) Fix 'openssl passwd -1'.
7821 *) Change PKCS12_key_gen_asc() so it can cope with non null
7822 terminated strings whose length is passed in the passlen
7823 parameter, for example from PEM callbacks. This was done
7824 by adding an extra length parameter to asc2uni().
7825 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7827 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7828 call failed, free the DSA structure.
7831 *) Fix to uni2asc() to cope with zero length Unicode strings.
7832 These are present in some PKCS#12 files.
7835 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7836 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7837 when writing a 32767 byte record.
7838 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7840 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7841 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7843 (RSA objects have a reference count access to which is protected
7844 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7845 so they are meant to be shared between threads.)
7846 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7847 "Reddie, Steven" <Steven.Reddie@ca.com>]
7849 *) Fix a deadlock in CRYPTO_mem_leaks().
7852 *) Use better test patterns in bntest.
7855 *) rand_win.c fix for Borland C.
7858 *) BN_rshift bugfix for n == 0.
7861 *) Add a 'bctest' script that checks for some known 'bc' bugs
7862 so that 'make test' does not abort just because 'bc' is broken.
7865 *) Store verify_result within SSL_SESSION also for client side to
7866 avoid potential security hole. (Re-used sessions on the client side
7867 always resulted in verify_result==X509_V_OK, not using the original
7868 result of the server certificate verification.)
7871 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7872 SSL3_RT_APPLICATION_DATA, return 0.
7873 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7877 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7878 releases, have been re-implemented by renaming the previous
7879 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7880 and ssl3_read_internal, respectively, and adding 'peek' parameters
7881 to them. The new ssl[23]_{read,peek} functions are calls to
7882 ssl[23]_read_internal with the 'peek' flag set appropriately.
7883 A 'peek' parameter has also been added to ssl3_read_bytes, which
7884 does the actual work for ssl3_read_internal.
7887 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7888 the method-specific "init()" handler. Also clean up ex_data after
7889 calling the method-specific "finish()" handler. Previously, this was
7890 happening the other way round.
7893 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7894 The previous value, 12, was not always sufficient for BN_mod_exp().
7897 *) Make sure that shared libraries get the internal name engine with
7898 the full version number and not just 0. This should mark the
7899 shared libraries as not backward compatible. Of course, this should
7900 be changed again when we can guarantee backward binary compatibility.
7903 *) Fix typo in get_cert_by_subject() in by_dir.c
7904 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7906 *) Rework the system to generate shared libraries:
7908 - Make note of the expected extension for the shared libraries and
7909 if there is a need for symbolic links from for example libcrypto.so.0
7910 to libcrypto.so.0.9.7. There is extended info in Configure for
7913 - Make as few rebuilds of the shared libraries as possible.
7915 - Still avoid linking the OpenSSL programs with the shared libraries.
7917 - When installing, install the shared libraries separately from the
7921 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7923 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7924 and not in SSL_clear because the latter is also used by the
7925 accept/connect functions; previously, the settings made by
7926 SSL_set_read_ahead would be lost during the handshake.
7927 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7929 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7930 Previously, it would create entries for disableed algorithms no
7934 *) Added several new manual pages for SSL_* function.
7937 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7939 *) In ssl23_get_client_hello, generate an error message when faced
7940 with an initial SSL 3.0/TLS record that is too small to contain the
7941 first two bytes of the ClientHello message, i.e. client_version.
7942 (Note that this is a pathologic case that probably has never happened
7943 in real life.) The previous approach was to use the version number
7944 from the record header as a substitute; but our protocol choice
7945 should not depend on that one because it is not authenticated
7946 by the Finished messages.
7949 *) More robust randomness gathering functions for Windows.
7950 [Jeffrey Altman <jaltman@columbia.edu>]
7952 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7953 not set then we don't setup the error code for issuer check errors
7954 to avoid possibly overwriting other errors which the callback does
7955 handle. If an application does set the flag then we assume it knows
7956 what it is doing and can handle the new informational codes
7960 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7961 a general "ANY" type, as such it should be able to decode anything
7962 including tagged types. However it didn't check the class so it would
7963 wrongly interpret tagged types in the same way as their universal
7964 counterpart and unknown types were just rejected. Changed so that the
7965 tagged and unknown types are handled in the same way as a SEQUENCE:
7966 that is the encoding is stored intact. There is also a new type
7967 "V_ASN1_OTHER" which is used when the class is not universal, in this
7968 case we have no idea what the actual type is so we just lump them all
7972 *) On VMS, stdout may very well lead to a file that is written to
7973 in a record-oriented fashion. That means that every write() will
7974 write a separate record, which will be read separately by the
7975 programs trying to read from it. This can be very confusing.
7977 The solution is to put a BIO filter in the way that will buffer
7978 text until a linefeed is reached, and then write everything a
7979 line at a time, so every record written will be an actual line,
7980 not chunks of lines and not (usually doesn't happen, but I've
7981 seen it once) several lines in one record. BIO_f_linebuffer() is
7984 Currently, it's a VMS-only method, because that's where it has
7985 been tested well enough.
7988 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7989 it can return incorrect results.
7990 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7991 but it was in 0.9.6-beta[12].)
7994 *) Disable the check for content being present when verifying detached
7995 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7996 include zero length content when signing messages.
7999 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8000 BIO_ctrl (for BIO pairs).
8003 *) Add DSO method for VMS.
8006 *) Bug fix: Montgomery multiplication could produce results with the
8010 *) Add RPM specification openssl.spec and modify it to build three
8011 packages. The default package contains applications, application
8012 documentation and run-time libraries. The devel package contains
8013 include files, static libraries and function documentation. The
8014 doc package contains the contents of the doc directory. The original
8015 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8018 *) Add a large number of documentation files for many SSL routines.
8019 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8021 *) Add a configuration entry for Sony News 4.
8022 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8024 *) Don't set the two most significant bits to one when generating a
8025 random number < q in the DSA library.
8028 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8029 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8030 the underlying transport is blocking) if a handshake took place.
8031 (The default behaviour is needed by applications such as s_client
8032 and s_server that use select() to determine when to use SSL_read;
8033 but for applications that know in advance when to expect data, it
8034 just makes things more complicated.)
8037 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8041 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8042 work better on such systems.
8043 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8045 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8046 Update PKCS12_parse() so it copies the friendlyName and the
8047 keyid to the certificates aux info.
8050 *) Fix bug in PKCS7_verify() which caused an infinite loop
8051 if there was more than one signature.
8052 [Sven Uszpelkat <su@celocom.de>]
8054 *) Major change in util/mkdef.pl to include extra information
8055 about each symbol, as well as presentig variables as well
8056 as functions. This change means that there's n more need
8057 to rebuild the .num files when some algorithms are excluded.
8060 *) Allow the verify time to be set by an application,
8061 rather than always using the current time.
8064 *) Phase 2 verify code reorganisation. The certificate
8065 verify code now looks up an issuer certificate by a
8066 number of criteria: subject name, authority key id
8067 and key usage. It also verifies self signed certificates
8068 by the same criteria. The main comparison function is
8069 X509_check_issued() which performs these checks.
8071 Lot of changes were necessary in order to support this
8072 without completely rewriting the lookup code.
8074 Authority and subject key identifier are now cached.
8076 The LHASH 'certs' is X509_STORE has now been replaced
8077 by a STACK_OF(X509_OBJECT). This is mainly because an
8078 LHASH can't store or retrieve multiple objects with
8079 the same hash value.
8081 As a result various functions (which were all internal
8082 use only) have changed to handle the new X509_STORE
8083 structure. This will break anything that messed round
8084 with X509_STORE internally.
8086 The functions X509_STORE_add_cert() now checks for an
8087 exact match, rather than just subject name.
8089 The X509_STORE API doesn't directly support the retrieval
8090 of multiple certificates matching a given criteria, however
8091 this can be worked round by performing a lookup first
8092 (which will fill the cache with candidate certificates)
8093 and then examining the cache for matches. This is probably
8094 the best we can do without throwing out X509_LOOKUP
8095 entirely (maybe later...).
8097 The X509_VERIFY_CTX structure has been enhanced considerably.
8099 All certificate lookup operations now go via a get_issuer()
8100 callback. Although this currently uses an X509_STORE it
8101 can be replaced by custom lookups. This is a simple way
8102 to bypass the X509_STORE hackery necessary to make this
8103 work and makes it possible to use more efficient techniques
8104 in future. A very simple version which uses a simple
8105 STACK for its trusted certificate store is also provided
8106 using X509_STORE_CTX_trusted_stack().
8108 The verify_cb() and verify() callbacks now have equivalents
8109 in the X509_STORE_CTX structure.
8111 X509_STORE_CTX also has a 'flags' field which can be used
8112 to customise the verify behaviour.
8115 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8116 excludes S/MIME capabilities.
8119 *) When a certificate request is read in keep a copy of the
8120 original encoding of the signed data and use it when outputing
8121 again. Signatures then use the original encoding rather than
8122 a decoded, encoded version which may cause problems if the
8123 request is improperly encoded.
8126 *) For consistency with other BIO_puts implementations, call
8127 buffer_write(b, ...) directly in buffer_puts instead of calling
8130 In BIO_puts, increment b->num_write as in BIO_write.
8131 [Peter.Sylvester@EdelWeb.fr]
8133 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8134 BN_zero, we may not return a BIGNUM with an array consisting of
8138 *) Avoid calling abort() from within the library when problems are
8139 detected, except if preprocessor symbols have been defined
8140 (such as REF_CHECK, BN_DEBUG etc.).
8143 *) New openssl application 'rsautl'. This utility can be
8144 used for low level RSA operations. DER public key
8145 BIO/fp routines also added.
8148 *) New Configure entry and patches for compiling on QNX 4.
8149 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8151 *) A demo state-machine implementation was sponsored by
8152 Nuron (http://www.nuron.com/) and is now available in
8153 demos/state_machine.
8156 *) New options added to the 'dgst' utility for signature
8157 generation and verification.
8160 *) Unrecognized PKCS#7 content types are now handled via a
8161 catch all ASN1_TYPE structure. This allows unsupported
8162 types to be stored as a "blob" and an application can
8163 encode and decode it manually.
8166 *) Fix various signed/unsigned issues to make a_strex.c
8168 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8170 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8171 length if passed a buffer. ASN1_INTEGER_to_BN failed
8172 if passed a NULL BN and its argument was negative.
8173 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8175 *) Modification to PKCS#7 encoding routines to output definite
8176 length encoding. Since currently the whole structures are in
8177 memory there's not real point in using indefinite length
8178 constructed encoding. However if OpenSSL is compiled with
8179 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8182 *) Added BIO_vprintf() and BIO_vsnprintf().
8185 *) Added more prefixes to parse for in the the strings written
8186 through a logging bio, to cover all the levels that are available
8187 through syslog. The prefixes are now:
8189 PANIC, EMERG, EMR => LOG_EMERG
8190 ALERT, ALR => LOG_ALERT
8191 CRIT, CRI => LOG_CRIT
8192 ERROR, ERR => LOG_ERR
8193 WARNING, WARN, WAR => LOG_WARNING
8194 NOTICE, NOTE, NOT => LOG_NOTICE
8195 INFO, INF => LOG_INFO
8196 DEBUG, DBG => LOG_DEBUG
8198 and as before, if none of those prefixes are present at the
8199 beginning of the string, LOG_ERR is chosen.
8201 On Win32, the LOG_* levels are mapped according to this:
8203 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8204 LOG_WARNING => EVENTLOG_WARNING_TYPE
8205 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8209 *) Made it possible to reconfigure with just the configuration
8210 argument "reconf" or "reconfigure". The command line arguments
8211 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8212 and are retrieved from there when reconfiguring.
8216 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8218 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8221 *) The obj_dat.pl script was messing up the sorting of object
8222 names. The reason was that it compared the quoted version
8223 of strings as a result "OCSP" > "OCSP Signing" because
8224 " > SPACE. Changed script to store unquoted versions of
8225 names and add quotes on output. It was also omitting some
8226 names from the lookup table if they were given a default
8227 value (that is if SN is missing it is given the same
8228 value as LN and vice versa), these are now added on the
8229 grounds that if an object has a name we should be able to
8230 look it up. Finally added warning output when duplicate
8231 short or long names are found.
8234 *) Changes needed for Tandem NSK.
8235 [Scott Uroff <scott@xypro.com>]
8237 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8238 RSA_padding_check_SSLv23(), special padding was never detected
8239 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8240 version rollback attacks was not effective.
8242 In s23_clnt.c, don't use special rollback-attack detection padding
8243 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8244 client; similarly, in s23_srvr.c, don't do the rollback check if
8245 SSL 2.0 is the only protocol enabled in the server.
8248 *) Make it possible to get hexdumps of unprintable data with 'openssl
8249 asn1parse'. By implication, the functions ASN1_parse_dump() and
8250 BIO_dump_indent() are added.
8253 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8254 these print out strings and name structures based on various
8255 flags including RFC2253 support and proper handling of
8256 multibyte characters. Added options to the 'x509' utility
8257 to allow the various flags to be set.
8260 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8261 Also change the functions X509_cmp_current_time() and
8262 X509_gmtime_adj() work with an ASN1_TIME structure,
8263 this will enable certificates using GeneralizedTime in validity
8264 dates to be checked.
8267 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8268 negative public key encodings) on by default,
8269 NO_NEG_PUBKEY_BUG can be set to disable it.
8272 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8273 content octets. An i2c_ASN1_OBJECT is unnecessary because
8274 the encoding can be trivially obtained from the structure.
8277 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8278 not read locks (CRYPTO_r_[un]lock).
8281 *) A first attempt at creating official support for shared
8282 libraries through configuration. I've kept it so the
8283 default is static libraries only, and the OpenSSL programs
8284 are always statically linked for now, but there are
8285 preparations for dynamic linking in place.
8286 This has been tested on Linux and Tru64.
8289 *) Randomness polling function for Win9x, as described in:
8290 Peter Gutmann, Software Generation of Practically Strong
8294 *) Fix so PRNG is seeded in req if using an already existing
8298 *) New options to smime application. -inform and -outform
8299 allow alternative formats for the S/MIME message including
8300 PEM and DER. The -content option allows the content to be
8301 specified separately. This should allow things like Netscape
8302 form signing output easier to verify.
8305 *) Fix the ASN1 encoding of tags using the 'long form'.
8308 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8309 STRING types. These convert content octets to and from the
8310 underlying type. The actual tag and length octets are
8311 already assumed to have been read in and checked. These
8312 are needed because all other string types have virtually
8313 identical handling apart from the tag. By having versions
8314 of the ASN1 functions that just operate on content octets
8315 IMPLICIT tagging can be handled properly. It also allows
8316 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8317 and ASN1_INTEGER are identical apart from the tag.
8320 *) Change the handling of OID objects as follows:
8322 - New object identifiers are inserted in objects.txt, following
8323 the syntax given in objects.README.
8324 - objects.pl is used to process obj_mac.num and create a new
8326 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8329 This is currently kind of a hack, and the perl code in objects.pl
8330 isn't very elegant, but it works as I intended. The simplest way
8331 to check that it worked correctly is to look in obj_dat.h and
8332 check the array nid_objs and make sure the objects haven't moved
8333 around (this is important!). Additions are OK, as well as
8334 consistent name changes.
8337 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8340 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8341 The given file adds to whatever has already been seeded into the
8342 random pool through the RANDFILE configuration file option or
8343 environment variable, or the default random state file.
8346 *) mkstack.pl now sorts each macro group into lexical order.
8347 Previously the output order depended on the order the files
8348 appeared in the directory, resulting in needless rewriting
8352 *) Patches to make OpenSSL compile under Win32 again. Mostly
8353 work arounds for the VC++ problem that it treats func() as
8354 func(void). Also stripped out the parts of mkdef.pl that
8355 added extra typesafe functions: these no longer exist.
8358 *) Reorganisation of the stack code. The macros are now all
8359 collected in safestack.h . Each macro is defined in terms of
8360 a "stack macro" of the form SKM_<name>(type, a, b). The
8361 DEBUG_SAFESTACK is now handled in terms of function casts,
8362 this has the advantage of retaining type safety without the
8363 use of additional functions. If DEBUG_SAFESTACK is not defined
8364 then the non typesafe macros are used instead. Also modified the
8365 mkstack.pl script to handle the new form. Needs testing to see
8366 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8367 the default if no major problems. Similar behaviour for ASN1_SET_OF
8368 and PKCS12_STACK_OF.
8371 *) When some versions of IIS use the 'NET' form of private key the
8372 key derivation algorithm is different. Normally MD5(password) is
8373 used as a 128 bit RC4 key. In the modified case
8374 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
8375 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8376 as the old Netscape_RSA functions except they have an additional
8377 'sgckey' parameter which uses the modified algorithm. Also added
8378 an -sgckey command line option to the rsa utility. Thanks to
8379 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8380 algorithm to openssl-dev.
8383 *) The evp_local.h macros were using 'c.##kname' which resulted in
8384 invalid expansion on some systems (SCO 5.0.5 for example).
8385 Corrected to 'c.kname'.
8386 [Phillip Porch <root@theporch.com>]
8388 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8389 a STACK of email addresses from a certificate or request, these look
8390 in the subject name and the subject alternative name extensions and
8391 omit any duplicate addresses.
8394 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8395 This makes DSA verification about 2 % faster.
8398 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8399 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8400 plus overhead for 1024 bit moduli).
8401 This makes exponentiations about 0.5 % faster for 1024 bit
8402 exponents (as measured by "openssl speed rsa2048").
8405 *) Rename memory handling macros to avoid conflicts with other
8407 Malloc => OPENSSL_malloc
8408 Malloc_locked => OPENSSL_malloc_locked
8409 Realloc => OPENSSL_realloc
8410 Free => OPENSSL_free
8413 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8414 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8417 *) CygWin32 support.
8418 [John Jarvie <jjarvie@newsguy.com>]
8420 *) The type-safe stack code has been rejigged. It is now only compiled
8421 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8422 by default all type-specific stack functions are "#define"d back to
8423 standard stack functions. This results in more streamlined output
8424 but retains the type-safety checking possibilities of the original
8428 *) The STACK code has been cleaned up, and certain type declarations
8429 that didn't make a lot of sense have been brought in line. This has
8430 also involved a cleanup of sorts in safestack.h to more correctly
8431 map type-safe stack functions onto their plain stack counterparts.
8432 This work has also resulted in a variety of "const"ifications of
8433 lots of the code, especially "_cmp" operations which should normally
8434 be prototyped with "const" parameters anyway.
8437 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8438 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8439 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8440 where all of 'md' is used each time the PRNG is used, but 'state'
8441 is used only indexed by a cyclic counter. As entropy may not be
8442 well distributed from the beginning, 'md' is important as a
8443 chaining variable. However, the output function chains only half
8444 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8445 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8446 in all of 'state' being rewritten, with the new values depending
8447 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8450 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8451 the handshake is continued after ssl_verify_cert_chain();
8452 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8453 can lead to 'unexplainable' connection aborts later.
8454 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8456 *) Major EVP API cipher revision.
8457 Add hooks for extra EVP features. This allows various cipher
8458 parameters to be set in the EVP interface. Support added for variable
8459 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8460 setting of RC2 and RC5 parameters.
8462 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8465 Remove lots of duplicated code from the EVP library. For example *every*
8466 cipher init() function handles the 'iv' in the same way according to the
8467 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8468 for CFB and OFB modes they zero ctx->num.
8470 New functionality allows removal of S/MIME code RC2 hack.
8472 Most of the routines have the same form and so can be declared in terms
8475 By shifting this to the top level EVP_CipherInit() it can be removed from
8476 all individual ciphers. If the cipher wants to handle IVs or keys
8477 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8480 Change lots of functions like EVP_EncryptUpdate() to now return a
8481 value: although software versions of the algorithms cannot fail
8482 any installed hardware versions can.
8485 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8486 this option is set, tolerate broken clients that send the negotiated
8487 protocol version number instead of the requested protocol version
8491 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8492 i.e. non-zero for export ciphersuites, zero otherwise.
8493 Previous versions had this flag inverted, inconsistent with
8494 rsa_tmp_cb (..._TMP_RSA_CB).
8495 [Bodo Moeller; problem reported by Amit Chopra]
8497 *) Add missing DSA library text string. Work around for some IIS
8498 key files with invalid SEQUENCE encoding.
8501 *) Add a document (doc/standards.txt) that list all kinds of standards
8502 and so on that are implemented in OpenSSL.
8505 *) Enhance c_rehash script. Old version would mishandle certificates
8506 with the same subject name hash and wouldn't handle CRLs at all.
8507 Added -fingerprint option to crl utility, to support new c_rehash
8511 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8514 *) Fix for SSL server purpose checking. Server checking was
8515 rejecting certificates which had extended key usage present
8516 but no ssl client purpose.
8517 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8519 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8520 is a little unclear about how a blank password is handled.
8521 Since the password in encoded as a BMPString with terminating
8522 double NULL a zero length password would end up as just the
8523 double NULL. However no password at all is different and is
8524 handled differently in the PKCS#12 key generation code. NS
8525 treats a blank password as zero length. MSIE treats it as no
8526 password on export: but it will try both on import. We now do
8527 the same: PKCS12_parse() tries zero length and no password if
8528 the password is set to "" or NULL (NULL is now a valid password:
8529 it wasn't before) as does the pkcs12 application.
8532 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8533 perror when PEM_read_bio_X509_REQ fails, the error message must
8534 be obtained from the error queue.
8537 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8538 it in ERR_remove_state if appropriate, and change ERR_get_state
8539 accordingly to avoid race conditions (this is necessary because
8540 thread_hash is no longer constant once set).
8543 *) Bugfix for linux-elf makefile.one.
8546 *) RSA_get_default_method() will now cause a default
8547 RSA_METHOD to be chosen if one doesn't exist already.
8548 Previously this was only set during a call to RSA_new()
8549 or RSA_new_method(NULL) meaning it was possible for
8550 RSA_get_default_method() to return NULL.
8553 *) Added native name translation to the existing DSO code
8554 that will convert (if the flag to do so is set) filenames
8555 that are sufficiently small and have no path information
8556 into a canonical native form. Eg. "blah" converted to
8557 "libblah.so" or "blah.dll" etc.
8560 *) New function ERR_error_string_n(e, buf, len) which is like
8561 ERR_error_string(e, buf), but writes at most 'len' bytes
8562 including the 0 terminator. For ERR_error_string_n, 'buf'
8564 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8566 *) CONF library reworked to become more general. A new CONF
8567 configuration file reader "class" is implemented as well as a
8568 new functions (NCONF_*, for "New CONF") to handle it. The now
8569 old CONF_* functions are still there, but are reimplemented to
8570 work in terms of the new functions. Also, a set of functions
8571 to handle the internal storage of the configuration data is
8572 provided to make it easier to write new configuration file
8573 reader "classes" (I can definitely see something reading a
8574 configuration file in XML format, for example), called _CONF_*,
8575 or "the configuration storage API"...
8577 The new configuration file reading functions are:
8579 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8580 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8582 NCONF_default, NCONF_WIN32
8584 NCONF_dump_fp, NCONF_dump_bio
8586 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8587 NCONF_new creates a new CONF object. This works in the same way
8588 as other interfaces in OpenSSL, like the BIO interface.
8589 NCONF_dump_* dump the internal storage of the configuration file,
8590 which is useful for debugging. All other functions take the same
8591 arguments as the old CONF_* functions wth the exception of the
8592 first that must be a `CONF *' instead of a `LHASH *'.
8594 To make it easer to use the new classes with the old CONF_* functions,
8595 the function CONF_set_default_method is provided.
8598 *) Add '-tls1' option to 'openssl ciphers', which was already
8599 mentioned in the documentation but had not been implemented.
8600 (This option is not yet really useful because even the additional
8601 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8604 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8605 OpenSSL-based applications) load shared libraries and bind to
8606 them in a portable way.
8607 [Geoff Thorpe, with contributions from Richard Levitte]
8609 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8611 *) Make sure _lrotl and _lrotr are only used with MSVC.
8613 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8614 (the default implementation of RAND_status).
8616 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8617 to '-clrext' (= clear extensions), as intended and documented.
8618 [Bodo Moeller; inconsistency pointed out by Michael Attili
8621 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8622 was larger than the MD block size.
8623 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8625 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8626 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8627 using the passed key: if the passed key was a private key the result
8628 of X509_print(), for example, would be to print out all the private key
8632 *) des_quad_cksum() byte order bug fix.
8633 [Ulf Möller, using the problem description in krb4-0.9.7, where
8634 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8636 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8638 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8640 *) For easily testing in shell scripts whether some command
8641 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8642 returns with exit code 0 iff no command of the given name is available.
8643 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8644 the output goes to stdout and nothing is printed to stderr.
8645 Additional arguments are always ignored.
8647 Since for each cipher there is a command of the same name,
8648 the 'no-cipher' compilation switches can be tested this way.
8650 ('openssl no-XXX' is not able to detect pseudo-commands such
8651 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8654 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8657 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8658 is set; it will be thrown away anyway because each handshake creates
8660 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8661 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8662 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
8663 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8666 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8667 'Q' and 'R' lose their special meanings (quit/renegotiate).
8668 This is part of what -quiet does; unlike -quiet, -ign_eof
8669 does not suppress any output.
8672 *) Add compatibility options to the purpose and trust code. The
8673 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8674 accepts a certificate or CA, this was the previous behaviour,
8675 with all the associated security issues.
8677 X509_TRUST_COMPAT is the old trust behaviour: only and
8678 automatically trust self signed roots in certificate store. A
8679 new trust setting X509_TRUST_DEFAULT is used to specify that
8680 a purpose has no associated trust setting and it should instead
8681 use the value in the default purpose.
8684 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8685 and fix a memory leak.
8688 *) In util/mkerr.pl (which implements 'make errors'), preserve
8689 reason strings from the previous version of the .c file, as
8690 the default to have only downcase letters (and digits) in
8691 automatically generated reasons codes is not always appropriate.
8694 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8695 using strerror. Previously, ERR_reason_error_string() returned
8696 library names as reason strings for SYSerr; but SYSerr is a special
8697 case where small numbers are errno values, not library numbers.
8700 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8701 converts DSA parameters into DH parameters. (When creating parameters,
8702 DSA_generate_parameters is used.)
8705 *) Include 'length' (recommended exponent length) in C code generated
8706 by 'openssl dhparam -C'.
8709 *) The second argument to set_label in perlasm was already being used
8710 so couldn't be used as a "file scope" flag. Moved to third argument
8714 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8715 instead of RAND_bytes for encryption IVs and salts.
8718 *) Include RAND_status() into RAND_METHOD instead of implementing
8719 it only for md_rand.c Otherwise replacing the PRNG by calling
8720 RAND_set_rand_method would be impossible.
8723 *) Don't let DSA_generate_key() enter an infinite loop if the random
8724 number generation fails.
8727 *) New 'rand' application for creating pseudo-random output.
8730 *) Added configuration support for Linux/IA64
8731 [Rolf Haberrecker <rolf@suse.de>]
8733 *) Assembler module support for Mingw32.
8736 *) Shared library support for HPUX (in shlib/).
8737 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8739 *) Shared library support for Solaris gcc.
8740 [Lutz Behnke <behnke@trustcenter.de>]
8742 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8744 *) PKCS7_encrypt() was adding text MIME headers twice because they
8745 were added manually and by SMIME_crlf_copy().
8748 *) In bntest.c don't call BN_rand with zero bits argument.
8749 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8751 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8752 case was implemented. This caused BN_div_recp() to fail occasionally.
8755 *) Add an optional second argument to the set_label() in the perl
8756 assembly language builder. If this argument exists and is set
8757 to 1 it signals that the assembler should use a symbol whose
8758 scope is the entire file, not just the current function. This
8759 is needed with MASM which uses the format label:: for this scope.
8760 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8762 *) Change the ASN1 types so they are typedefs by default. Before
8763 almost all types were #define'd to ASN1_STRING which was causing
8764 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8768 *) Change names of new functions to the new get1/get0 naming
8769 convention: After 'get1', the caller owns a reference count
8770 and has to call ..._free; 'get0' returns a pointer to some
8771 data structure without incrementing reference counters.
8772 (Some of the existing 'get' functions increment a reference
8773 counter, some don't.)
8774 Similarly, 'set1' and 'add1' functions increase reference
8775 counters or duplicate objects.
8778 *) Allow for the possibility of temp RSA key generation failure:
8779 the code used to assume it always worked and crashed on failure.
8782 *) Fix potential buffer overrun problem in BIO_printf().
8783 [Ulf Möller, using public domain code by Patrick Powell; problem
8784 pointed out by David Sacerdote <das33@cornell.edu>]
8786 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8787 RAND_egd() and RAND_status(). In the command line application,
8788 the EGD socket can be specified like a seed file using RANDFILE
8792 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8793 Some CAs (e.g. Verisign) distribute certificates in this form.
8796 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8797 list to exclude them. This means that no special compilation option
8798 is needed to use anonymous DH: it just needs to be included in the
8802 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8803 EVP_MD_type. The old functionality is available in a new macro called
8804 EVP_MD_md(). Change code that uses it and update docs.
8807 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8808 where the 'void *' argument is replaced by a function pointer argument.
8809 Previously 'void *' was abused to point to functions, which works on
8810 many platforms, but is not correct. As these functions are usually
8811 called by macros defined in OpenSSL header files, most source code
8812 should work without changes.
8815 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8816 sections with information on -D... compiler switches used for
8817 compiling the library so that applications can see them. To enable
8818 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8819 must be defined. E.g.,
8820 #define OPENSSL_ALGORITHM_DEFINES
8821 #include <openssl/opensslconf.h>
8822 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8823 [Richard Levitte, Ulf and Bodo Möller]
8825 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8829 *) Change the 'other' type in certificate aux info to a STACK_OF
8830 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8831 the required ASN1 format: arbitrary types determined by an OID.
8834 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8835 argument to 'req'. This is not because the function is newer or
8836 better than others it just uses the work 'NEW' in the certificate
8837 request header lines. Some software needs this.
8840 *) Reorganise password command line arguments: now passwords can be
8841 obtained from various sources. Delete the PEM_cb function and make
8842 it the default behaviour: i.e. if the callback is NULL and the
8843 usrdata argument is not NULL interpret it as a null terminated pass
8844 phrase. If usrdata and the callback are NULL then the pass phrase
8845 is prompted for as usual.
8848 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8849 the support is automatically enabled. The resulting binaries will
8850 autodetect the card and use it if present.
8851 [Ben Laurie and Compaq Inc.]
8853 *) Work around for Netscape hang bug. This sends certificate request
8854 and server done in one record. Since this is perfectly legal in the
8855 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8856 the bugs/SSLv3 entry for more info.
8859 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8862 *) Add -rand argument to smime and pkcs12 applications and read/write
8866 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8869 *) Add command line password options to the remaining applications.
8872 *) Bug fix for BN_div_recp() for numerators with an even number of
8876 *) More tests in bntest.c, and changed test_bn output.
8879 *) ./config recognizes MacOS X now.
8882 *) Bug fix for BN_div() when the first words of num and divsor are
8883 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8886 *) Add support for various broken PKCS#8 formats, and command line
8887 options to produce them.
8890 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8891 get temporary BIGNUMs from a BN_CTX.
8894 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8898 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8899 include a #define from the old name to the new. The original intent
8900 was that statically linked binaries could for example just call
8901 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8902 link with digests. This never worked becayse SSLeay_add_all_digests()
8903 and SSLeay_add_all_ciphers() were in the same source file so calling
8904 one would link with the other. They are now in separate source files.
8907 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8910 *) Use a less unusual form of the Miller-Rabin primality test (it used
8911 a binary algorithm for exponentiation integrated into the Miller-Rabin
8912 loop, our standard modexp algorithms are faster).
8915 *) Support for the EBCDIC character set completed.
8916 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8918 *) Source code cleanups: use const where appropriate, eliminate casts,
8919 use void * instead of char * in lhash.
8922 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8923 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8924 this the server could overwrite ephemeral keys that the client
8928 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8929 using 50 iterations of the Rabin-Miller test.
8931 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8932 iterations of the Rabin-Miller test as required by the appendix
8933 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8934 As BN_is_prime_fasttest includes trial division, DSA parameter
8935 generation becomes much faster.
8937 This implies a change for the callback functions in DSA_is_prime
8938 and DSA_generate_parameters: The callback function is called once
8939 for each positive witness in the Rabin-Miller test, not just
8940 occasionally in the inner loop; and the parameters to the
8941 callback function now provide an iteration count for the outer
8942 loop rather than for the current invocation of the inner loop.
8943 DSA_generate_parameters additionally can call the callback
8944 function with an 'iteration count' of -1, meaning that a
8945 candidate has passed the trial division test (when q is generated
8946 from an application-provided seed, trial division is skipped).
8949 *) New function BN_is_prime_fasttest that optionally does trial
8950 division before starting the Rabin-Miller test and has
8951 an additional BN_CTX * argument (whereas BN_is_prime always
8952 has to allocate at least one BN_CTX).
8953 'callback(1, -1, cb_arg)' is called when a number has passed the
8954 trial division stage.
8957 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8961 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8964 *) New function BN_pseudo_rand().
8967 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8968 bignum version of BN_from_montgomery() with the working code from
8969 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8973 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8974 made it impossible to use the same SSL_SESSION data structure in
8975 SSL2 clients in multiple threads.
8978 *) The return value of RAND_load_file() no longer counts bytes obtained
8979 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8980 to seed the PRNG (previously an explicit byte count was required).
8981 [Ulf Möller, Bodo Möller]
8983 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8984 used (char *) instead of (void *) and had casts all over the place.
8987 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8990 *) Retain source code compatibility for BN_prime_checks macro:
8991 BN_is_prime(..., BN_prime_checks, ...) now uses
8992 BN_prime_checks_for_size to determine the appropriate number of
8993 Rabin-Miller iterations.
8996 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8997 DH_CHECK_P_NOT_SAFE_PRIME.
8998 (Check if this is true? OpenPGP calls them "strong".)
9001 *) Merge the functionality of "dh" and "gendh" programs into a new program
9002 "dhparam". The old programs are retained for now but will handle DH keys
9003 (instead of parameters) in future.
9006 *) Make the ciphers, s_server and s_client programs check the return values
9007 when a new cipher list is set.
9010 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9011 ciphers. Before when the 56bit ciphers were enabled the sorting was
9014 The syntax for the cipher sorting has been extended to support sorting by
9015 cipher-strength (using the strength_bits hard coded in the tables).
9016 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9018 Fix a bug in the cipher-command parser: when supplying a cipher command
9019 string with an "undefined" symbol (neither command nor alphanumeric
9020 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9021 an error is flagged.
9023 Due to the strength-sorting extension, the code of the
9024 ssl_create_cipher_list() function was completely rearranged. I hope that
9025 the readability was also increased :-)
9026 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9028 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9029 for the first serial number and places 2 in the serial number file. This
9030 avoids problems when the root CA is created with serial number zero and
9031 the first user certificate has the same issuer name and serial number
9035 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9036 the new code. Add documentation for this stuff.
9039 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9040 X509_*() to X509at_*() on the grounds that they don't handle X509
9041 structures and behave in an analagous way to the X509v3 functions:
9042 they shouldn't be called directly but wrapper functions should be used
9045 So we also now have some wrapper functions that call the X509at functions
9046 when passed certificate requests. (TO DO: similar things can be done with
9047 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9048 things. Some of these need some d2i or i2d and print functionality
9049 because they handle more complex structures.)
9052 *) Add missing #ifndefs that caused missing symbols when building libssl
9053 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9054 NO_RSA in ssl/s2*.c.
9055 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9057 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9058 has a return value which indicates the quality of the random data
9059 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9060 error queue. New function RAND_pseudo_bytes() generates output that is
9061 guaranteed to be unique but not unpredictable. RAND_add is like
9062 RAND_seed, but takes an extra argument for an entropy estimate
9063 (RAND_seed always assumes full entropy).
9066 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9067 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9068 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9069 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9070 false-positive rate of at most 2^-80 for random input.
9073 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9076 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9077 in the 0.9.5 release), this returns the chain
9078 from an X509_CTX structure with a dup of the stack and all
9079 the X509 reference counts upped: so the stack will exist
9080 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9083 Also make SSL_SESSION_print() print out the verify return
9087 *) Add manpage for the pkcs12 command. Also change the default
9088 behaviour so MAC iteration counts are used unless the new
9089 -nomaciter option is used. This improves file security and
9090 only older versions of MSIE (4.0 for example) need it.
9093 *) Honor the no-xxx Configure options when creating .DEF files.
9096 *) Add PKCS#10 attributes to field table: challengePassword,
9097 unstructuredName and unstructuredAddress. These are taken from
9098 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9099 international characters are used.
9101 More changes to X509_ATTRIBUTE code: allow the setting of types
9102 based on strings. Remove the 'loc' parameter when adding
9103 attributes because these will be a SET OF encoding which is sorted
9107 *) Initial changes to the 'req' utility to allow request generation
9108 automation. This will allow an application to just generate a template
9109 file containing all the field values and have req construct the
9112 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9113 used all over the place including certificate requests and PKCS#7
9114 structures. They are currently handled manually where necessary with
9115 some primitive wrappers for PKCS#7. The new functions behave in a
9116 manner analogous to the X509 extension functions: they allow
9117 attributes to be looked up by NID and added.
9119 Later something similar to the X509V3 code would be desirable to
9120 automatically handle the encoding, decoding and printing of the
9121 more complex types. The string types like challengePassword can
9122 be handled by the string table functions.
9124 Also modified the multi byte string table handling. Now there is
9125 a 'global mask' which masks out certain types. The table itself
9126 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9127 is useful when for example there is only one permissible type
9128 (as in countryName) and using the mask might result in no valid
9132 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9133 SSL_get_peer_finished to allow applications to obtain the latest
9134 Finished messages sent to the peer or expected from the peer,
9135 respectively. (SSL_get_peer_finished is usually the Finished message
9136 actually received from the peer, otherwise the protocol will be aborted.)
9138 As the Finished message are message digests of the complete handshake
9139 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9140 be used for external authentication procedures when the authentication
9141 provided by SSL/TLS is not desired or is not enough.
9144 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9145 the host supports BWX extension and if Compaq C is present on the
9146 $PATH. Just exploiting of the BWX extension results in 20-30%
9147 performance kick for some algorithms, e.g. DES and RC4 to mention
9148 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9152 *) Add support for MS "fast SGC". This is arguably a violation of the
9153 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9154 weak crypto and after checking the certificate is SGC a second one
9155 with strong crypto. MS SGC stops the first handshake after receiving
9156 the server certificate message and sends a second client hello. Since
9157 a server will typically do all the time consuming operations before
9158 expecting any further messages from the client (server key exchange
9159 is the most expensive) there is little difference between the two.
9161 To get OpenSSL to support MS SGC we have to permit a second client
9162 hello message after we have sent server done. In addition we have to
9163 reset the MAC if we do get this second client hello.
9166 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9167 if a DER encoded private key is RSA or DSA traditional format. Changed
9168 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9169 format DER encoded private key. Newer code should use PKCS#8 format which
9170 has the key type encoded in the ASN1 structure. Added DER private key
9171 support to pkcs8 application.
9174 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9175 ciphersuites has been selected (as required by the SSL 3/TLS 1
9176 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9177 is set, we interpret this as a request to violate the specification
9178 (the worst that can happen is a handshake failure, and 'correct'
9179 behaviour would result in a handshake failure anyway).
9182 *) In SSL_CTX_add_session, take into account that there might be multiple
9183 SSL_SESSION structures with the same session ID (e.g. when two threads
9184 concurrently obtain them from an external cache).
9185 The internal cache can handle only one SSL_SESSION with a given ID,
9186 so if there's a conflict, we now throw out the old one to achieve
9190 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9191 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9192 some routines that use cipher OIDs: some ciphers do not have OIDs
9193 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9197 *) Simplify the trust setting structure and code. Now we just have
9198 two sequences of OIDs for trusted and rejected settings. These will
9199 typically have values the same as the extended key usage extension
9200 and any application specific purposes.
9202 The trust checking code now has a default behaviour: it will just
9203 check for an object with the same NID as the passed id. Functions can
9204 be provided to override either the default behaviour or the behaviour
9205 for a given id. SSL client, server and email already have functions
9206 in place for compatibility: they check the NID and also return "trusted"
9207 if the certificate is self signed.
9210 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9211 traditional format into an EVP_PKEY structure.
9214 *) Add a password callback function PEM_cb() which either prompts for
9215 a password if usr_data is NULL or otherwise assumes it is a null
9216 terminated password. Allow passwords to be passed on command line
9217 environment or config files in a few more utilities.
9220 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9221 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9222 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9223 Update documentation.
9226 *) Support for ASN1 "NULL" type. This could be handled before by using
9227 ASN1_TYPE but there wasn't any function that would try to read a NULL
9228 and produce an error if it couldn't. For compatibility we also have
9229 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9230 don't allocate anything because they don't need to.
9233 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9235 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9237 *) Rebuild of the memory allocation routines used by OpenSSL code and
9238 possibly others as well. The purpose is to make an interface that
9239 provide hooks so anyone can build a separate set of allocation and
9240 deallocation routines to be used by OpenSSL, for example memory
9241 pool implementations, or something else, which was previously hard
9242 since Malloc(), Realloc() and Free() were defined as macros having
9243 the values malloc, realloc and free, respectively (except for Win32
9244 compilations). The same is provided for memory debugging code.
9245 OpenSSL already comes with functionality to find memory leaks, but
9246 this gives people a chance to debug other memory problems.
9248 With these changes, a new set of functions and macros have appeared:
9250 CRYPTO_set_mem_debug_functions() [F]
9251 CRYPTO_get_mem_debug_functions() [F]
9252 CRYPTO_dbg_set_options() [F]
9253 CRYPTO_dbg_get_options() [F]
9254 CRYPTO_malloc_debug_init() [M]
9256 The memory debug functions are NULL by default, unless the library
9257 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9258 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9259 gives the standard debugging functions that come with OpenSSL) or
9260 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9261 provided by the library user) must be used. When the standard
9262 debugging functions are used, CRYPTO_dbg_set_options can be used to
9263 request additional information:
9264 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9265 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9267 Also, things like CRYPTO_set_mem_functions will always give the
9268 expected result (the new set of functions is used for allocation
9269 and deallocation) at all times, regardless of platform and compiler
9272 To finish it up, some functions that were never use in any other
9273 way than through macros have a new API and new semantic:
9276 CRYPTO_dbg_realloc()
9279 All macros of value have retained their old syntax.
9280 [Richard Levitte and Bodo Moeller]
9282 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9283 ordering of SMIMECapabilities wasn't in "strength order" and there
9284 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9288 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9289 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9290 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9292 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9293 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9294 functionality to handle multipart/signed properly) and a utility
9295 called 'smime' to call all this stuff. This is based on code I
9296 originally wrote for Celo who have kindly allowed it to be
9297 included in OpenSSL.
9300 *) Add variants des_set_key_checked and des_set_key_unchecked of
9301 des_set_key (aka des_key_sched). Global variable des_check_key
9302 decides which of these is called by des_set_key; this way
9303 des_check_key behaves as it always did, but applications and
9304 the library itself, which was buggy for des_check_key == 1,
9305 have a cleaner way to pick the version they need.
9308 *) New function PKCS12_newpass() which changes the password of a
9312 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9313 dynamic mix. In both cases the ids can be used as an index into the
9314 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9315 functions so they accept a list of the field values and the
9316 application doesn't need to directly manipulate the X509_TRUST
9320 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9324 *) Modify the way the V3 extension code looks up extensions. This now
9325 works in a similar way to the object code: we have some "standard"
9326 extensions in a static table which is searched with OBJ_bsearch()
9327 and the application can add dynamic ones if needed. The file
9328 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9329 updated whenever a new extension is added to the core code and kept
9330 in ext_nid order. There is a simple program 'tabtest.c' which checks
9331 this. New extensions are not added too often so this file can readily
9332 be maintained manually.
9334 There are two big advantages in doing things this way. The extensions
9335 can be looked up immediately and no longer need to be "added" using
9336 X509V3_add_standard_extensions(): this function now does nothing.
9337 [Side note: I get *lots* of email saying the extension code doesn't
9338 work because people forget to call this function]
9339 Also no dynamic allocation is done unless new extensions are added:
9340 so if we don't add custom extensions there is no need to call
9341 X509V3_EXT_cleanup().
9344 *) Modify enc utility's salting as follows: make salting the default. Add a
9345 magic header, so unsalted files fail gracefully instead of just decrypting
9346 to garbage. This is because not salting is a big security hole, so people
9347 should be discouraged from doing it.
9350 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9351 digest to be passed on the command line but it only used this
9352 parameter when signing a certificate. Modified so all relevant
9353 operations are affected by the digest parameter including the
9354 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9355 DSA key was used because it didn't fix the digest.
9358 *) Initial certificate chain verify code. Currently tests the untrusted
9359 certificates for consistency with the verify purpose (which is set
9360 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9362 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9363 this is because it will reject chains with invalid extensions whereas
9364 every previous version of OpenSSL and SSLeay made no checks at all.
9366 Trust code: checks the root CA for the relevant trust settings. Trust
9367 settings have an initial value consistent with the verify purpose: e.g.
9368 if the verify purpose is for SSL client use it expects the CA to be
9369 trusted for SSL client use. However the default value can be changed to
9370 permit custom trust settings: one example of this would be to only trust
9371 certificates from a specific "secure" set of CAs.
9373 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9374 which should be used for version portability: especially since the
9375 verify structure is likely to change more often now.
9377 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9378 to set them. If not set then assume SSL clients will verify SSL servers
9381 Two new options to the verify program: -untrusted allows a set of
9382 untrusted certificates to be passed in and -purpose which sets the
9383 intended purpose of the certificate. If a purpose is set then the
9384 new chain verify code is used to check extension consistency.
9387 *) Support for the authority information access extension.
9390 *) Modify RSA and DSA PEM read routines to transparently handle
9391 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9392 public keys in a format compatible with certificate
9393 SubjectPublicKeyInfo structures. Unfortunately there were already
9394 functions called *_PublicKey_* which used various odd formats so
9395 these are retained for compatibility: however the DSA variants were
9396 never in a public release so they have been deleted. Changed dsa/rsa
9397 utilities to handle the new format: note no releases ever handled public
9398 keys so we should be OK.
9400 The primary motivation for this change is to avoid the same fiasco
9401 that dogs private keys: there are several incompatible private key
9402 formats some of which are standard and some OpenSSL specific and
9403 require various evil hacks to allow partial transparent handling and
9404 even then it doesn't work with DER formats. Given the option anything
9405 other than PKCS#8 should be dumped: but the other formats have to
9406 stay in the name of compatibility.
9408 With public keys and the benefit of hindsight one standard format
9409 is used which works with EVP_PKEY, RSA or DSA structures: though
9410 it clearly returns an error if you try to read the wrong kind of key.
9412 Added a -pubkey option to the 'x509' utility to output the public key.
9413 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9414 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9415 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9416 that do the same as the EVP_PKEY_assign_*() except they up the
9417 reference count of the added key (they don't "swallow" the
9421 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9422 CRLs would fail if the file contained no certificates or no CRLs:
9423 added a new function to read in both types and return the number
9424 read: this means that if none are read it will be an error. The
9425 DER versions of the certificate and CRL reader would always fail
9426 because it isn't possible to mix certificates and CRLs in DER format
9427 without choking one or the other routine. Changed this to just read
9428 a certificate: this is the best we can do. Also modified the code
9429 in apps/verify.c to take notice of return codes: it was previously
9430 attempting to read in certificates from NULL pointers and ignoring
9431 any errors: this is one reason why the cert and CRL reader seemed
9432 to work. It doesn't check return codes from the default certificate
9433 routines: these may well fail if the certificates aren't installed.
9436 *) Code to support otherName option in GeneralName.
9439 *) First update to verify code. Change the verify utility
9440 so it warns if it is passed a self signed certificate:
9441 for consistency with the normal behaviour. X509_verify
9442 has been modified to it will now verify a self signed
9443 certificate if *exactly* the same certificate appears
9444 in the store: it was previously impossible to trust a
9445 single self signed certificate. This means that:
9446 openssl verify ss.pem
9447 now gives a warning about a self signed certificate but
9448 openssl verify -CAfile ss.pem ss.pem
9452 *) For servers, store verify_result in SSL_SESSION data structure
9453 (and add it to external session representation).
9454 This is needed when client certificate verifications fails,
9455 but an application-provided verification callback (set by
9456 SSL_CTX_set_cert_verify_callback) allows accepting the session
9457 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9458 but returns 1): When the session is reused, we have to set
9459 ssl->verify_result to the appropriate error code to avoid
9461 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9463 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9464 case in PKCS7_dataInit() where the signed PKCS7 structure
9465 didn't contain any existing data because it was being created.
9466 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9468 *) Add a salt to the key derivation routines in enc.c. This
9469 forms the first 8 bytes of the encrypted file. Also add a
9470 -S option to allow a salt to be input on the command line.
9473 *) New function X509_cmp(). Oddly enough there wasn't a function
9474 to compare two certificates. We do this by working out the SHA1
9475 hash and comparing that. X509_cmp() will be needed by the trust
9479 *) SSL_get1_session() is like SSL_get_session(), but increments
9480 the reference count in the SSL_SESSION returned.
9481 [Geoff Thorpe <geoff@eu.c2.net>]
9483 *) Fix for 'req': it was adding a null to request attributes.
9484 Also change the X509_LOOKUP and X509_INFO code to handle
9485 certificate auxiliary information.
9488 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9492 *) Add the possibility to add extra information to the memory leak
9493 detecting output, to form tracebacks, showing from where each
9494 allocation was originated: CRYPTO_push_info("constant string") adds
9495 the string plus current file name and line number to a per-thread
9496 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9497 is like calling CYRPTO_pop_info() until the stack is empty.
9498 Also updated memory leak detection code to be multi-thread-safe.
9501 *) Add options -text and -noout to pkcs7 utility and delete the
9502 encryption options which never did anything. Update docs.
9505 *) Add options to some of the utilities to allow the pass phrase
9506 to be included on either the command line (not recommended on
9507 OSes like Unix) or read from the environment. Update the
9508 manpages and fix a few bugs.
9511 *) Add a few manpages for some of the openssl commands.
9514 *) Fix the -revoke option in ca. It was freeing up memory twice,
9515 leaking and not finding already revoked certificates.
9518 *) Extensive changes to support certificate auxiliary information.
9519 This involves the use of X509_CERT_AUX structure and X509_AUX
9520 functions. An X509_AUX function such as PEM_read_X509_AUX()
9521 can still read in a certificate file in the usual way but it
9522 will also read in any additional "auxiliary information". By
9523 doing things this way a fair degree of compatibility can be
9524 retained: existing certificates can have this information added
9525 using the new 'x509' options.
9527 Current auxiliary information includes an "alias" and some trust
9528 settings. The trust settings will ultimately be used in enhanced
9529 certificate chain verification routines: currently a certificate
9530 can only be trusted if it is self signed and then it is trusted
9534 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9535 The problem was that one of the replacement routines had not been working
9536 since SSLeay releases. For now the offending routine has been replaced
9537 with non-optimised assembler. Even so, this now gives around 95%
9538 performance improvement for 1024 bit RSA signs.
9541 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9542 handling. Most clients have the effective key size in bits equal to
9543 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9544 A few however don't do this and instead use the size of the decrypted key
9545 to determine the RC2 key length and the AlgorithmIdentifier to determine
9546 the effective key length. In this case the effective key length can still
9547 be 40 bits but the key length can be 168 bits for example. This is fixed
9548 by manually forcing an RC2 key into the EVP_PKEY structure because the
9549 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9550 the key length and effective key length are equal.
9553 *) Add a bunch of functions that should simplify the creation of
9554 X509_NAME structures. Now you should be able to do:
9555 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9556 and have it automatically work out the correct field type and fill in
9557 the structures. The more adventurous can try:
9558 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9559 and it will (hopefully) work out the correct multibyte encoding.
9562 *) Change the 'req' utility to use the new field handling and multibyte
9563 copy routines. Before the DN field creation was handled in an ad hoc
9564 way in req, ca, and x509 which was rather broken and didn't support
9565 BMPStrings or UTF8Strings. Since some software doesn't implement
9566 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9567 using the dirstring_type option. See the new comment in the default
9568 openssl.cnf for more info.
9571 *) Make crypto/rand/md_rand.c more robust:
9572 - Assure unique random numbers after fork().
9573 - Make sure that concurrent threads access the global counter and
9574 md serializably so that we never lose entropy in them
9575 or use exactly the same state in multiple threads.
9576 Access to the large state is not always serializable because
9577 the additional locking could be a performance killer, and
9578 md should be large enough anyway.
9581 *) New file apps/app_rand.c with commonly needed functionality
9582 for handling the random seed file.
9584 Use the random seed file in some applications that previously did not:
9586 dsaparam -genkey (which also ignored its '-rand' option),
9589 x509 (when signing).
9590 Except on systems with /dev/urandom, it is crucial to have a random
9591 seed file at least for key creation, DSA signing, and for DH exchanges;
9592 for RSA signatures we could do without one.
9594 gendh and gendsa (unlike genrsa) used to read only the first byte
9595 of each file listed in the '-rand' option. The function as previously
9596 found in genrsa is now in app_rand.c and is used by all programs
9597 that support '-rand'.
9600 *) In RAND_write_file, use mode 0600 for creating files;
9601 don't just chmod when it may be too late.
9604 *) Report an error from X509_STORE_load_locations
9605 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9608 *) New function ASN1_mbstring_copy() this copies a string in either
9609 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9610 into an ASN1_STRING type. A mask of permissible types is passed
9611 and it chooses the "minimal" type to use or an error if not type
9615 *) Add function equivalents to the various macros in asn1.h. The old
9616 macros are retained with an M_ prefix. Code inside the library can
9617 use the M_ macros. External code (including the openssl utility)
9618 should *NOT* in order to be "shared library friendly".
9621 *) Add various functions that can check a certificate's extensions
9622 to see if it usable for various purposes such as SSL client,
9623 server or S/MIME and CAs of these types. This is currently
9624 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9625 verification. Also added a -purpose flag to x509 utility to
9626 print out all the purposes.
9629 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9633 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9634 for, obtain and decode and extension and obtain its critical flag.
9635 This allows all the necessary extension code to be handled in a
9636 single function call.
9639 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9640 platforms. See crypto/rc4/rc4_enc.c for further details.
9643 *) New -noout option to asn1parse. This causes no output to be produced
9644 its main use is when combined with -strparse and -out to extract data
9645 from a file (which may not be in ASN.1 format).
9648 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9649 when producing the local key id.
9650 [Richard Levitte <levitte@stacken.kth.se>]
9652 *) New option -dhparam in s_server. This allows a DH parameter file to be
9653 stated explicitly. If it is not stated then it tries the first server
9654 certificate file. The previous behaviour hard coded the filename
9658 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9659 a public key to be input or output. For example:
9660 openssl rsa -in key.pem -pubout -out pubkey.pem
9661 Also added necessary DSA public key functions to handle this.
9664 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9665 in the message. This was handled by allowing
9666 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9667 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9669 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9670 to the end of the strings whereas this didn't. This would cause problems
9671 if strings read with d2i_ASN1_bytes() were later modified.
9672 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9674 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9675 data and it contains EOF it will end up returning an error. This is
9676 caused by input 46 bytes long. The cause is due to the way base64
9677 BIOs find the start of base64 encoded data. They do this by trying a
9678 trial decode on each line until they find one that works. When they
9679 do a flag is set and it starts again knowing it can pass all the
9680 data directly through the decoder. Unfortunately it doesn't reset
9681 the context it uses. This means that if EOF is reached an attempt
9682 is made to pass two EOFs through the context and this causes the
9683 resulting error. This can also cause other problems as well. As is
9684 usual with these problems it takes *ages* to find and the fix is
9685 trivial: move one line.
9686 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9688 *) Ugly workaround to get s_client and s_server working under Windows. The
9689 old code wouldn't work because it needed to select() on sockets and the
9690 tty (for keypresses and to see if data could be written). Win32 only
9691 supports select() on sockets so we select() with a 1s timeout on the
9692 sockets and then see if any characters are waiting to be read, if none
9693 are present then we retry, we also assume we can always write data to
9694 the tty. This isn't nice because the code then blocks until we've
9695 received a complete line of data and it is effectively polling the
9696 keyboard at 1s intervals: however it's quite a bit better than not
9697 working at all :-) A dedicated Windows application might handle this
9698 with an event loop for example.
9701 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9702 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9703 will be called when RSA_sign() and RSA_verify() are used. This is useful
9704 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9705 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9706 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9707 This necessitated the support of an extra signature type NID_md5_sha1
9708 for SSL signatures and modifications to the SSL library to use it instead
9709 of calling RSA_public_decrypt() and RSA_private_encrypt().
9712 *) Add new -verify -CAfile and -CApath options to the crl program, these
9713 will lookup a CRL issuers certificate and verify the signature in a
9714 similar way to the verify program. Tidy up the crl program so it
9715 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9716 less strict. It will now permit CRL extensions even if it is not
9717 a V2 CRL: this will allow it to tolerate some broken CRLs.
9720 *) Initialize all non-automatic variables each time one of the openssl
9721 sub-programs is started (this is necessary as they may be started
9722 multiple times from the "OpenSSL>" prompt).
9723 [Lennart Bang, Bodo Moeller]
9725 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9726 removing all other RSA functionality (this is what NO_RSA does). This
9727 is so (for example) those in the US can disable those operations covered
9728 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9732 *) Non-copying interface to BIO pairs.
9733 (still largely untested)
9736 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9737 ASCII string. This was handled independently in various places before.
9740 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9741 UTF8 strings a character at a time.
9744 *) Use client_version from client hello to select the protocol
9745 (s23_srvr.c) and for RSA client key exchange verification
9746 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9749 *) Add various utility functions to handle SPKACs, these were previously
9750 handled by poking round in the structure internals. Added new function
9751 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9752 print, verify and generate SPKACs. Based on an original idea from
9753 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9756 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9759 *) Allow the config file extension section to be overwritten on the
9760 command line. Based on an original idea from Massimiliano Pala
9761 <madwolf@comune.modena.it>. The new option is called -extensions
9762 and can be applied to ca, req and x509. Also -reqexts to override
9763 the request extensions in req and -crlexts to override the crl extensions
9767 *) Add new feature to the SPKAC handling in ca. Now you can include
9768 the same field multiple times by preceding it by "XXXX." for example:
9771 this is the same syntax as used in the req config file.
9774 *) Allow certificate extensions to be added to certificate requests. These
9775 are specified in a 'req_extensions' option of the req section of the
9776 config file. They can be printed out with the -text option to req but
9777 are otherwise ignored at present.
9780 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9781 data read consists of only the final block it would not decrypted because
9782 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9783 A misplaced 'break' also meant the decrypted final block might not be
9784 copied until the next read.
9787 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9788 a few extra parameters to the DH structure: these will be useful if
9789 for example we want the value of 'q' or implement X9.42 DH.
9792 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9793 provides hooks that allow the default DSA functions or functions on a
9794 "per key" basis to be replaced. This allows hardware acceleration and
9795 hardware key storage to be handled without major modification to the
9796 library. Also added low level modexp hooks and CRYPTO_EX structure and
9797 associated functions.
9800 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9801 as "read only": it can't be written to and the buffer it points to will
9802 not be freed. Reading from a read only BIO is much more efficient than
9803 a normal memory BIO. This was added because there are several times when
9804 an area of memory needs to be read from a BIO. The previous method was
9805 to create a memory BIO and write the data to it, this results in two
9806 copies of the data and an O(n^2) reading algorithm. There is a new
9807 function BIO_new_mem_buf() which creates a read only memory BIO from
9808 an area of memory. Also modified the PKCS#7 routines to use read only
9812 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9813 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9814 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9815 but a retry condition occured while trying to read the rest.
9818 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9819 NID_pkcs7_encrypted by default: this was wrong since this should almost
9820 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9821 the encrypted data type: this is a more sensible place to put it and it
9822 allows the PKCS#12 code to be tidied up that duplicated this
9826 *) Changed obj_dat.pl script so it takes its input and output files on
9827 the command line. This should avoid shell escape redirection problems
9831 *) Initial support for certificate extension requests, these are included
9832 in things like Xenroll certificate requests. Included functions to allow
9833 extensions to be obtained and added.
9836 *) -crlf option to s_client and s_server for sending newlines as
9837 CRLF (as required by many protocols).
9840 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9842 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9843 [Ralf S. Engelschall]
9845 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9846 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9848 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9852 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9853 DH parameters/keys (q is lost during that conversion, but the resulting
9854 DH parameters contain its length).
9856 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9857 much faster than DH_generate_parameters (which creates parameters
9858 where p = 2*q + 1), and also the smaller q makes DH computations
9859 much more efficient (160-bit exponentiation instead of 1024-bit
9860 exponentiation); so this provides a convenient way to support DHE
9861 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9862 utter importance to use
9863 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9865 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9866 when such DH parameters are used, because otherwise small subgroup
9867 attacks may become possible!
9870 *) Avoid memory leak in i2d_DHparams.
9873 *) Allow the -k option to be used more than once in the enc program:
9874 this allows the same encrypted message to be read by multiple recipients.
9877 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9878 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9879 it will always use the numerical form of the OID, even if it has a short
9883 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9884 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9885 otherwise bn_mod_exp was called. In the case of hardware keys for example
9886 no private key components need be present and it might store extra data
9887 in the RSA structure, which cannot be accessed from bn_mod_exp.
9888 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9889 private key operations.
9892 *) Added support for SPARC Linux.
9895 *) pem_password_cb function type incompatibly changed from
9896 typedef int pem_password_cb(char *buf, int size, int rwflag);
9898 ....(char *buf, int size, int rwflag, void *userdata);
9899 so that applications can pass data to their callbacks:
9900 The PEM[_ASN1]_{read,write}... functions and macros now take an
9901 additional void * argument, which is just handed through whenever
9902 the password callback is called.
9903 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9905 New function SSL_CTX_set_default_passwd_cb_userdata.
9907 Compatibility note: As many C implementations push function arguments
9908 onto the stack in reverse order, the new library version is likely to
9909 interoperate with programs that have been compiled with the old
9910 pem_password_cb definition (PEM_whatever takes some data that
9911 happens to be on the stack as its last argument, and the callback
9912 just ignores this garbage); but there is no guarantee whatsoever that
9915 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9916 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9917 problems not only on Windows, but also on some Unix platforms.
9918 To avoid problematic command lines, these definitions are now in an
9919 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9920 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9923 *) MIPS III/IV assembler module is reimplemented.
9926 *) More DES library cleanups: remove references to srand/rand and
9927 delete an unused file.
9930 *) Add support for the the free Netwide assembler (NASM) under Win32,
9931 since not many people have MASM (ml) and it can be hard to obtain.
9932 This is currently experimental but it seems to work OK and pass all
9933 the tests. Check out INSTALL.W32 for info.
9936 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9937 without temporary keys kept an extra copy of the server key,
9938 and connections with temporary keys did not free everything in case
9942 *) New function RSA_check_key and new openssl rsa option -check
9943 for verifying the consistency of RSA keys.
9944 [Ulf Moeller, Bodo Moeller]
9946 *) Various changes to make Win32 compile work:
9947 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9948 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9949 comparison" warnings.
9950 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9953 *) Add a debugging option to PKCS#5 v2 key generation function: when
9954 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9955 derived keys are printed to stderr.
9958 *) Copy the flags in ASN1_STRING_dup().
9959 [Roman E. Pavlov <pre@mo.msk.ru>]
9961 *) The x509 application mishandled signing requests containing DSA
9962 keys when the signing key was also DSA and the parameters didn't match.
9964 It was supposed to omit the parameters when they matched the signing key:
9965 the verifying software was then supposed to automatically use the CA's
9966 parameters if they were absent from the end user certificate.
9968 Omitting parameters is no longer recommended. The test was also
9969 the wrong way round! This was probably due to unusual behaviour in
9970 EVP_cmp_parameters() which returns 1 if the parameters match.
9971 This meant that parameters were omitted when they *didn't* match and
9972 the certificate was useless. Certificates signed with 'ca' didn't have
9974 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9976 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9977 The interface is as follows:
9978 Applications can use
9979 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9980 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9981 "off" is now the default.
9982 The library internally uses
9983 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9984 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9985 to disable memory-checking temporarily.
9987 Some inconsistent states that previously were possible (and were
9988 even the default) are now avoided.
9990 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9991 with each memory chunk allocated; this is occasionally more helpful
9992 than just having a counter.
9994 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9996 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10000 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10001 which largely parallels "options", but is for changing API behaviour,
10002 whereas "options" are about protocol behaviour.
10003 Initial "mode" flags are:
10005 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10006 a single record has been written.
10007 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10008 retries use the same buffer location.
10009 (But all of the contents must be
10013 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10016 *) Fix problems with no-hmac etc.
10017 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10019 *) New functions RSA_get_default_method(), RSA_set_method() and
10020 RSA_get_method(). These allows replacement of RSA_METHODs without having
10021 to mess around with the internals of an RSA structure.
10024 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10025 Also really enable memory leak checks in openssl.c and in some
10027 [Chad C. Mulligan, Bodo Moeller]
10029 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10030 up the length of negative integers. This has now been simplified to just
10031 store the length when it is first determined and use it later, rather
10032 than trying to keep track of where data is copied and updating it to
10034 [Steve Henson, reported by Brien Wheeler
10035 <bwheeler@authentica-security.com>]
10037 *) Add a new function PKCS7_signatureVerify. This allows the verification
10038 of a PKCS#7 signature but with the signing certificate passed to the
10039 function itself. This contrasts with PKCS7_dataVerify which assumes the
10040 certificate is present in the PKCS#7 structure. This isn't always the
10041 case: certificates can be omitted from a PKCS#7 structure and be
10042 distributed by "out of band" means (such as a certificate database).
10045 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10046 function prototypes in pem.h, also change util/mkdef.pl to add the
10047 necessary function names.
10050 *) mk1mf.pl (used by Windows builds) did not properly read the
10051 options set by Configure in the top level Makefile, and Configure
10052 was not even able to write more than one option correctly.
10053 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10056 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10057 file to be loaded from a BIO or FILE pointer. The BIO version will
10058 for example allow memory BIOs to contain config info.
10061 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10062 Whoever hopes to achieve shared-library compatibility across versions
10063 must use this, not the compile-time macro.
10064 (Exercise 0.9.4: Which is the minimum library version required by
10066 Note: All this applies only to multi-threaded programs, others don't
10070 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10071 through a BIO pair triggered the default case, i.e.
10072 SSLerr(...,SSL_R_UNKNOWN_STATE).
10075 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10076 can use the SSL library even if none of the specific BIOs is
10080 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10081 for the encoded length.
10082 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10084 *) Add initial documentation of the X509V3 functions.
10087 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10088 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10089 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10090 secure PKCS#8 private key format with a high iteration count.
10093 *) Fix determination of Perl interpreter: A perl or perl5
10094 _directory_ in $PATH was also accepted as the interpreter.
10095 [Ralf S. Engelschall]
10097 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10098 wrong with it but it was very old and did things like calling
10099 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10100 unusual formatting.
10103 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10104 to use the new extension code.
10107 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10108 with macros. This should make it easier to change their form, add extra
10109 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10113 *) Add to configuration table a new entry that can specify an alternative
10114 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10115 according to Mark Crispin <MRC@Panda.COM>.
10119 *) DES CBC did not update the IV. Weird.
10122 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10123 Changing the behaviour of the former might break existing programs --
10124 where IV updating is needed, des_ncbc_encrypt can be used.
10127 *) When bntest is run from "make test" it drives bc to check its
10128 calculations, as well as internally checking them. If an internal check
10129 fails, it needs to cause bc to give a non-zero result or make test carries
10130 on without noticing the failure. Fixed.
10133 *) DES library cleanups.
10136 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10137 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10138 ciphers. NOTE: although the key derivation function has been verified
10139 against some published test vectors it has not been extensively tested
10140 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10144 *) Instead of "mkdir -p", which is not fully portable, use new
10145 Perl script "util/mkdir-p.pl".
10148 *) Rewrite the way password based encryption (PBE) is handled. It used to
10149 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10150 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10151 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10152 the 'parameter' field of the AlgorithmIdentifier is passed to the
10153 underlying key generation function so it must do its own ASN1 parsing.
10154 This has also changed the EVP_PBE_CipherInit() function which now has a
10155 'parameter' argument instead of literal salt and iteration count values
10156 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10159 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10160 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10161 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10162 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10163 value was just used as a "magic string" and not used directly its
10164 value doesn't matter.
10167 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10171 *) "linux-sparc64" configuration (ultrapenguin).
10172 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10173 "linux-sparc" configuration.
10174 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10176 *) config now generates no-xxx options for missing ciphers.
10179 *) Support the EBCDIC character set (work in progress).
10180 File ebcdic.c not yet included because it has a different license.
10181 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10183 *) Support BS2000/OSD-POSIX.
10184 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10186 *) Make callbacks for key generation use void * instead of char *.
10189 *) Make S/MIME samples compile (not yet tested).
10192 *) Additional typesafe stacks.
10195 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10199 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10201 *) New configuration variant "sco5-gcc".
10203 *) Updated some demos.
10204 [Sean O Riordain, Wade Scholine]
10206 *) Add missing BIO_free at exit of pkcs12 application.
10209 *) Fix memory leak in conf.c.
10212 *) Updates for Win32 to assembler version of MD5.
10215 *) Set #! path to perl in apps/der_chop to where we found it
10216 instead of using a fixed path.
10219 *) SHA library changes for irix64-mips4-cc.
10222 *) Improvements for VMS support.
10226 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10228 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10229 This also avoids the problems with SC4.2 and unpatched SC5.
10230 [Andy Polyakov <appro@fy.chalmers.se>]
10232 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10233 These are required because of the typesafe stack would otherwise break
10234 existing code. If old code used a structure member which used to be STACK
10235 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10236 sk_num or sk_value it would produce an error because the num, data members
10237 are not present in STACK_OF. Now it just produces a warning. sk_set
10238 replaces the old method of assigning a value to sk_value
10239 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10240 that does this will no longer work (and should use sk_set instead) but
10241 this could be regarded as a "questionable" behaviour anyway.
10244 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10245 correctly handle encrypted S/MIME data.
10248 *) Change type of various DES function arguments from des_cblock
10249 (which means, in function argument declarations, pointer to char)
10250 to des_cblock * (meaning pointer to array with 8 char elements),
10251 which allows the compiler to do more typechecking; it was like
10252 that back in SSLeay, but with lots of ugly casts.
10254 Introduce new type const_des_cblock.
10257 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10258 problems: find RecipientInfo structure that matches recipient certificate
10259 and initialise the ASN1 structures properly based on passed cipher.
10262 *) Belatedly make the BN tests actually check the results.
10265 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10266 to and from BNs: it was completely broken. New compilation option
10267 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10268 key elements as negative integers.
10271 *) Reorganize and speed up MD5.
10272 [Andy Polyakov <appro@fy.chalmers.se>]
10275 [Richard Levitte <richard@levitte.org>]
10277 *) New option -out to asn1parse to allow the parsed structure to be
10278 output to a file. This is most useful when combined with the -strparse
10279 option to examine the output of things like OCTET STRINGS.
10282 *) Make SSL library a little more fool-proof by not requiring any longer
10283 that SSL_set_{accept,connect}_state be called before
10284 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10285 in many applications because usually everything *appeared* to work as
10286 intended anyway -- now it really works as intended).
10289 *) Move openssl.cnf out of lib/.
10292 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10293 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10294 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10295 [Ralf S. Engelschall]
10297 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10298 handle PKCS#7 enveloped data properly.
10299 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10301 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10302 copying pointers. The cert_st handling is changed by this in
10303 various ways (and thus what used to be known as ctx->default_cert
10304 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10305 any longer when s->cert does not give us what we need).
10306 ssl_cert_instantiate becomes obsolete by this change.
10307 As soon as we've got the new code right (possibly it already is?),
10308 we have solved a couple of bugs of the earlier code where s->cert
10309 was used as if it could not have been shared with other SSL structures.
10311 Note that using the SSL API in certain dirty ways now will result
10312 in different behaviour than observed with earlier library versions:
10313 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10314 does not influence s as it used to.
10316 In order to clean up things more thoroughly, inside SSL_SESSION
10317 we don't use CERT any longer, but a new structure SESS_CERT
10318 that holds per-session data (if available); currently, this is
10319 the peer's certificate chain and, for clients, the server's certificate
10320 and temporary key. CERT holds only those values that can have
10321 meaningful defaults in an SSL_CTX.
10324 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10325 from the internal representation. Various PKCS#7 fixes: remove some
10326 evil casts and set the enc_dig_alg field properly based on the signing
10330 *) Allow PKCS#12 password to be set from the command line or the
10331 environment. Let 'ca' get its config file name from the environment
10332 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10336 *) Allow certificate policies extension to use an IA5STRING for the
10337 organization field. This is contrary to the PKIX definition but
10338 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10342 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10343 without disallowing inline assembler and the like for non-pedantic builds.
10346 *) Support Borland C++ builder.
10347 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10349 *) Support Mingw32.
10352 *) SHA-1 cleanups and performance enhancements.
10353 [Andy Polyakov <appro@fy.chalmers.se>]
10355 *) Sparc v8plus assembler for the bignum library.
10356 [Andy Polyakov <appro@fy.chalmers.se>]
10358 *) Accept any -xxx and +xxx compiler options in Configure.
10361 *) Update HPUX configuration.
10364 *) Add missing sk_<type>_unshift() function to safestack.h
10365 [Ralf S. Engelschall]
10367 *) New function SSL_CTX_use_certificate_chain_file that sets the
10368 "extra_cert"s in addition to the certificate. (This makes sense
10369 only for "PEM" format files, as chains as a whole are not
10373 *) Support verify_depth from the SSL API.
10374 x509_vfy.c had what can be considered an off-by-one-error:
10375 Its depth (which was not part of the external interface)
10376 was actually counting the number of certificates in a chain;
10377 now it really counts the depth.
10380 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10381 instead of X509err, which often resulted in confusing error
10382 messages since the error codes are not globally unique
10383 (e.g. an alleged error in ssl3_accept when a certificate
10384 didn't match the private key).
10386 *) New function SSL_CTX_set_session_id_context that allows to set a default
10387 value (so that you don't need SSL_set_session_id_context for each
10388 connection using the SSL_CTX).
10391 *) OAEP decoding bug fix.
10394 *) Support INSTALL_PREFIX for package builders, as proposed by
10398 *) New Configure options "threads" and "no-threads". For systems
10399 where the proper compiler options are known (currently Solaris
10400 and Linux), "threads" is the default.
10403 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10406 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10407 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10408 such as /usr/local/bin.
10411 *) "make linux-shared" to build shared libraries.
10412 [Niels Poppe <niels@netbox.org>]
10414 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10417 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10418 extension adding in x509 utility.
10421 *) Remove NOPROTO sections and error code comments.
10424 *) Partial rewrite of the DEF file generator to now parse the ANSI
10428 *) New Configure options --prefix=DIR and --openssldir=DIR.
10431 *) Complete rewrite of the error code script(s). It is all now handled
10432 by one script at the top level which handles error code gathering,
10433 header rewriting and C source file generation. It should be much better
10434 than the old method: it now uses a modified version of Ulf's parser to
10435 read the ANSI prototypes in all header files (thus the old K&R definitions
10436 aren't needed for error creation any more) and do a better job of
10437 translating function codes into names. The old 'ASN1 error code imbedded
10438 in a comment' is no longer necessary and it doesn't use .err files which
10439 have now been deleted. Also the error code call doesn't have to appear all
10440 on one line (which resulted in some large lines...).
10443 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10446 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10447 0 (which usually indicates a closed connection), but continue reading.
10450 *) Fix some race conditions.
10453 *) Add support for CRL distribution points extension. Add Certificate
10454 Policies and CRL distribution points documentation.
10457 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10460 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10461 8 of keying material. Merlin has also confirmed interop with this fix
10462 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10463 [Merlin Hughes <merlin@baltimore.ie>]
10465 *) Fix lots of warnings.
10466 [Richard Levitte <levitte@stacken.kth.se>]
10468 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10469 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10470 [Richard Levitte <levitte@stacken.kth.se>]
10472 *) Fix problems with sizeof(long) == 8.
10473 [Andy Polyakov <appro@fy.chalmers.se>]
10475 *) Change functions to ANSI C.
10478 *) Fix typos in error codes.
10479 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10481 *) Remove defunct assembler files from Configure.
10484 *) SPARC v8 assembler BIGNUM implementation.
10485 [Andy Polyakov <appro@fy.chalmers.se>]
10487 *) Support for Certificate Policies extension: both print and set.
10488 Various additions to support the r2i method this uses.
10491 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10492 return a const string when you are expecting an allocated buffer.
10495 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10496 types DirectoryString and DisplayText.
10499 *) Add code to allow r2i extensions to access the configuration database,
10500 add an LHASH database driver and add several ctx helper functions.
10503 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10504 fail when they extended the size of a BIGNUM.
10507 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10508 support typesafe stack.
10511 *) Fix typo in SSL_[gs]et_options().
10512 [Nils Frostberg <nils@medcom.se>]
10514 *) Delete various functions and files that belonged to the (now obsolete)
10515 old X509V3 handling code.
10518 *) New Configure option "rsaref".
10521 *) Don't auto-generate pem.h.
10524 *) Introduce type-safe ASN.1 SETs.
10527 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10528 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10530 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10531 that links with OpenSSL (well at least cause lots of warnings), but fear
10532 not: the conversion is trivial, and it eliminates loads of evil casts. A
10533 few STACKed things have been converted already. Feel free to convert more.
10534 In the fullness of time, I'll do away with the STACK type altogether.
10537 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10538 specified in <certfile> by updating the entry in the index.txt file.
10539 This way one no longer has to edit the index.txt file manually for
10540 revoking a certificate. The -revoke option does the gory details now.
10541 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10543 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10544 `-text' option at all and this way the `-noout -text' combination was
10545 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10546 [Ralf S. Engelschall]
10548 *) Make sure a corresponding plain text error message exists for the
10549 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10550 verify callback function determined that a certificate was revoked.
10551 [Ralf S. Engelschall]
10553 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10554 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10555 all available cipers including rc5, which was forgotten until now.
10556 In order to let the testing shell script know which algorithms
10557 are available, a new (up to now undocumented) command
10558 "openssl list-cipher-commands" is used.
10561 *) Bugfix: s_client occasionally would sleep in select() when
10562 it should have checked SSL_pending() first.
10565 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10566 the raw DSA values prior to ASN.1 encoding.
10569 *) Tweaks to Configure
10570 [Niels Poppe <niels@netbox.org>]
10572 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10576 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10579 *) New config option to avoid instructions that are illegal on the 80386.
10580 The default code is faster, but requires at least a 486.
10583 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10584 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10585 same as SSL2_VERSION anyway.
10588 *) New "-showcerts" option for s_client.
10591 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10592 application. Various cleanups and fixes.
10595 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10596 modify error routines to work internally. Add error codes and PBE init
10597 to library startup routines.
10600 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10601 packing functions to asn1 and evp. Changed function names and error
10602 codes along the way.
10605 *) PKCS12 integration: and so it begins... First of several patches to
10606 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10607 objects to objects.h
10610 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10611 and display support for Thawte strong extranet extension.
10614 *) Add LinuxPPC support.
10615 [Jeff Dubrule <igor@pobox.org>]
10617 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10618 bn_div_words in alpha.s.
10619 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10621 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10622 OAEP isn't supported when OpenSSL is built with RSAref.
10623 [Ulf Moeller <ulf@fitug.de>]
10625 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10626 so they no longer are missing under -DNOPROTO.
10627 [Soren S. Jorvang <soren@t.dk>]
10630 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10632 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10633 doesn't work when the session is reused. Coming soon!
10636 *) Fix a security hole, that allows sessions to be reused in the wrong
10637 context thus bypassing client cert protection! All software that uses
10638 client certs and session caches in multiple contexts NEEDS PATCHING to
10639 allow session reuse! A fuller solution is in the works.
10640 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10642 *) Some more source tree cleanups (removed obsolete files
10643 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10644 permission on "config" script to be executable) and a fix for the INSTALL
10646 [Ulf Moeller <ulf@fitug.de>]
10648 *) Remove some legacy and erroneous uses of malloc, free instead of
10650 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10652 *) Make rsa_oaep_test return non-zero on error.
10653 [Ulf Moeller <ulf@fitug.de>]
10655 *) Add support for native Solaris shared libraries. Configure
10656 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10657 if someone would make that last step automatic.
10658 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10660 *) ctx_size was not built with the right compiler during "make links". Fixed.
10663 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10664 except NULL ciphers". This means the default cipher list will no longer
10665 enable NULL ciphers. They need to be specifically enabled e.g. with
10666 the string "DEFAULT:eNULL".
10669 *) Fix to RSA private encryption routines: if p < q then it would
10670 occasionally produce an invalid result. This will only happen with
10671 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10674 *) Be less restrictive and allow also `perl util/perlpath.pl
10675 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10676 because this way one can also use an interpreter named `perl5' (which is
10677 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10678 installed as `perl').
10679 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10681 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10682 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10684 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10685 advapi32.lib to Win32 build and change the pem test comparision
10686 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10687 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10688 and crypto/des/ede_cbcm_enc.c.
10691 *) DES quad checksum was broken on big-endian architectures. Fixed.
10694 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10695 Win32 test batch file so it (might) work again. The Win32 test batch file
10696 is horrible: I feel ill....
10699 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10700 in e_os.h. Audit of header files to check ANSI and non ANSI
10701 sections: 10 functions were absent from non ANSI section and not exported
10702 from Windows DLLs. Fixed up libeay.num for new functions.
10705 *) Make `openssl version' output lines consistent.
10706 [Ralf S. Engelschall]
10708 *) Fix Win32 symbol export lists for BIO functions: Added
10709 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10710 to ms/libeay{16,32}.def.
10711 [Ralf S. Engelschall]
10713 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10714 fine under Unix and passes some trivial tests I've now added. But the
10715 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10716 added to make sure no one expects that this stuff really works in the
10717 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10718 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10720 [Ralf S. Engelschall]
10722 *) Fix the generation of two part addresses in perl.
10723 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10725 *) Add config entry for Linux on MIPS.
10726 [John Tobey <jtobey@channel1.com>]
10728 *) Make links whenever Configure is run, unless we are on Windoze.
10731 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10732 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10736 *) Add a useful kludge to allow package maintainers to specify compiler and
10737 other platforms details on the command line without having to patch the
10738 Configure script everytime: One now can use ``perl Configure
10739 <id>:<details>'', i.e. platform ids are allowed to have details appended
10740 to them (seperated by colons). This is treated as there would be a static
10741 pre-configured entry in Configure's %table under key <id> with value
10742 <details> and ``perl Configure <id>'' is called. So, when you want to
10743 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10744 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10745 now, which overrides the FreeBSD-elf entry on-the-fly.
10746 [Ralf S. Engelschall]
10748 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10751 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10752 on the `perl Configure ...' command line. This way one can compile
10753 OpenSSL libraries with Position Independent Code (PIC) which is needed
10754 for linking it into DSOs.
10755 [Ralf S. Engelschall]
10757 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10761 *) Cleaned up the LICENSE document: The official contact for any license
10762 questions now is the OpenSSL core team under openssl-core@openssl.org.
10763 And add a paragraph about the dual-license situation to make sure people
10764 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10765 to the OpenSSL toolkit.
10766 [Ralf S. Engelschall]
10768 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10769 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10770 Additonally cleaned up the `make links' target: Remove unnecessary
10771 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10772 to speed processing and no longer clutter the display with confusing
10773 stuff. Instead only the actually done links are displayed.
10774 [Ralf S. Engelschall]
10776 *) Permit null encryption ciphersuites, used for authentication only. It used
10777 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10778 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10782 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10783 signed attributes when verifying signatures (this would break them),
10784 the detached data encoding was wrong and public keys obtained using
10785 X509_get_pubkey() weren't freed.
10788 *) Add text documentation for the BUFFER functions. Also added a work around
10789 to a Win95 console bug. This was triggered by the password read stuff: the
10790 last character typed gets carried over to the next fread(). If you were
10791 generating a new cert request using 'req' for example then the last
10792 character of the passphrase would be CR which would then enter the first
10796 *) Added the new `Includes OpenSSL Cryptography Software' button as
10797 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10798 button and can be used by applications based on OpenSSL to show the
10799 relationship to the OpenSSL project.
10800 [Ralf S. Engelschall]
10802 *) Remove confusing variables in function signatures in files
10803 ssl/ssl_lib.c and ssl/ssl.h.
10804 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10806 *) Don't install bss_file.c under PREFIX/include/
10807 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10809 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10810 functions that return function pointers and has support for NT specific
10811 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10812 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10813 unsigned to signed types: this was killing the Win32 compile.
10816 *) Add new certificate file to stack functions,
10817 SSL_add_dir_cert_subjects_to_stack() and
10818 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10819 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10820 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10821 This means that Apache-SSL and similar packages don't have to mess around
10822 to add as many CAs as they want to the preferred list.
10825 *) Experiment with doxygen documentation. Currently only partially applied to
10827 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10828 openssl.doxy as the configuration file.
10831 *) Get rid of remaining C++-style comments which strict C compilers hate.
10832 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10834 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10835 compiled in by default: it has problems with large keys.
10838 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10839 DH private keys and/or callback functions which directly correspond to
10840 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10841 is needed for applications which have to configure certificates on a
10842 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10844 For the RSA certificate situation is makes no difference, but
10845 for the DSA certificate situation this fixes the "no shared cipher"
10846 problem where the OpenSSL cipher selection procedure failed because the
10847 temporary keys were not overtaken from the context and the API provided
10848 no way to reconfigure them.
10849 The new functions now let applications reconfigure the stuff and they
10850 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10851 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10852 non-public-API function ssl_cert_instantiate() is used as a helper
10853 function and also to reduce code redundancy inside ssl_rsa.c.
10854 [Ralf S. Engelschall]
10856 *) Move s_server -dcert and -dkey options out of the undocumented feature
10857 area because they are useful for the DSA situation and should be
10858 recognized by the users.
10859 [Ralf S. Engelschall]
10861 *) Fix the cipher decision scheme for export ciphers: the export bits are
10862 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10863 SSL_EXP_MASK. So, the original variable has to be used instead of the
10864 already masked variable.
10865 [Richard Levitte <levitte@stacken.kth.se>]
10867 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10868 [Richard Levitte <levitte@stacken.kth.se>]
10870 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10871 from `int' to `unsigned int' because it's a length and initialized by
10872 EVP_DigestFinal() which expects an `unsigned int *'.
10873 [Richard Levitte <levitte@stacken.kth.se>]
10875 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10876 script. Instead use the usual Shell->Perl transition trick.
10877 [Ralf S. Engelschall]
10879 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10880 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10881 -noout -modulus' as it's already the case for `openssl rsa -noout
10882 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10883 currently the public key is printed (a decision which was already done by
10884 `openssl dsa -modulus' in the past) which serves a similar purpose.
10885 Additionally the NO_RSA no longer completely removes the whole -modulus
10886 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10888 [Ralf S. Engelschall]
10890 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10891 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10892 [Arne Ansper <arne@ats.cyber.ee>]
10894 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10895 to be added. Now both 'req' and 'ca' can use new objects defined in the
10899 *) Add cool BIO that does syslog (or event log on NT).
10900 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10902 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10903 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10904 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10905 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10908 *) Add preliminary config info for new extension code.
10911 *) Make RSA_NO_PADDING really use no padding.
10912 [Ulf Moeller <ulf@fitug.de>]
10914 *) Generate errors when private/public key check is done.
10917 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10918 for some CRL extensions and new objects added.
10921 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10922 key usage extension and fuller support for authority key id.
10925 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10926 padding method for RSA, which is recommended for new applications in PKCS
10927 #1 v2.0 (RFC 2437, October 1998).
10928 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10929 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10930 against Bleichbacher's attack on RSA.
10931 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10934 *) Updates to the new SSL compression code
10935 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10937 *) Fix so that the version number in the master secret, when passed
10938 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10939 (because the server will not accept higher), that the version number
10940 is 0x03,0x01, not 0x03,0x00
10941 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10943 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10944 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10945 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10948 *) Support for RAW extensions where an arbitrary extension can be
10949 created by including its DER encoding. See apps/openssl.cnf for
10953 *) Make sure latest Perl versions don't interpret some generated C array
10954 code as Perl array code in the crypto/err/err_genc.pl script.
10955 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10957 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10958 not many people have the assembler. Various Win32 compilation fixes and
10959 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10960 build instructions.
10963 *) Modify configure script 'Configure' to automatically create crypto/date.h
10964 file under Win32 and also build pem.h from pem.org. New script
10965 util/mkfiles.pl to create the MINFO file on environments that can't do a
10966 'make files': perl util/mkfiles.pl >MINFO should work.
10969 *) Major rework of DES function declarations, in the pursuit of correctness
10970 and purity. As a result, many evil casts evaporated, and some weirdness,
10971 too. You may find this causes warnings in your code. Zapping your evil
10972 casts will probably fix them. Mostly.
10975 *) Fix for a typo in asn1.h. Bug fix to object creation script
10976 obj_dat.pl. It considered a zero in an object definition to mean
10977 "end of object": none of the objects in objects.h have any zeros
10978 so it wasn't spotted.
10979 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10981 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10982 Masking (CBCM). In the absence of test vectors, the best I have been able
10983 to do is check that the decrypt undoes the encrypt, so far. Send me test
10984 vectors if you have them.
10987 *) Correct calculation of key length for export ciphers (too much space was
10988 allocated for null ciphers). This has not been tested!
10991 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10992 message is now correct (it understands "crypto" and "ssl" on its
10993 command line). There is also now an "update" option. This will update
10994 the util/ssleay.num and util/libeay.num files with any new functions.
10996 perl util/mkdef.pl crypto ssl update
10997 it will update them.
11000 *) Overhauled the Perl interface (perl/*):
11001 - ported BN stuff to OpenSSL's different BN library
11002 - made the perl/ source tree CVS-aware
11003 - renamed the package from SSLeay to OpenSSL (the files still contain
11004 their history because I've copied them in the repository)
11005 - removed obsolete files (the test scripts will be replaced
11006 by better Test::Harness variants in the future)
11007 [Ralf S. Engelschall]
11009 *) First cut for a very conservative source tree cleanup:
11010 1. merge various obsolete readme texts into doc/ssleay.txt
11011 where we collect the old documents and readme texts.
11012 2. remove the first part of files where I'm already sure that we no
11013 longer need them because of three reasons: either they are just temporary
11014 files which were left by Eric or they are preserved original files where
11015 I've verified that the diff is also available in the CVS via "cvs diff
11016 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11017 the crypto/md/ stuff).
11018 [Ralf S. Engelschall]
11020 *) More extension code. Incomplete support for subject and issuer alt
11021 name, issuer and authority key id. Change the i2v function parameters
11022 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11023 what that's for :-) Fix to ASN1 macro which messed up
11024 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11027 *) Preliminary support for ENUMERATED type. This is largely copied from the
11031 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11032 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11034 *) Make sure `make rehash' target really finds the `openssl' program.
11035 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11037 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11038 like to hear about it if this slows down other processors.
11041 *) Add CygWin32 platform information to Configure script.
11042 [Alan Batie <batie@aahz.jf.intel.com>]
11044 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11045 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11047 *) New program nseq to manipulate netscape certificate sequences
11050 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11054 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11055 but the BN code had some problems that would cause failures when
11056 doing certificate verification and some other functions.
11057 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11059 *) Add ASN1 and PEM code to support netscape certificate sequences.
11062 *) Add ASN1 and PEM code to support netscape certificate sequences.
11065 *) Add several PKIX and private extended key usage OIDs.
11068 *) Modify the 'ca' program to handle the new extension code. Modify
11069 openssl.cnf for new extension format, add comments.
11072 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11073 and add a sample to openssl.cnf so req -x509 now adds appropriate
11077 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11078 error code, add initial support to X509_print() and x509 application.
11081 *) Takes a deep breath and start addding X509 V3 extension support code. Add
11082 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11083 stuff is currently isolated and isn't even compiled yet.
11086 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11087 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11088 Removed the versions check from X509 routines when loading extensions:
11089 this allows certain broken certificates that don't set the version
11090 properly to be processed.
11093 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11094 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11095 can still be regenerated with "make depend".
11098 *) Spelling mistake in C version of CAST-128.
11099 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11101 *) Changes to the error generation code. The perl script err-code.pl
11102 now reads in the old error codes and retains the old numbers, only
11103 adding new ones if necessary. It also only changes the .err files if new
11104 codes are added. The makefiles have been modified to only insert errors
11105 when needed (to avoid needlessly modifying header files). This is done
11106 by only inserting errors if the .err file is newer than the auto generated
11107 C file. To rebuild all the error codes from scratch (the old behaviour)
11108 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11109 or delete all the .err files.
11112 *) CAST-128 was incorrectly implemented for short keys. The C version has
11113 been fixed, but is untested. The assembler versions are also fixed, but
11114 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11115 to regenerate it if needed.
11116 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11117 Hagino <itojun@kame.net>]
11119 *) File was opened incorrectly in randfile.c.
11120 [Ulf Möller <ulf@fitug.de>]
11122 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11123 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11124 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11125 al: it's just almost always a UTCTime. Note this patch adds new error
11126 codes so do a "make errors" if there are problems.
11129 *) Correct Linux 1 recognition in config.
11130 [Ulf Möller <ulf@fitug.de>]
11132 *) Remove pointless MD5 hash when using DSA keys in ca.
11133 [Anonymous <nobody@replay.com>]
11135 *) Generate an error if given an empty string as a cert directory. Also
11136 generate an error if handed NULL (previously returned 0 to indicate an
11137 error, but didn't set one).
11138 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11140 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11143 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11144 parameters. This was causing a warning which killed off the Win32 compile.
11147 *) Remove C++ style comments from crypto/bn/bn_local.h.
11148 [Neil Costigan <neil.costigan@celocom.com>]
11150 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11151 based on a text string, looking up short and long names and finally
11152 "dot" format. The "dot" format stuff didn't work. Added new function
11153 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11154 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11155 OID is not part of the table.
11158 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11159 X509_LOOKUP_by_alias().
11162 *) Sort openssl functions by name.
11165 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11166 encryption from sample DSA keys (in case anyone is interested the password
11170 *) Make _all_ *_free functions accept a NULL pointer.
11171 [Frans Heymans <fheymans@isaserver.be>]
11173 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11175 [Anonymous <nobody@replay.com>]
11177 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11178 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11180 *) Don't blow it for numeric -newkey arguments to apps/req.
11181 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11183 *) Temp key "for export" tests were wrong in s3_srvr.c.
11184 [Anonymous <nobody@replay.com>]
11186 *) Add prototype for temp key callback functions
11187 SSL_CTX_set_tmp_{rsa,dh}_callback().
11190 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11191 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11194 *) X509_name_add_entry() freed the wrong thing after an error.
11195 [Arne Ansper <arne@ats.cyber.ee>]
11197 *) rsa_eay.c would attempt to free a NULL context.
11198 [Arne Ansper <arne@ats.cyber.ee>]
11200 *) BIO_s_socket() had a broken should_retry() on Windoze.
11201 [Arne Ansper <arne@ats.cyber.ee>]
11203 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11204 [Arne Ansper <arne@ats.cyber.ee>]
11206 *) Make sure the already existing X509_STORE->depth variable is initialized
11207 in X509_STORE_new(), but document the fact that this variable is still
11208 unused in the certificate verification process.
11209 [Ralf S. Engelschall]
11211 *) Fix the various library and apps files to free up pkeys obtained from
11212 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11215 *) Fix reference counting in X509_PUBKEY_get(). This makes
11216 demos/maurice/example2.c work, amongst others, probably.
11217 [Steve Henson and Ben Laurie]
11219 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11220 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11221 are no longer created. This way we have a single and consistent command
11222 line interface `openssl <command>', similar to `cvs <command>'.
11223 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11225 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11226 BIT STRING wrapper always have zero unused bits.
11229 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11232 *) Make the top-level INSTALL documentation easier to understand.
11235 *) Makefiles updated to exit if an error occurs in a sub-directory
11236 make (including if user presses ^C) [Paul Sutton]
11238 *) Make Montgomery context stuff explicit in RSA data structure.
11241 *) Fix build order of pem and err to allow for generated pem.h.
11244 *) Fix renumbering bug in X509_NAME_delete_entry().
11247 *) Enhanced the err-ins.pl script so it makes the error library number
11248 global and can add a library name. This is needed for external ASN1 and
11249 other error libraries.
11252 *) Fixed sk_insert which never worked properly.
11255 *) Fix ASN1 macros so they can handle indefinite length construted
11256 EXPLICIT tags. Some non standard certificates use these: they can now
11260 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11261 into a single doc/ssleay.txt bundle. This way the information is still
11262 preserved but no longer messes up this directory. Now it's new room for
11263 the new set of documenation files.
11264 [Ralf S. Engelschall]
11266 *) SETs were incorrectly DER encoded. This was a major pain, because they
11267 shared code with SEQUENCEs, which aren't coded the same. This means that
11268 almost everything to do with SETs or SEQUENCEs has either changed name or
11269 number of arguments.
11270 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11272 *) Fix test data to work with the above.
11275 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11276 was already fixed by Eric for 0.9.1 it seems.
11277 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11279 *) Autodetect FreeBSD3.
11282 *) Fix various bugs in Configure. This affects the following platforms:
11286 unixware-2.0-pentium
11290 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11291 before they are needed.
11294 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11298 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11300 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11301 changed SSLeay to OpenSSL in version strings.
11302 [Ralf S. Engelschall]
11304 *) Some fixups to the top-level documents.
11307 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11308 because the symlink to include/ was missing.
11309 [Ralf S. Engelschall]
11311 *) Incorporated the popular no-RSA/DSA-only patches
11312 which allow to compile a RSA-free SSLeay.
11313 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11315 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11316 when "ssleay" is still not found.
11317 [Ralf S. Engelschall]
11319 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11320 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11322 *) Updated the README file.
11323 [Ralf S. Engelschall]
11325 *) Added various .cvsignore files in the CVS repository subdirs
11326 to make a "cvs update" really silent.
11327 [Ralf S. Engelschall]
11329 *) Recompiled the error-definition header files and added
11330 missing symbols to the Win32 linker tables.
11331 [Ralf S. Engelschall]
11333 *) Cleaned up the top-level documents;
11334 o new files: CHANGES and LICENSE
11335 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11336 o merged COPYRIGHT into LICENSE
11337 o removed obsolete TODO file
11338 o renamed MICROSOFT to INSTALL.W32
11339 [Ralf S. Engelschall]
11341 *) Removed dummy files from the 0.9.1b source tree:
11342 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11343 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11344 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11345 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11346 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11347 [Ralf S. Engelschall]
11349 *) Added various platform portability fixes.
11352 *) The Genesis of the OpenSSL rpject:
11353 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11354 Young and Tim J. Hudson created while they were working for C2Net until
11356 [The OpenSSL Project]
11359 Changes between 0.9.0b and 0.9.1b [not released]
11361 *) Updated a few CA certificates under certs/
11364 *) Changed some BIGNUM api stuff.
11367 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11368 DGUX x86, Linux Alpha, etc.
11371 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11372 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11376 *) Add -strparse option to asn1pars program which parses nested
11378 [Dr Stephen Henson <shenson@bigfoot.com>]
11380 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11383 *) DSA fix for "ca" program.
11386 *) Added "-genkey" option to "dsaparam" program.
11389 *) Added RIPE MD160 (rmd160) message digest.
11392 *) Added -a (all) option to "ssleay version" command.
11395 *) Added PLATFORM define which is the id given to Configure.
11398 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11401 *) Extended the ASN.1 parser routines.
11404 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11407 *) Added a BN_CTX to the BN library.
11410 *) Fixed the weak key values in DES library
11413 *) Changed API in EVP library for cipher aliases.
11416 *) Added support for RC2/64bit cipher.
11419 *) Converted the lhash library to the crypto/mem.c functions.
11422 *) Added more recognized ASN.1 object ids.
11425 *) Added more RSA padding checks for SSL/TLS.
11428 *) Added BIO proxy/filter functionality.
11431 *) Added extra_certs to SSL_CTX which can be used
11432 send extra CA certificates to the client in the CA cert chain sending
11433 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11436 *) Now Fortezza is denied in the authentication phase because
11437 this is key exchange mechanism is not supported by SSLeay at all.
11440 *) Additional PKCS1 checks.
11443 *) Support the string "TLSv1" for all TLS v1 ciphers.
11446 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11447 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11450 *) Fixed a few memory leaks.
11453 *) Fixed various code and comment typos.
11456 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11457 bytes sent in the client random.
11458 [Edward Bishop <ebishop@spyglass.com>]