5 Changes between 1.0.2k and 1.0.2l [xx XXX xxxx]
9 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
10 platform rather than 'mingw'.
13 Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
15 *) Truncated packet could crash via OOB read
17 If one side of an SSL/TLS path is running on a 32-bit host and a specific
18 cipher is being used, then a truncated packet can cause that host to
19 perform an out-of-bounds read, usually resulting in a crash.
21 This issue was reported to OpenSSL by Robert Święcki of Google.
25 *) BN_mod_exp may produce incorrect results on x86_64
27 There is a carry propagating bug in the x86_64 Montgomery squaring
28 procedure. No EC algorithms are affected. Analysis suggests that attacks
29 against RSA and DSA as a result of this defect would be very difficult to
30 perform and are not believed likely. Attacks against DH are considered just
31 feasible (although very difficult) because most of the work necessary to
32 deduce information about a private key may be performed offline. The amount
33 of resources required for such an attack would be very significant and
34 likely only accessible to a limited number of attackers. An attacker would
35 additionally need online access to an unpatched system using the target
36 private key in a scenario with persistent DH parameters and a private
37 key that is shared between multiple clients. For example this can occur by
38 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
39 similar to CVE-2015-3193 but must be treated as a separate problem.
41 This issue was reported to OpenSSL by the OSS-Fuzz project.
45 *) Montgomery multiplication may produce incorrect results
47 There is a carry propagating bug in the Broadwell-specific Montgomery
48 multiplication procedure that handles input lengths divisible by, but
49 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
50 and DH private keys are impossible. This is because the subroutine in
51 question is not used in operations with the private key itself and an input
52 of the attacker's direct choice. Otherwise the bug can manifest itself as
53 transient authentication and key negotiation failures or reproducible
54 erroneous outcome of public-key operations with specially crafted input.
55 Among EC algorithms only Brainpool P-512 curves are affected and one
56 presumably can attack ECDH key negotiation. Impact was not analyzed in
57 detail, because pre-requisites for attack are considered unlikely. Namely
58 multiple clients have to choose the curve in question and the server has to
59 share the private key among them, neither of which is default behaviour.
60 Even then only clients that chose the curve will be affected.
62 This issue was publicly reported as transient failures and was not
63 initially recognized as a security issue. Thanks to Richard Morgan for
64 providing reproducible case.
68 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
69 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
70 prevent issues where no progress is being made and the peer continually
71 sends unrecognised record types, using up resources processing them.
74 Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
76 *) Missing CRL sanity check
78 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
79 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
80 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
82 This issue only affects the OpenSSL 1.0.2i
86 Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
88 *) OCSP Status Request extension unbounded memory growth
90 A malicious client can send an excessively large OCSP Status Request
91 extension. If that client continually requests renegotiation, sending a
92 large OCSP Status Request extension each time, then there will be unbounded
93 memory growth on the server. This will eventually lead to a Denial Of
94 Service attack through memory exhaustion. Servers with a default
95 configuration are vulnerable even if they do not support OCSP. Builds using
96 the "no-ocsp" build time option are not affected.
98 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
102 *) In order to mitigate the SWEET32 attack, the DES ciphers were moved from
105 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
110 *) OOB write in MDC2_Update()
112 An overflow can occur in MDC2_Update() either if called directly or
113 through the EVP_DigestUpdate() function using MDC2. If an attacker
114 is able to supply very large amounts of input data after a previous
115 call to EVP_EncryptUpdate() with a partial block then a length check
116 can overflow resulting in a heap corruption.
118 The amount of data needed is comparable to SIZE_MAX which is impractical
121 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
125 *) Malformed SHA512 ticket DoS
127 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
128 DoS attack where a malformed ticket will result in an OOB read which will
131 The use of SHA512 in TLS session tickets is comparatively rare as it requires
132 a custom server callback and ticket lookup mechanism.
134 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
138 *) OOB write in BN_bn2dec()
140 The function BN_bn2dec() does not check the return value of BN_div_word().
141 This can cause an OOB write if an application uses this function with an
142 overly large BIGNUM. This could be a problem if an overly large certificate
143 or CRL is printed out from an untrusted source. TLS is not affected because
144 record limits will reject an oversized certificate before it is parsed.
146 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
150 *) OOB read in TS_OBJ_print_bio()
152 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
153 the total length the OID text representation would use and not the amount
154 of data written. This will result in OOB reads when large OIDs are
157 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
161 *) Pointer arithmetic undefined behaviour
163 Avoid some undefined pointer arithmetic
165 A common idiom in the codebase is to check limits in the following manner:
168 Where "p" points to some malloc'd data of SIZE bytes and
171 "len" here could be from some externally supplied data (e.g. from a TLS
174 The rules of C pointer arithmetic are such that "p + len" is only well
175 defined where len <= SIZE. Therefore the above idiom is actually
178 For example this could cause problems if some malloc implementation
179 provides an address for "p" such that "p + len" actually overflows for
180 values of len that are too big and therefore p + len < limit.
182 This issue was reported to OpenSSL by Guido Vranken
186 *) Constant time flag not preserved in DSA signing
188 Operations in the DSA signing algorithm should run in constant time in
189 order to avoid side channel attacks. A flaw in the OpenSSL DSA
190 implementation means that a non-constant time codepath is followed for
191 certain operations. This has been demonstrated through a cache-timing
192 attack to be sufficient for an attacker to recover the private DSA key.
194 This issue was reported by César Pereida (Aalto University), Billy Brumley
195 (Tampere University of Technology), and Yuval Yarom (The University of
200 *) DTLS buffered message DoS
202 In a DTLS connection where handshake messages are delivered out-of-order
203 those messages that OpenSSL is not yet ready to process will be buffered
204 for later use. Under certain circumstances, a flaw in the logic means that
205 those messages do not get removed from the buffer even though the handshake
206 has been completed. An attacker could force up to approx. 15 messages to
207 remain in the buffer when they are no longer required. These messages will
208 be cleared when the DTLS connection is closed. The default maximum size for
209 a message is 100k. Therefore the attacker could force an additional 1500k
210 to be consumed per connection. By opening many simulataneous connections an
211 attacker could cause a DoS attack through memory exhaustion.
213 This issue was reported to OpenSSL by Quan Luo.
217 *) DTLS replay protection DoS
219 A flaw in the DTLS replay attack protection mechanism means that records
220 that arrive for future epochs update the replay protection "window" before
221 the MAC for the record has been validated. This could be exploited by an
222 attacker by sending a record for the next epoch (which does not have to
223 decrypt or have a valid MAC), with a very large sequence number. This means
224 that all subsequent legitimate packets are dropped causing a denial of
225 service for a specific DTLS connection.
227 This issue was reported to OpenSSL by the OCAP audit team.
231 *) Certificate message OOB reads
233 In OpenSSL 1.0.2 and earlier some missing message length checks can result
234 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
235 theoretical DoS risk but this has not been observed in practice on common
238 The messages affected are client certificate, client certificate request
239 and server certificate. As a result the attack can only be performed
240 against a client or a server which enables client authentication.
242 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
246 Changes between 1.0.2g and 1.0.2h [3 May 2016]
248 *) Prevent padding oracle in AES-NI CBC MAC check
250 A MITM attacker can use a padding oracle attack to decrypt traffic
251 when the connection uses an AES CBC cipher and the server support
254 This issue was introduced as part of the fix for Lucky 13 padding
255 attack (CVE-2013-0169). The padding check was rewritten to be in
256 constant time by making sure that always the same bytes are read and
257 compared against either the MAC or padding bytes. But it no longer
258 checked that there was enough data to have both the MAC and padding
261 This issue was reported by Juraj Somorovsky using TLS-Attacker.
265 *) Fix EVP_EncodeUpdate overflow
267 An overflow can occur in the EVP_EncodeUpdate() function which is used for
268 Base64 encoding of binary data. If an attacker is able to supply very large
269 amounts of input data then a length check can overflow resulting in a heap
272 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
273 the PEM_write_bio* family of functions. These are mainly used within the
274 OpenSSL command line applications, so any application which processes data
275 from an untrusted source and outputs it as a PEM file should be considered
276 vulnerable to this issue. User applications that call these APIs directly
277 with large amounts of untrusted data may also be vulnerable.
279 This issue was reported by Guido Vranken.
283 *) Fix EVP_EncryptUpdate overflow
285 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
286 is able to supply very large amounts of input data after a previous call to
287 EVP_EncryptUpdate() with a partial block then a length check can overflow
288 resulting in a heap corruption. Following an analysis of all OpenSSL
289 internal usage of the EVP_EncryptUpdate() function all usage is one of two
290 forms. The first form is where the EVP_EncryptUpdate() call is known to be
291 the first called function after an EVP_EncryptInit(), and therefore that
292 specific call must be safe. The second form is where the length passed to
293 EVP_EncryptUpdate() can be seen from the code to be some small value and
294 therefore there is no possibility of an overflow. Since all instances are
295 one of these two forms, it is believed that there can be no overflows in
296 internal code due to this problem. It should be noted that
297 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
298 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
299 of these calls have also been analysed too and it is believed there are no
300 instances in internal usage where an overflow could occur.
302 This issue was reported by Guido Vranken.
306 *) Prevent ASN.1 BIO excessive memory allocation
308 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
309 a short invalid encoding can casuse allocation of large amounts of memory
310 potentially consuming excessive resources or exhausting memory.
312 Any application parsing untrusted data through d2i BIO functions is
313 affected. The memory based functions such as d2i_X509() are *not* affected.
314 Since the memory based functions are used by the TLS library, TLS
315 applications are not affected.
317 This issue was reported by Brian Carpenter.
323 ASN1 Strings that are over 1024 bytes can cause an overread in applications
324 using the X509_NAME_oneline() function on EBCDIC systems. This could result
325 in arbitrary stack data being returned in the buffer.
327 This issue was reported by Guido Vranken.
331 *) Modify behavior of ALPN to invoke callback after SNI/servername
332 callback, such that updates to the SSL_CTX affect ALPN.
335 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
339 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
340 methods are enabled and ssl2 is disabled the methods return NULL.
343 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
345 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
346 Builds that are not configured with "enable-weak-ssl-ciphers" will not
347 provide any "EXPORT" or "LOW" strength ciphers.
350 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
351 is by default disabled at build-time. Builds that are not configured with
352 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
353 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
354 will need to explicitly call either of:
356 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
358 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
360 as appropriate. Even if either of those is used, or the application
361 explicitly uses the version-specific SSLv2_method() or its client and
362 server variants, SSLv2 ciphers vulnerable to exhaustive search key
363 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
364 ciphers, and SSLv2 56-bit DES are no longer available.
368 *) Fix a double-free in DSA code
370 A double free bug was discovered when OpenSSL parses malformed DSA private
371 keys and could lead to a DoS attack or memory corruption for applications
372 that receive DSA private keys from untrusted sources. This scenario is
375 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
380 *) Disable SRP fake user seed to address a server memory leak.
382 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
384 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
385 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
386 was changed to ignore the "fake user" SRP seed, even if the seed
389 Users should use SRP_VBASE_get1_by_user instead. Note that in
390 SRP_VBASE_get1_by_user, caller must free the returned value. Note
391 also that even though configuring the SRP seed attempts to hide
392 invalid usernames by continuing the handshake with fake
393 credentials, this behaviour is not constant time and no strong
394 guarantees are made that the handshake is indistinguishable from
395 that of a valid user.
399 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
401 In the BN_hex2bn function the number of hex digits is calculated using an
402 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
403 large values of |i| this can result in |bn_expand| not allocating any
404 memory because |i * 4| is negative. This can leave the internal BIGNUM data
405 field as NULL leading to a subsequent NULL ptr deref. For very large values
406 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
407 In this case memory is allocated to the internal BIGNUM data field, but it
408 is insufficiently sized leading to heap corruption. A similar issue exists
409 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
410 is ever called by user applications with very large untrusted hex/dec data.
411 This is anticipated to be a rare occurrence.
413 All OpenSSL internal usage of these functions use data that is not expected
414 to be untrusted, e.g. config file data or application command line
415 arguments. If user developed applications generate config file data based
416 on untrusted data then it is possible that this could also lead to security
417 consequences. This is also anticipated to be rare.
419 This issue was reported to OpenSSL by Guido Vranken.
423 *) Fix memory issues in BIO_*printf functions
425 The internal |fmtstr| function used in processing a "%s" format string in
426 the BIO_*printf functions could overflow while calculating the length of a
427 string and cause an OOB read when printing very long strings.
429 Additionally the internal |doapr_outch| function can attempt to write to an
430 OOB memory location (at an offset from the NULL pointer) in the event of a
431 memory allocation failure. In 1.0.2 and below this could be caused where
432 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
433 could be in processing a very long "%s" format string. Memory leaks can
436 The first issue may mask the second issue dependent on compiler behaviour.
437 These problems could enable attacks where large amounts of untrusted data
438 is passed to the BIO_*printf functions. If applications use these functions
439 in this way then they could be vulnerable. OpenSSL itself uses these
440 functions when printing out human-readable dumps of ASN.1 data. Therefore
441 applications that print this data could be vulnerable if the data is from
442 untrusted sources. OpenSSL command line applications could also be
443 vulnerable where they print out ASN.1 data, or if untrusted data is passed
444 as command line arguments.
446 Libssl is not considered directly vulnerable. Additionally certificates etc
447 received via remote connections via libssl are also unlikely to be able to
448 trigger these issues because of message size limits enforced within libssl.
450 This issue was reported to OpenSSL Guido Vranken.
454 *) Side channel attack on modular exponentiation
456 A side-channel attack was found which makes use of cache-bank conflicts on
457 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
458 of RSA keys. The ability to exploit this issue is limited as it relies on
459 an attacker who has control of code in a thread running on the same
460 hyper-threaded core as the victim thread which is performing decryptions.
462 This issue was reported to OpenSSL by Yuval Yarom, The University of
463 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
464 Nadia Heninger, University of Pennsylvania with more information at
465 http://cachebleed.info.
469 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
470 if no keysize is specified with default_bits. This fixes an
471 omission in an earlier change that changed all RSA/DSA key generation
472 apps to use 2048 bits by default.
475 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
477 *) DH small subgroups
479 Historically OpenSSL only ever generated DH parameters based on "safe"
480 primes. More recently (in version 1.0.2) support was provided for
481 generating X9.42 style parameter files such as those required for RFC 5114
482 support. The primes used in such files may not be "safe". Where an
483 application is using DH configured with parameters based on primes that are
484 not "safe" then an attacker could use this fact to find a peer's private
485 DH exponent. This attack requires that the attacker complete multiple
486 handshakes in which the peer uses the same private DH exponent. For example
487 this could be used to discover a TLS server's private DH exponent if it's
488 reusing the private DH exponent or it's using a static DH ciphersuite.
490 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
491 TLS. It is not on by default. If the option is not set then the server
492 reuses the same private DH exponent for the life of the server process and
493 would be vulnerable to this attack. It is believed that many popular
494 applications do set this option and would therefore not be at risk.
496 The fix for this issue adds an additional check where a "q" parameter is
497 available (as is the case in X9.42 based parameters). This detects the
498 only known attack, and is the only possible defense for static DH
499 ciphersuites. This could have some performance impact.
501 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
502 default and cannot be disabled. This could have some performance impact.
504 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
508 *) SSLv2 doesn't block disabled ciphers
510 A malicious client can negotiate SSLv2 ciphers that have been disabled on
511 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
512 been disabled, provided that the SSLv2 protocol was not also disabled via
515 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
516 and Sebastian Schinzel.
520 *) Reject DH handshakes with parameters shorter than 1024 bits.
523 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
525 *) BN_mod_exp may produce incorrect results on x86_64
527 There is a carry propagating bug in the x86_64 Montgomery squaring
528 procedure. No EC algorithms are affected. Analysis suggests that attacks
529 against RSA and DSA as a result of this defect would be very difficult to
530 perform and are not believed likely. Attacks against DH are considered just
531 feasible (although very difficult) because most of the work necessary to
532 deduce information about a private key may be performed offline. The amount
533 of resources required for such an attack would be very significant and
534 likely only accessible to a limited number of attackers. An attacker would
535 additionally need online access to an unpatched system using the target
536 private key in a scenario with persistent DH parameters and a private
537 key that is shared between multiple clients. For example this can occur by
538 default in OpenSSL DHE based SSL/TLS ciphersuites.
540 This issue was reported to OpenSSL by Hanno Böck.
544 *) Certificate verify crash with missing PSS parameter
546 The signature verification routines will crash with a NULL pointer
547 dereference if presented with an ASN.1 signature using the RSA PSS
548 algorithm and absent mask generation function parameter. Since these
549 routines are used to verify certificate signature algorithms this can be
550 used to crash any certificate verification operation and exploited in a
551 DoS attack. Any application which performs certificate verification is
552 vulnerable including OpenSSL clients and servers which enable client
555 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
559 *) X509_ATTRIBUTE memory leak
561 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
562 memory. This structure is used by the PKCS#7 and CMS routines so any
563 application which reads PKCS#7 or CMS data from untrusted sources is
564 affected. SSL/TLS is not affected.
566 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
571 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
572 This changes the decoding behaviour for some invalid messages,
573 though the change is mostly in the more lenient direction, and
574 legacy behaviour is preserved as much as possible.
577 *) In DSA_generate_parameters_ex, if the provided seed is too short,
578 use a random seed, as already documented.
579 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
581 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
583 *) Alternate chains certificate forgery
585 During certificate verfification, OpenSSL will attempt to find an
586 alternative certificate chain if the first attempt to build such a chain
587 fails. An error in the implementation of this logic can mean that an
588 attacker could cause certain checks on untrusted certificates to be
589 bypassed, such as the CA flag, enabling them to use a valid leaf
590 certificate to act as a CA and "issue" an invalid certificate.
592 This issue was reported to OpenSSL by Adam Langley/David Benjamin
597 *) Race condition handling PSK identify hint
599 If PSK identity hints are received by a multi-threaded client then
600 the values are wrongly updated in the parent SSL_CTX structure. This can
601 result in a race condition potentially leading to a double free of the
606 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
608 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
609 incompatibility in the handling of HMAC. The previous ABI has now been
612 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
614 *) Malformed ECParameters causes infinite loop
616 When processing an ECParameters structure OpenSSL enters an infinite loop
617 if the curve specified is over a specially malformed binary polynomial
620 This can be used to perform denial of service against any
621 system which processes public keys, certificate requests or
622 certificates. This includes TLS clients and TLS servers with
623 client authentication enabled.
625 This issue was reported to OpenSSL by Joseph Barr-Pixton.
629 *) Exploitable out-of-bounds read in X509_cmp_time
631 X509_cmp_time does not properly check the length of the ASN1_TIME
632 string and can read a few bytes out of bounds. In addition,
633 X509_cmp_time accepts an arbitrary number of fractional seconds in the
636 An attacker can use this to craft malformed certificates and CRLs of
637 various sizes and potentially cause a segmentation fault, resulting in
638 a DoS on applications that verify certificates or CRLs. TLS clients
639 that verify CRLs are affected. TLS clients and servers with client
640 authentication enabled may be affected if they use custom verification
643 This issue was reported to OpenSSL by Robert Swiecki (Google), and
644 independently by Hanno Böck.
648 *) PKCS7 crash with missing EnvelopedContent
650 The PKCS#7 parsing code does not handle missing inner EncryptedContent
651 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
652 with missing content and trigger a NULL pointer dereference on parsing.
654 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
655 structures from untrusted sources are affected. OpenSSL clients and
656 servers are not affected.
658 This issue was reported to OpenSSL by Michal Zalewski (Google).
662 *) CMS verify infinite loop with unknown hash function
664 When verifying a signedData message the CMS code can enter an infinite loop
665 if presented with an unknown hash function OID. This can be used to perform
666 denial of service against any system which verifies signedData messages using
668 This issue was reported to OpenSSL by Johannes Bauer.
672 *) Race condition handling NewSessionTicket
674 If a NewSessionTicket is received by a multi-threaded client when attempting to
675 reuse a previous ticket then a race condition can occur potentially leading to
676 a double free of the ticket data.
680 *) Removed support for the two export grade static DH ciphersuites
681 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
682 were newly added (along with a number of other static DH ciphersuites) to
683 1.0.2. However the two export ones have *never* worked since they were
684 introduced. It seems strange in any case to be adding new export
685 ciphersuites, and given "logjam" it also does not seem correct to fix them.
688 *) Only support 256-bit or stronger elliptic curves with the
689 'ecdh_auto' setting (server) or by default (client). Of supported
690 curves, prefer P-256 (both).
693 *) Reject DH handshakes with parameters shorter than 768 bits.
694 [Kurt Roeckx and Emilia Kasper]
696 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
698 *) ClientHello sigalgs DoS fix
700 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
701 invalid signature algorithms extension a NULL pointer dereference will
702 occur. This can be exploited in a DoS attack against the server.
704 This issue was was reported to OpenSSL by David Ramos of Stanford
707 [Stephen Henson and Matt Caswell]
709 *) Multiblock corrupted pointer fix
711 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
712 feature only applies on 64 bit x86 architecture platforms that support AES
713 NI instructions. A defect in the implementation of "multiblock" can cause
714 OpenSSL's internal write buffer to become incorrectly set to NULL when
715 using non-blocking IO. Typically, when the user application is using a
716 socket BIO for writing, this will only result in a failed connection.
717 However if some other BIO is used then it is likely that a segmentation
718 fault will be triggered, thus enabling a potential DoS attack.
720 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
724 *) Segmentation fault in DTLSv1_listen fix
726 The DTLSv1_listen function is intended to be stateless and processes the
727 initial ClientHello from many peers. It is common for user code to loop
728 over the call to DTLSv1_listen until a valid ClientHello is received with
729 an associated cookie. A defect in the implementation of DTLSv1_listen means
730 that state is preserved in the SSL object from one invocation to the next
731 that can lead to a segmentation fault. Errors processing the initial
732 ClientHello can trigger this scenario. An example of such an error could be
733 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
736 This issue was reported to OpenSSL by Per Allansson.
740 *) Segmentation fault in ASN1_TYPE_cmp fix
742 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
743 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
744 certificate signature algorithm consistency this can be used to crash any
745 certificate verification operation and exploited in a DoS attack. Any
746 application which performs certificate verification is vulnerable including
747 OpenSSL clients and servers which enable client authentication.
751 *) Segmentation fault for invalid PSS parameters fix
753 The signature verification routines will crash with a NULL pointer
754 dereference if presented with an ASN.1 signature using the RSA PSS
755 algorithm and invalid parameters. Since these routines are used to verify
756 certificate signature algorithms this can be used to crash any
757 certificate verification operation and exploited in a DoS attack. Any
758 application which performs certificate verification is vulnerable including
759 OpenSSL clients and servers which enable client authentication.
761 This issue was was reported to OpenSSL by Brian Carpenter.
765 *) ASN.1 structure reuse memory corruption fix
767 Reusing a structure in ASN.1 parsing may allow an attacker to cause
768 memory corruption via an invalid write. Such reuse is and has been
769 strongly discouraged and is believed to be rare.
771 Applications that parse structures containing CHOICE or ANY DEFINED BY
772 components may be affected. Certificate parsing (d2i_X509 and related
773 functions) are however not affected. OpenSSL clients and servers are
778 *) PKCS7 NULL pointer dereferences fix
780 The PKCS#7 parsing code does not handle missing outer ContentInfo
781 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
782 missing content and trigger a NULL pointer dereference on parsing.
784 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
785 otherwise parse PKCS#7 structures from untrusted sources are
786 affected. OpenSSL clients and servers are not affected.
788 This issue was reported to OpenSSL by Michal Zalewski (Google).
792 *) DoS via reachable assert in SSLv2 servers fix
794 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
795 servers that both support SSLv2 and enable export cipher suites by sending
796 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
798 This issue was discovered by Sean Burford (Google) and Emilia Käsper
799 (OpenSSL development team).
803 *) Empty CKE with client auth and DHE fix
805 If client auth is used then a server can seg fault in the event of a DHE
806 ciphersuite being selected and a zero length ClientKeyExchange message
807 being sent by the client. This could be exploited in a DoS attack.
811 *) Handshake with unseeded PRNG fix
813 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
814 with an unseeded PRNG. The conditions are:
815 - The client is on a platform where the PRNG has not been seeded
816 automatically, and the user has not seeded manually
817 - A protocol specific client method version has been used (i.e. not
818 SSL_client_methodv23)
819 - A ciphersuite is used that does not require additional random data from
820 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
822 If the handshake succeeds then the client random that has been used will
823 have been generated from a PRNG with insufficient entropy and therefore the
824 output may be predictable.
826 For example using the following command with an unseeded openssl will
827 succeed on an unpatched platform:
829 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
833 *) Use After Free following d2i_ECPrivatekey error fix
835 A malformed EC private key file consumed via the d2i_ECPrivateKey function
836 could cause a use after free condition. This, in turn, could cause a double
837 free in several private key parsing functions (such as d2i_PrivateKey
838 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
839 for applications that receive EC private keys from untrusted
840 sources. This scenario is considered rare.
842 This issue was discovered by the BoringSSL project and fixed in their
847 *) X509_to_X509_REQ NULL pointer deref fix
849 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
850 the certificate key is invalid. This function is rarely used in practice.
852 This issue was discovered by Brian Carpenter.
856 *) Removed the export ciphers from the DEFAULT ciphers
859 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
861 *) Change RSA and DH/DSA key generation apps to generate 2048-bit
865 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
866 ARMv5 through ARMv8, as opposite to "locking" it to single one.
867 So far those who have to target multiple plaforms would compromise
868 and argue that binary targeting say ARMv5 would still execute on
869 ARMv8. "Universal" build resolves this compromise by providing
870 near-optimal performance even on newer platforms.
873 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
874 (other platforms pending).
875 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
877 *) Add support for the SignedCertificateTimestampList certificate and
878 OCSP response extensions from RFC6962.
881 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
882 for corner cases. (Certain input points at infinity could lead to
883 bogus results, with non-infinity inputs mapped to infinity too.)
886 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
887 This covers AES, SHA256/512 and GHASH. "Initial" means that most
888 common cases are optimized and there still is room for further
889 improvements. Vector Permutation AES for Altivec is also added.
892 *) Add support for little-endian ppc64 Linux target.
893 [Marcelo Cerri (IBM)]
895 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
896 SHA1, SHA256 and GHASH. "Initial" means that most common cases
897 are optimized and there still is room for further improvements.
898 Both 32- and 64-bit modes are supported.
899 [Andy Polyakov, Ard Biesheuvel (Linaro)]
901 *) Improved ARMv7 NEON support.
904 *) Support for SPARC Architecture 2011 crypto extensions, first
905 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
906 SHA256/512, MD5, GHASH and modular exponentiation.
907 [Andy Polyakov, David Miller]
909 *) Accelerated modular exponentiation for Intel processors, a.k.a.
911 [Shay Gueron & Vlad Krasnov (Intel Corp)]
913 *) Support for new and upcoming Intel processors, including AVX2,
914 BMI and SHA ISA extensions. This includes additional "stitched"
915 implementations, AESNI-SHA256 and GCM, and multi-buffer support
918 This work was sponsored by Intel Corp.
921 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
922 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
923 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
926 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
927 this fixes a limiation in previous versions of OpenSSL.
930 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
931 MGF1 digest and OAEP label.
934 *) Add EVP support for key wrapping algorithms, to avoid problems with
935 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
936 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
937 algorithms and include tests cases.
940 *) Add functions to allocate and set the fields of an ECDSA_METHOD
942 [Douglas E. Engert, Steve Henson]
944 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
945 difference in days and seconds between two tm or ASN1_TIME structures.
948 *) Add -rev test option to s_server to just reverse order of characters
949 received by client and send back to server. Also prints an abbreviated
950 summary of the connection parameters.
953 *) New option -brief for s_client and s_server to print out a brief summary
954 of connection parameters.
957 *) Add callbacks for arbitrary TLS extensions.
958 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
960 *) New option -crl_download in several openssl utilities to download CRLs
961 from CRLDP extension in certificates.
964 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
967 *) New function X509_CRL_diff to generate a delta CRL from the difference
968 of two full CRLs. Add support to "crl" utility.
971 *) New functions to set lookup_crls function and to retrieve
972 X509_STORE from X509_STORE_CTX.
975 *) Print out deprecated issuer and subject unique ID fields in
979 *) Extend OCSP I/O functions so they can be used for simple general purpose
980 HTTP as well as OCSP. New wrapper function which can be used to download
981 CRLs using the OCSP API.
984 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
987 *) SSL_CONF* functions. These provide a common framework for application
988 configuration using configuration files or command lines.
991 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
992 message callback and prints the results. Needs compile time option
993 "enable-ssl-trace". New options to s_client and s_server to enable
997 *) New ctrl and macro to retrieve supported points extensions.
998 Print out extension in s_server and s_client.
1001 *) New functions to retrieve certificate signature and signature
1005 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1009 *) New Suite B modes for TLS code. These use and enforce the requirements
1010 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1011 only use Suite B curves. The Suite B modes can be set by using the
1012 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1015 *) New chain verification flags for Suite B levels of security. Check
1016 algorithms are acceptable when flags are set in X509_verify_cert.
1019 *) Make tls1_check_chain return a set of flags indicating checks passed
1020 by a certificate chain. Add additional tests to handle client
1021 certificates: checks for matching certificate type and issuer name
1025 *) If an attempt is made to use a signature algorithm not in the peer
1026 preference list abort the handshake. If client has no suitable
1027 signature algorithms in response to a certificate request do not
1028 use the certificate.
1031 *) If server EC tmp key is not in client preference list abort handshake.
1034 *) Add support for certificate stores in CERT structure. This makes it
1035 possible to have different stores per SSL structure or one store in
1036 the parent SSL_CTX. Include distint stores for certificate chain
1037 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1038 to build and store a certificate chain in CERT structure: returing
1039 an error if the chain cannot be built: this will allow applications
1040 to test if a chain is correctly configured.
1042 Note: if the CERT based stores are not set then the parent SSL_CTX
1043 store is used to retain compatibility with existing behaviour.
1047 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1048 mask based on the current session, check mask when sending client
1049 hello and checking the requested ciphersuite.
1052 *) New ctrls to retrieve and set certificate types in a certificate
1053 request message. Print out received values in s_client. If certificate
1054 types is not set with custom values set sensible values based on
1055 supported signature algorithms.
1058 *) Support for distinct client and server supported signature algorithms.
1061 *) Add certificate callback. If set this is called whenever a certificate
1062 is required by client or server. An application can decide which
1063 certificate chain to present based on arbitrary criteria: for example
1064 supported signature algorithms. Add very simple example to s_server.
1065 This fixes many of the problems and restrictions of the existing client
1066 certificate callback: for example you can now clear an existing
1067 certificate and specify the whole chain.
1070 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1071 the certificate can be used for (if anything). Set valid_flags field
1072 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1073 to have similar checks in it.
1075 Add new "cert_flags" field to CERT structure and include a "strict mode".
1076 This enforces some TLS certificate requirements (such as only permitting
1077 certificate signature algorithms contained in the supported algorithms
1078 extension) which some implementations ignore: this option should be used
1079 with caution as it could cause interoperability issues.
1082 *) Update and tidy signature algorithm extension processing. Work out
1083 shared signature algorithms based on preferences and peer algorithms
1084 and print them out in s_client and s_server. Abort handshake if no
1085 shared signature algorithms.
1088 *) Add new functions to allow customised supported signature algorithms
1089 for SSL and SSL_CTX structures. Add options to s_client and s_server
1093 *) New function SSL_certs_clear() to delete all references to certificates
1094 from an SSL structure. Before this once a certificate had been added
1095 it couldn't be removed.
1098 *) Integrate hostname, email address and IP address checking with certificate
1099 verification. New verify options supporting checking in opensl utility.
1102 *) Fixes and wildcard matching support to hostname and email checking
1103 functions. Add manual page.
1104 [Florian Weimer (Red Hat Product Security Team)]
1106 *) New functions to check a hostname email or IP address against a
1107 certificate. Add options x509 utility to print results of checks against
1111 *) Fix OCSP checking.
1112 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1114 *) Initial experimental support for explicitly trusted non-root CAs.
1115 OpenSSL still tries to build a complete chain to a root but if an
1116 intermediate CA has a trust setting included that is used. The first
1117 setting is used: whether to trust (e.g., -addtrust option to the x509
1121 *) Add -trusted_first option which attempts to find certificates in the
1122 trusted store even if an untrusted chain is also supplied.
1125 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1126 platform support for Linux and Android.
1129 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1132 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1133 When in FIPS mode the approved implementations are used as normal,
1134 when not in FIPS mode the internal unapproved versions are used instead.
1135 This means that the FIPS capable OpenSSL isn't forced to use the
1136 (often lower perfomance) FIPS implementations outside FIPS mode.
1139 *) Transparently support X9.42 DH parameters when calling
1140 PEM_read_bio_DHparameters. This means existing applications can handle
1141 the new parameter format automatically.
1144 *) Initial experimental support for X9.42 DH parameter format: mainly
1145 to support use of 'q' parameter for RFC5114 parameters.
1148 *) Add DH parameters from RFC5114 including test data to dhtest.
1151 *) Support for automatic EC temporary key parameter selection. If enabled
1152 the most preferred EC parameters are automatically used instead of
1153 hardcoded fixed parameters. Now a server just has to call:
1154 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1155 support ECDH and use the most appropriate parameters.
1158 *) Enhance and tidy EC curve and point format TLS extension code. Use
1159 static structures instead of allocation if default values are used.
1160 New ctrls to set curves we wish to support and to retrieve shared curves.
1161 Print out shared curves in s_server. New options to s_server and s_client
1162 to set list of supported curves.
1165 *) New ctrls to retrieve supported signature algorithms and
1166 supported curve values as an array of NIDs. Extend openssl utility
1167 to print out received values.
1170 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1171 between NIDs and the more common NIST names such as "P-256". Enhance
1172 ecparam utility and ECC method to recognise the NIST names for curves.
1175 *) Enhance SSL/TLS certificate chain handling to support different
1176 chains for each certificate instead of one chain in the parent SSL_CTX.
1179 *) Support for fixed DH ciphersuite client authentication: where both
1180 server and client use DH certificates with common parameters.
1183 *) Support for fixed DH ciphersuites: those requiring DH server
1187 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1189 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1190 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1191 X509_CINF_get_signature were reverted post internal team review.
1193 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1195 *) Build fixes for the Windows and OpenVMS platforms
1196 [Matt Caswell and Richard Levitte]
1198 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1200 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1201 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1202 dereference. This could lead to a Denial Of Service attack. Thanks to
1203 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1207 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1208 dtls1_buffer_record function under certain conditions. In particular this
1209 could occur if an attacker sent repeated DTLS records with the same
1210 sequence number but for the next epoch. The memory leak could be exploited
1211 by an attacker in a Denial of Service attack through memory exhaustion.
1212 Thanks to Chris Mueller for reporting this issue.
1216 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1217 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1218 method would be set to NULL which could later result in a NULL pointer
1219 dereference. Thanks to Frank Schmirler for reporting this issue.
1223 *) Abort handshake if server key exchange message is omitted for ephemeral
1226 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1227 reporting this issue.
1231 *) Remove non-export ephemeral RSA code on client and server. This code
1232 violated the TLS standard by allowing the use of temporary RSA keys in
1233 non-export ciphersuites and could be used by a server to effectively
1234 downgrade the RSA key length used to a value smaller than the server
1235 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1236 INRIA or reporting this issue.
1240 *) Fixed issue where DH client certificates are accepted without verification.
1241 An OpenSSL server will accept a DH certificate for client authentication
1242 without the certificate verify message. This effectively allows a client to
1243 authenticate without the use of a private key. This only affects servers
1244 which trust a client certificate authority which issues certificates
1245 containing DH keys: these are extremely rare and hardly ever encountered.
1246 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1251 *) Ensure that the session ID context of an SSL is updated when its
1252 SSL_CTX is updated via SSL_set_SSL_CTX.
1254 The session ID context is typically set from the parent SSL_CTX,
1255 and can vary with the CTX.
1258 *) Fix various certificate fingerprint issues.
1260 By using non-DER or invalid encodings outside the signed portion of a
1261 certificate the fingerprint can be changed without breaking the signature.
1262 Although no details of the signed portion of the certificate can be changed
1263 this can cause problems with some applications: e.g. those using the
1264 certificate fingerprint for blacklists.
1266 1. Reject signatures with non zero unused bits.
1268 If the BIT STRING containing the signature has non zero unused bits reject
1269 the signature. All current signature algorithms require zero unused bits.
1271 2. Check certificate algorithm consistency.
1273 Check the AlgorithmIdentifier inside TBS matches the one in the
1274 certificate signature. NB: this will result in signature failure
1275 errors for some broken certificates.
1277 Thanks to Konrad Kraszewski from Google for reporting this issue.
1279 3. Check DSA/ECDSA signatures use DER.
1281 Reencode DSA/ECDSA signatures and compare with the original received
1282 signature. Return an error if there is a mismatch.
1284 This will reject various cases including garbage after signature
1285 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1286 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1287 (negative or with leading zeroes).
1289 Further analysis was conducted and fixes were developed by Stephen Henson
1290 of the OpenSSL core team.
1295 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1296 results on some platforms, including x86_64. This bug occurs at random
1297 with a very low probability, and is not known to be exploitable in any
1298 way, though its exact impact is difficult to determine. Thanks to Pieter
1299 Wuille (Blockstream) who reported this issue and also suggested an initial
1300 fix. Further analysis was conducted by the OpenSSL development team and
1301 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1302 the OpenSSL core team.
1306 *) Do not resume sessions on the server if the negotiated protocol
1307 version does not match the session's version. Resuming with a different
1308 version, while not strictly forbidden by the RFC, is of questionable
1309 sanity and breaks all known clients.
1310 [David Benjamin, Emilia Käsper]
1312 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1313 early CCS messages during renegotiation. (Note that because
1314 renegotiation is encrypted, this early CCS was not exploitable.)
1317 *) Tighten client-side session ticket handling during renegotiation:
1318 ensure that the client only accepts a session ticket if the server sends
1319 the extension anew in the ServerHello. Previously, a TLS client would
1320 reuse the old extension state and thus accept a session ticket if one was
1321 announced in the initial ServerHello.
1323 Similarly, ensure that the client requires a session ticket if one
1324 was advertised in the ServerHello. Previously, a TLS client would
1325 ignore a missing NewSessionTicket message.
1328 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1330 *) SRTP Memory Leak.
1332 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1333 sends a carefully crafted handshake message, to cause OpenSSL to fail
1334 to free up to 64k of memory causing a memory leak. This could be
1335 exploited in a Denial Of Service attack. This issue affects OpenSSL
1336 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1337 whether SRTP is used or configured. Implementations of OpenSSL that
1338 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1340 The fix was developed by the OpenSSL team.
1344 *) Session Ticket Memory Leak.
1346 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1347 integrity of that ticket is first verified. In the event of a session
1348 ticket integrity check failing, OpenSSL will fail to free memory
1349 causing a memory leak. By sending a large number of invalid session
1350 tickets an attacker could exploit this issue in a Denial Of Service
1355 *) Build option no-ssl3 is incomplete.
1357 When OpenSSL is configured with "no-ssl3" as a build option, servers
1358 could accept and complete a SSL 3.0 handshake, and clients could be
1359 configured to send them.
1361 [Akamai and the OpenSSL team]
1363 *) Add support for TLS_FALLBACK_SCSV.
1364 Client applications doing fallback retries should call
1365 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1367 [Adam Langley, Bodo Moeller]
1369 *) Add additional DigestInfo checks.
1371 Reencode DigestInto in DER and check against the original when
1372 verifying RSA signature: this will reject any improperly encoded
1373 DigestInfo structures.
1375 Note: this is a precautionary measure and no attacks are currently known.
1379 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1381 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1382 SRP code can be overrun an internal buffer. Add sanity check that
1383 g, A, B < N to SRP code.
1385 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1386 Group for discovering this issue.
1390 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1391 TLS 1.0 instead of higher protocol versions when the ClientHello message
1392 is badly fragmented. This allows a man-in-the-middle attacker to force a
1393 downgrade to TLS 1.0 even if both the server and the client support a
1394 higher protocol version, by modifying the client's TLS records.
1396 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1397 researching this issue.
1401 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1402 to a denial of service attack. A malicious server can crash the client
1403 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1404 ciphersuite and sending carefully crafted handshake messages.
1406 Thanks to Felix Gröbert (Google) for discovering and researching this
1411 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1412 to leak memory. This can be exploited through a Denial of Service attack.
1413 Thanks to Adam Langley for discovering and researching this issue.
1417 *) An attacker can force openssl to consume large amounts of memory whilst
1418 processing DTLS handshake messages. This can be exploited through a
1419 Denial of Service attack.
1420 Thanks to Adam Langley for discovering and researching this issue.
1424 *) An attacker can force an error condition which causes openssl to crash
1425 whilst processing DTLS packets due to memory being freed twice. This
1426 can be exploited through a Denial of Service attack.
1427 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1432 *) If a multithreaded client connects to a malicious server using a resumed
1433 session and the server sends an ec point format extension it could write
1434 up to 255 bytes to freed memory.
1436 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1441 *) A malicious server can crash an OpenSSL client with a null pointer
1442 dereference (read) by specifying an SRP ciphersuite even though it was not
1443 properly negotiated with the client. This can be exploited through a
1444 Denial of Service attack.
1446 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1447 discovering and researching this issue.
1451 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1452 X509_name_oneline, X509_name_print_ex et al. to leak some information
1453 from the stack. Applications may be affected if they echo pretty printing
1454 output to the attacker.
1456 Thanks to Ivan Fratric (Google) for discovering this issue.
1458 [Emilia Käsper, and Steve Henson]
1460 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1461 for corner cases. (Certain input points at infinity could lead to
1462 bogus results, with non-infinity inputs mapped to infinity too.)
1465 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1467 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1468 handshake can force the use of weak keying material in OpenSSL
1469 SSL/TLS clients and servers.
1471 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1472 researching this issue. (CVE-2014-0224)
1473 [KIKUCHI Masashi, Steve Henson]
1475 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1476 OpenSSL DTLS client the code can be made to recurse eventually crashing
1479 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1481 [Imre Rad, Steve Henson]
1483 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1484 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1485 client or server. This is potentially exploitable to run arbitrary
1486 code on a vulnerable client or server.
1488 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1489 [Jüri Aedla, Steve Henson]
1491 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1492 are subject to a denial of service attack.
1494 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1495 this issue. (CVE-2014-3470)
1496 [Felix Gröbert, Ivan Fratric, Steve Henson]
1498 *) Harmonize version and its documentation. -f flag is used to display
1500 [mancha <mancha1@zoho.com>]
1502 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1503 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1504 [mancha <mancha1@zoho.com>]
1506 *) Fix some double frees. These are not thought to be exploitable.
1507 [mancha <mancha1@zoho.com>]
1509 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1511 *) A missing bounds check in the handling of the TLS heartbeat extension
1512 can be used to reveal up to 64k of memory to a connected client or
1515 Thanks for Neel Mehta of Google Security for discovering this bug and to
1516 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1517 preparing the fix (CVE-2014-0160)
1518 [Adam Langley, Bodo Moeller]
1520 *) Fix for the attack described in the paper "Recovering OpenSSL
1521 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1522 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1523 http://eprint.iacr.org/2014/140
1525 Thanks to Yuval Yarom and Naomi Benger for discovering this
1526 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1527 [Yuval Yarom and Naomi Benger]
1529 *) TLS pad extension: draft-agl-tls-padding-03
1531 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1532 TLS client Hello record length value would otherwise be > 255 and
1533 less that 512 pad with a dummy extension containing zeroes so it
1534 is at least 512 bytes long.
1536 [Adam Langley, Steve Henson]
1538 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1540 *) Fix for TLS record tampering bug. A carefully crafted invalid
1541 handshake could crash OpenSSL with a NULL pointer exception.
1542 Thanks to Anton Johansson for reporting this issues.
1545 *) Keep original DTLS digest and encryption contexts in retransmission
1546 structures so we can use the previous session parameters if they need
1547 to be resent. (CVE-2013-6450)
1550 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1551 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1552 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1553 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1554 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1555 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1556 [Rob Stradling, Adam Langley]
1558 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1560 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1561 supporting platforms or when small records were transferred.
1562 [Andy Polyakov, Steve Henson]
1564 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1566 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1568 This addresses the flaw in CBC record processing discovered by
1569 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1570 at: http://www.isg.rhul.ac.uk/tls/
1572 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1573 Security Group at Royal Holloway, University of London
1574 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1575 Emilia Käsper for the initial patch.
1577 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1579 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1580 ciphersuites which can be exploited in a denial of service attack.
1581 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1582 and detecting this bug and to Wolfgang Ettlinger
1583 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1587 *) Return an error when checking OCSP signatures when key is NULL.
1588 This fixes a DoS attack. (CVE-2013-0166)
1591 *) Make openssl verify return errors.
1592 [Chris Palmer <palmer@google.com> and Ben Laurie]
1594 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1595 the right response is stapled. Also change SSL_get_certificate()
1596 so it returns the certificate actually sent.
1597 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1598 [Rob Stradling <rob.stradling@comodo.com>]
1600 *) Fix possible deadlock when decoding public keys.
1603 *) Don't use TLS 1.0 record version number in initial client hello
1607 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1609 *) Sanity check record length before skipping explicit IV in TLS
1610 1.2, 1.1 and DTLS to fix DoS attack.
1612 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1613 fuzzing as a service testing platform.
1617 *) Initialise tkeylen properly when encrypting CMS messages.
1618 Thanks to Solar Designer of Openwall for reporting this issue.
1621 *) In FIPS mode don't try to use composite ciphers as they are not
1625 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1627 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1628 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1629 mean any application compiled against OpenSSL 1.0.0 headers setting
1630 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1631 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1632 0x10000000L Any application which was previously compiled against
1633 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1634 will need to be recompiled as a result. Letting be results in
1635 inability to disable specifically TLS 1.1 and in client context,
1636 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1639 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1640 disable just protocol X, but all protocols above X *if* there are
1641 protocols *below* X still enabled. In more practical terms it means
1642 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1643 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1644 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1648 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1650 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1651 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1652 in CRYPTO_realloc_clean.
1654 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1655 issue and to Adam Langley <agl@chromium.org> for fixing it.
1657 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1659 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1662 *) Workarounds for some broken servers that "hang" if a client hello
1663 record length exceeds 255 bytes.
1665 1. Do not use record version number > TLS 1.0 in initial client
1666 hello: some (but not all) hanging servers will now work.
1667 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1668 the number of ciphers sent in the client hello. This should be
1669 set to an even number, such as 50, for example by passing:
1670 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1671 Most broken servers should now work.
1672 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1673 TLS 1.2 client support entirely.
1676 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1679 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1681 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1682 STRING form instead of a DigestInfo.
1685 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1686 and the RSA_sign/RSA_verify functions. This was made more apparent when
1687 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1688 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1689 the correct format in RSA_verify so both forms transparently work.
1692 *) Some servers which support TLS 1.0 can choke if we initially indicate
1693 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1694 encrypted premaster secret. As a workaround use the maximum pemitted
1695 client version in client hello, this should keep such servers happy
1696 and still work with previous versions of OpenSSL.
1699 *) Add support for TLS/DTLS heartbeats.
1700 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1702 *) Add support for SCTP.
1703 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1705 *) Improved PRNG seeding for VOS.
1706 [Paul Green <Paul.Green@stratus.com>]
1708 *) Extensive assembler packs updates, most notably:
1710 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1711 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1712 - x86_64: bit-sliced AES implementation;
1713 - ARM: NEON support, contemporary platforms optimizations;
1714 - s390x: z196 support;
1715 - *: GHASH and GF(2^m) multiplication implementations;
1719 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1720 (removal of unnecessary code)
1721 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1723 *) Add TLS key material exporter from RFC 5705.
1726 *) Add DTLS-SRTP negotiation from RFC 5764.
1729 *) Add Next Protocol Negotiation,
1730 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1731 disabled with a no-npn flag to config or Configure. Code donated
1733 [Adam Langley <agl@google.com> and Ben Laurie]
1735 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1736 NIST-P256, NIST-P521, with constant-time single point multiplication on
1737 typical inputs. Compiler support for the nonstandard type __uint128_t is
1738 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1739 Code made available under Apache License version 2.0.
1741 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1742 line to include this in your build of OpenSSL, and run "make depend" (or
1743 "make update"). This enables the following EC_METHODs:
1745 EC_GFp_nistp224_method()
1746 EC_GFp_nistp256_method()
1747 EC_GFp_nistp521_method()
1749 EC_GROUP_new_by_curve_name() will automatically use these (while
1750 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1752 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1754 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1755 all platforms. Move ssize_t definition from e_os.h to the public
1756 header file e_os2.h as it now appears in public header file cms.h
1759 *) New -sigopt option to the ca, req and x509 utilities. Additional
1760 signature parameters can be passed using this option and in
1764 *) Add RSA PSS signing function. This will generate and set the
1765 appropriate AlgorithmIdentifiers for PSS based on those in the
1766 corresponding EVP_MD_CTX structure. No application support yet.
1769 *) Support for companion algorithm specific ASN1 signing routines.
1770 New function ASN1_item_sign_ctx() signs a pre-initialised
1771 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1772 the appropriate parameters.
1775 *) Add new algorithm specific ASN1 verification initialisation function
1776 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1777 handling will be the same no matter what EVP_PKEY_METHOD is used.
1778 Add a PSS handler to support verification of PSS signatures: checked
1779 against a number of sample certificates.
1782 *) Add signature printing for PSS. Add PSS OIDs.
1783 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1785 *) Add algorithm specific signature printing. An individual ASN1 method
1786 can now print out signatures instead of the standard hex dump.
1788 More complex signatures (e.g. PSS) can print out more meaningful
1789 information. Include DSA version that prints out the signature
1793 *) Password based recipient info support for CMS library: implementing
1797 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1798 neatly separates the code into cipher and PBE sections and is required
1799 for some algorithms that split PBES2 into separate pieces (such as
1800 password based CMS).
1803 *) Session-handling fixes:
1804 - Fix handling of connections that are resuming with a session ID,
1805 but also support Session Tickets.
1806 - Fix a bug that suppressed issuing of a new ticket if the client
1807 presented a ticket with an expired session.
1808 - Try to set the ticket lifetime hint to something reasonable.
1809 - Make tickets shorter by excluding irrelevant information.
1810 - On the client side, don't ignore renewed tickets.
1811 [Adam Langley, Bodo Moeller (Google)]
1813 *) Fix PSK session representation.
1816 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1818 This work was sponsored by Intel.
1821 *) Add GCM support to TLS library. Some custom code is needed to split
1822 the IV between the fixed (from PRF) and explicit (from TLS record)
1823 portions. This adds all GCM ciphersuites supported by RFC5288 and
1824 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
1825 add a special AESGCM string for GCM only.
1828 *) Expand range of ctrls for AES GCM. Permit setting invocation
1829 field on decrypt and retrieval of invocation field only on encrypt.
1832 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1833 As required by RFC5289 these ciphersuites cannot be used if for
1834 versions of TLS earlier than 1.2.
1837 *) For FIPS capable OpenSSL interpret a NULL default public key method
1838 as unset and return the appopriate default but do *not* set the default.
1839 This means we can return the appopriate method in applications that
1840 swicth between FIPS and non-FIPS modes.
1843 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1844 ENGINE is used then we cannot handle that in the FIPS module so we
1845 keep original code iff non-FIPS operations are allowed.
1848 *) Add -attime option to openssl utilities.
1849 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1851 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1854 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1855 FIPS EC methods unconditionally for now.
1858 *) New build option no-ec2m to disable characteristic 2 code.
1861 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1862 all cases can be covered as some introduce binary incompatibilities.
1865 *) Redirect RSA operations to FIPS module including keygen,
1866 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1869 *) Add similar low level API blocking to ciphers.
1872 *) Low level digest APIs are not approved in FIPS mode: any attempt
1873 to use these will cause a fatal error. Applications that *really* want
1874 to use them can use the private_* version instead.
1877 *) Redirect cipher operations to FIPS module for FIPS builds.
1880 *) Redirect digest operations to FIPS module for FIPS builds.
1883 *) Update build system to add "fips" flag which will link in fipscanister.o
1884 for static and shared library builds embedding a signature if needed.
1887 *) Output TLS supported curves in preference order instead of numerical
1888 order. This is currently hardcoded for the highest order curves first.
1889 This should be configurable so applications can judge speed vs strength.
1892 *) Add TLS v1.2 server support for client authentication.
1895 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1899 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1900 FIPS modules versions.
1903 *) Add TLS v1.2 client side support for client authentication. Keep cache
1904 of handshake records longer as we don't know the hash algorithm to use
1905 until after the certificate request message is received.
1908 *) Initial TLS v1.2 client support. Add a default signature algorithms
1909 extension including all the algorithms we support. Parse new signature
1910 format in client key exchange. Relax some ECC signing restrictions for
1911 TLS v1.2 as indicated in RFC5246.
1914 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1915 to new signature format when needed using client digest preference.
1916 All server ciphersuites should now work correctly in TLS v1.2. No client
1917 support yet and no support for client certificates.
1920 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1921 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1922 ciphersuites. At present only RSA key exchange ciphersuites work with
1923 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1924 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1925 and version checking.
1928 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1929 with this defined it will not be affected by any changes to ssl internal
1930 structures. Add several utility functions to allow openssl application
1931 to work with OPENSSL_NO_SSL_INTERN defined.
1935 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1937 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1940 *) Permit abbreviated handshakes when renegotiating using the function
1941 SSL_renegotiate_abbreviated().
1942 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1944 *) Add call to ENGINE_register_all_complete() to
1945 ENGINE_load_builtin_engines(), so some implementations get used
1946 automatically instead of needing explicit application support.
1949 *) Add support for TLS key exporter as described in RFC5705.
1950 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1952 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1953 a few changes are required:
1955 Add SSL_OP_NO_TLSv1_1 flag.
1956 Add TLSv1_1 methods.
1957 Update version checking logic to handle version 1.1.
1958 Add explicit IV handling (ported from DTLS code).
1959 Add command line options to s_client/s_server.
1962 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1964 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1965 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1966 content decryption and always return the same error. Note: this attack
1967 needs on average 2^20 messages so it only affects automated senders. The
1968 old behaviour can be reenabled in the CMS code by setting the
1969 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1970 an MMA defence is not necessary.
1971 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1972 this issue. (CVE-2012-0884)
1975 *) Fix CVE-2011-4619: make sure we really are receiving a
1976 client hello before rejecting multiple SGC restarts. Thanks to
1977 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1980 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1982 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1983 Thanks to Antonio Martin, Enterprise Secure Access Research and
1984 Development, Cisco Systems, Inc. for discovering this bug and
1985 preparing a fix. (CVE-2012-0050)
1988 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1990 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1991 of the Vaudenay padding oracle attack on CBC mode encryption
1992 which enables an efficient plaintext recovery attack against
1993 the OpenSSL implementation of DTLS. Their attack exploits timing
1994 differences arising during decryption processing. A research
1995 paper describing this attack can be found at:
1996 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1997 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1998 Security Group at Royal Holloway, University of London
1999 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2000 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2001 for preparing the fix. (CVE-2011-4108)
2002 [Robin Seggelmann, Michael Tuexen]
2004 *) Clear bytes used for block padding of SSL 3.0 records.
2006 [Adam Langley (Google)]
2008 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2009 Kadianakis <desnacked@gmail.com> for discovering this issue and
2010 Adam Langley for preparing the fix. (CVE-2011-4619)
2011 [Adam Langley (Google)]
2013 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2014 [Andrey Kulikov <amdeich@gmail.com>]
2016 *) Prevent malformed RFC3779 data triggering an assertion failure.
2017 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2018 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2019 [Rob Austein <sra@hactrn.net>]
2021 *) Improved PRNG seeding for VOS.
2022 [Paul Green <Paul.Green@stratus.com>]
2024 *) Fix ssl_ciph.c set-up race.
2025 [Adam Langley (Google)]
2027 *) Fix spurious failures in ecdsatest.c.
2028 [Emilia Käsper (Google)]
2030 *) Fix the BIO_f_buffer() implementation (which was mixing different
2031 interpretations of the '..._len' fields).
2032 [Adam Langley (Google)]
2034 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2035 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2036 threads won't reuse the same blinding coefficients.
2038 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2039 lock to call BN_BLINDING_invert_ex, and avoids one use of
2040 BN_BLINDING_update for each BN_BLINDING structure (previously,
2041 the last update always remained unused).
2042 [Emilia Käsper (Google)]
2044 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2045 [Bob Buckholz (Google)]
2047 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2049 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2050 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2051 [Kaspar Brand <ossl@velox.ch>]
2053 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2054 for multi-threaded use of ECDH. (CVE-2011-3210)
2055 [Adam Langley (Google)]
2057 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2060 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2061 signature public key algorithm by using OID xref utilities instead.
2062 Before this you could only use some ECC ciphersuites with SHA1 only.
2065 *) Add protection against ECDSA timing attacks as mentioned in the paper
2066 by Billy Bob Brumley and Nicola Tuveri, see:
2068 http://eprint.iacr.org/2011/232.pdf
2070 [Billy Bob Brumley and Nicola Tuveri]
2072 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2074 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2075 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2077 *) Fix bug in string printing code: if *any* escaping is enabled we must
2078 escape the escape character (backslash) or the resulting string is
2082 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2084 *) Disable code workaround for ancient and obsolete Netscape browsers
2085 and servers: an attacker can use it in a ciphersuite downgrade attack.
2086 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2089 *) Fixed J-PAKE implementation error, originally discovered by
2090 Sebastien Martini, further info and confirmation from Stefan
2091 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2094 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2096 *) Fix extension code to avoid race conditions which can result in a buffer
2097 overrun vulnerability: resumed sessions must not be modified as they can
2098 be shared by multiple threads. CVE-2010-3864
2101 *) Fix WIN32 build system to correctly link an ENGINE directory into
2105 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2107 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2109 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2111 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2113 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2114 context. The operation can be customised via the ctrl mechanism in
2115 case ENGINEs want to include additional functionality.
2118 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2121 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2122 output hashes compatible with older versions of OpenSSL.
2123 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2125 *) Fix compression algorithm handling: if resuming a session use the
2126 compression algorithm of the resumed session instead of determining
2127 it from client hello again. Don't allow server to change algorithm.
2130 *) Add load_crls() function to apps tidying load_certs() too. Add option
2131 to verify utility to allow additional CRLs to be included.
2134 *) Update OCSP request code to permit adding custom headers to the request:
2135 some responders need this.
2138 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2140 [Julia Lawall <julia@diku.dk>]
2142 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2143 needlessly dereferenced structures, used obsolete functions and
2144 didn't handle all updated verify codes correctly.
2147 *) Disable MD2 in the default configuration.
2150 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2151 indicate the initial BIO being pushed or popped. This makes it possible
2152 to determine whether the BIO is the one explicitly called or as a result
2153 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2154 it handles reference counts correctly and doesn't zero out the I/O bio
2155 when it is not being explicitly popped. WARNING: applications which
2156 included workarounds for the old buggy behaviour will need to be modified
2157 or they could free up already freed BIOs.
2160 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2161 renaming to all platforms (within the 0.9.8 branch, this was
2162 done conditionally on Netware platforms to avoid a name clash).
2163 [Guenter <lists@gknw.net>]
2165 *) Add ECDHE and PSK support to DTLS.
2166 [Michael Tuexen <tuexen@fh-muenster.de>]
2168 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2172 *) Add "missing" function EVP_MD_flags() (without this the only way to
2173 retrieve a digest flags is by accessing the structure directly. Update
2174 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2175 or cipher is registered as in the "from" argument. Print out all
2176 registered digests in the dgst usage message instead of manually
2177 attempting to work them out.
2180 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2181 this allows the use of compression and extensions. Change default cipher
2182 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2183 by default unless an application cipher string requests it.
2186 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2187 key ids to find matching certificates and keys but some PKCS#12 files
2188 don't follow the (somewhat unwritten) rules and this strategy fails.
2189 Now just gather all certificates together and the first private key
2190 then look for the first certificate that matches the key.
2193 *) Support use of registered digest and cipher names for dgst and cipher
2194 commands instead of having to add each one as a special case. So now
2201 openssl dgst -sha256 foo
2203 and this works for ENGINE based algorithms too.
2207 *) Update Gost ENGINE to support parameter files.
2208 [Victor B. Wagner <vitus@cryptocom.ru>]
2210 *) Support GeneralizedTime in ca utility.
2211 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2213 *) Enhance the hash format used for certificate directory links. The new
2214 form uses the canonical encoding (meaning equivalent names will work
2215 even if they aren't identical) and uses SHA1 instead of MD5. This form
2216 is incompatible with the older format and as a result c_rehash should
2217 be used to rebuild symbolic links.
2220 *) Make PKCS#8 the default write format for private keys, replacing the
2221 traditional format. This form is standardised, more secure and doesn't
2222 include an implicit MD5 dependency.
2225 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2226 committed to OpenSSL should pass this lot as a minimum.
2229 *) Add session ticket override functionality for use by EAP-FAST.
2230 [Jouni Malinen <j@w1.fi>]
2232 *) Modify HMAC functions to return a value. Since these can be implemented
2233 in an ENGINE errors can occur.
2236 *) Type-checked OBJ_bsearch_ex.
2239 *) Type-checked OBJ_bsearch. Also some constification necessitated
2240 by type-checking. Still to come: TXT_DB, bsearch(?),
2241 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2245 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2246 seconds to a tm structure directly, instead of going through OS
2247 specific date routines. This avoids any issues with OS routines such
2248 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2249 and X509_time_adj_ex() to cover the extended range. The existing
2250 X509_time_adj() is still usable and will no longer have any date issues.
2253 *) Delta CRL support. New use deltas option which will attempt to locate
2254 and search any appropriate delta CRLs available.
2256 This work was sponsored by Google.
2259 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2260 code and add additional score elements. Validate alternate CRL paths
2261 as part of the CRL checking and indicate a new error "CRL path validation
2262 error" in this case. Applications wanting additional details can use
2263 the verify callback and check the new "parent" field. If this is not
2264 NULL CRL path validation is taking place. Existing applications wont
2265 see this because it requires extended CRL support which is off by
2268 This work was sponsored by Google.
2271 *) Support for freshest CRL extension.
2273 This work was sponsored by Google.
2276 *) Initial indirect CRL support. Currently only supported in the CRLs
2277 passed directly and not via lookup. Process certificate issuer
2278 CRL entry extension and lookup CRL entries by bother issuer name
2279 and serial number. Check and process CRL issuer entry in IDP extension.
2281 This work was sponsored by Google.
2284 *) Add support for distinct certificate and CRL paths. The CRL issuer
2285 certificate is validated separately in this case. Only enabled if
2286 an extended CRL support flag is set: this flag will enable additional
2287 CRL functionality in future.
2289 This work was sponsored by Google.
2292 *) Add support for policy mappings extension.
2294 This work was sponsored by Google.
2297 *) Fixes to pathlength constraint, self issued certificate handling,
2298 policy processing to align with RFC3280 and PKITS tests.
2300 This work was sponsored by Google.
2303 *) Support for name constraints certificate extension. DN, email, DNS
2304 and URI types are currently supported.
2306 This work was sponsored by Google.
2309 *) To cater for systems that provide a pointer-based thread ID rather
2310 than numeric, deprecate the current numeric thread ID mechanism and
2311 replace it with a structure and associated callback type. This
2312 mechanism allows a numeric "hash" to be extracted from a thread ID in
2313 either case, and on platforms where pointers are larger than 'long',
2314 mixing is done to help ensure the numeric 'hash' is usable even if it
2315 can't be guaranteed unique. The default mechanism is to use "&errno"
2316 as a pointer-based thread ID to distinguish between threads.
2318 Applications that want to provide their own thread IDs should now use
2319 CRYPTO_THREADID_set_callback() to register a callback that will call
2320 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2322 Note that ERR_remove_state() is now deprecated, because it is tied
2323 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2324 to free the current thread's error state should be replaced by
2325 ERR_remove_thread_state(NULL).
2327 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2328 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2329 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2330 application was previously providing a numeric thread callback that
2331 was inappropriate for distinguishing threads, then uniqueness might
2332 have been obtained with &errno that happened immediately in the
2333 intermediate development versions of OpenSSL; this is no longer the
2334 case, the numeric thread callback will now override the automatic use
2336 [Geoff Thorpe, with help from Bodo Moeller]
2338 *) Initial support for different CRL issuing certificates. This covers a
2339 simple case where the self issued certificates in the chain exist and
2340 the real CRL issuer is higher in the existing chain.
2342 This work was sponsored by Google.
2345 *) Removed effectively defunct crypto/store from the build.
2348 *) Revamp of STACK to provide stronger type-checking. Still to come:
2349 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2350 ASN1_STRING, CONF_VALUE.
2353 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2354 RAM on SSL connections. This option can save about 34k per idle SSL.
2357 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2358 STACK, TXT_DB, bsearch, qsort.
2361 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2362 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2363 support for data, signedData, compressedData, digestedData and
2364 encryptedData, envelopedData types included. Scripts to check against
2365 RFC4134 examples draft and interop and consistency checks of many
2366 content types and variants.
2369 *) Add options to enc utility to support use of zlib compression BIO.
2372 *) Extend mk1mf to support importing of options and assembly language
2373 files from Configure script, currently only included in VC-WIN32.
2374 The assembly language rules can now optionally generate the source
2375 files from the associated perl scripts.
2378 *) Implement remaining functionality needed to support GOST ciphersuites.
2379 Interop testing has been performed using CryptoPro implementations.
2380 [Victor B. Wagner <vitus@cryptocom.ru>]
2382 *) s390x assembler pack.
2385 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2389 *) Implement Opaque PRF Input TLS extension as specified in
2390 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2391 official specification yet and no extension type assignment by
2392 IANA exists, this extension (for now) will have to be explicitly
2393 enabled when building OpenSSL by providing the extension number
2394 to use. For example, specify an option
2396 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2398 to the "config" or "Configure" script to enable the extension,
2399 assuming extension number 0x9527 (which is a completely arbitrary
2400 and unofficial assignment based on the MD5 hash of the Internet
2401 Draft). Note that by doing so, you potentially lose
2402 interoperability with other TLS implementations since these might
2403 be using the same extension number for other purposes.
2405 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2406 opaque PRF input value to use in the handshake. This will create
2407 an interal copy of the length-'len' string at 'src', and will
2408 return non-zero for success.
2410 To get more control and flexibility, provide a callback function
2413 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2414 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2418 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2421 Callback function 'cb' will be called in handshakes, and is
2422 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2423 Argument 'arg' is for application purposes (the value as given to
2424 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2425 be provided to the callback function). The callback function
2426 has to return non-zero to report success: usually 1 to use opaque
2427 PRF input just if possible, or 2 to enforce use of the opaque PRF
2428 input. In the latter case, the library will abort the handshake
2429 if opaque PRF input is not successfully negotiated.
2431 Arguments 'peerinput' and 'len' given to the callback function
2432 will always be NULL and 0 in the case of a client. A server will
2433 see the client's opaque PRF input through these variables if
2434 available (NULL and 0 otherwise). Note that if the server
2435 provides an opaque PRF input, the length must be the same as the
2436 length of the client's opaque PRF input.
2438 Note that the callback function will only be called when creating
2439 a new session (session resumption can resume whatever was
2440 previously negotiated), and will not be called in SSL 2.0
2441 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2442 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2443 for applications that need to enforce opaque PRF input.
2447 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2450 [Victor B. Wagner <vitus@cryptocom.ru>]
2452 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2453 RFC4507bis. The encrypted ticket format is an encrypted encoded
2454 SSL_SESSION structure, that way new session features are automatically
2457 If a client application caches session in an SSL_SESSION structure
2458 support is transparent because tickets are now stored in the encoded
2461 The SSL_CTX structure automatically generates keys for ticket
2462 protection in servers so again support should be possible
2463 with no application modification.
2465 If a client or server wishes to disable RFC4507 support then the option
2466 SSL_OP_NO_TICKET can be set.
2468 Add a TLS extension debugging callback to allow the contents of any client
2469 or server extensions to be examined.
2471 This work was sponsored by Google.
2474 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2475 OpenSSL should now compile cleanly on gcc 4.2
2476 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2478 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2479 support including streaming MAC support: this is required for GOST
2480 ciphersuite support.
2481 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2483 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2484 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2485 to output in BER and PEM format.
2488 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2489 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2490 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2491 ENGINE support for HMAC keys which are unextractable. New -mac and
2492 -macopt options to dgst utility.
2495 *) New option -sigopt to dgst utility. Update dgst to use
2496 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2497 alternative signing paramaters such as X9.31 or PSS in the dgst
2501 *) Change ssl_cipher_apply_rule(), the internal function that does
2502 the work each time a ciphersuite string requests enabling
2503 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2504 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2505 the order of disabled ciphersuites such that those ciphersuites
2506 that most recently went from enabled to disabled not only stay
2507 in order with respect to each other, but also have higher priority
2508 than other disabled ciphersuites the next time ciphersuites are
2511 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2512 the same ciphersuites as with "HIGH" alone, but in a specific
2513 order where the PSK ciphersuites come first (since they are the
2514 most recently disabled ciphersuites when "HIGH" is parsed).
2516 Also, change ssl_create_cipher_list() (using this new
2517 funcionality) such that between otherwise identical
2518 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2522 *) Change ssl_create_cipher_list() so that it automatically
2523 arranges the ciphersuites in reasonable order before starting
2524 to process the rule string. Thus, the definition for "DEFAULT"
2525 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2526 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2527 This makes it much easier to arrive at a reasonable default order
2528 in applications for which anonymous ciphers are OK (meaning
2529 that you can't actually use DEFAULT).
2530 [Bodo Moeller; suggested by Victor Duchovni]
2532 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2533 processing) into multiple integers instead of setting
2534 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2535 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2536 (These masks as well as the individual bit definitions are hidden
2537 away into the non-exported interface ssl/ssl_locl.h, so this
2538 change to the definition of the SSL_CIPHER structure shouldn't
2539 affect applications.) This give us more bits for each of these
2540 categories, so there is no longer a need to coagulate AES128 and
2541 AES256 into a single algorithm bit, and to coagulate Camellia128
2542 and Camellia256 into a single algorithm bit, which has led to all
2545 Thus, among other things, the kludge introduced in 0.9.7m and
2546 0.9.8e for masking out AES256 independently of AES128 or masking
2547 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2549 With the change, we also introduce new ciphersuite aliases that
2550 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2554 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2555 Use the leftmost N bytes of the signature input if the input is
2556 larger than the prime q (with N being the size in bytes of q).
2559 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2560 it yet and it is largely untested.
2563 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2566 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2567 some compilers (gcc 4.2 and later) reject their use. Safestack is
2568 reimplemented. Update ASN1 to avoid use of legacy functions.
2571 *) Win32/64 targets are linked with Winsock2.
2574 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2575 to external functions. This can be used to increase CRL handling
2576 efficiency especially when CRLs are very large by (for example) storing
2577 the CRL revoked certificates in a database.
2580 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2581 new CRLs added to a directory can be used. New command line option
2582 -verify_return_error to s_client and s_server. This causes real errors
2583 to be returned by the verify callback instead of carrying on no matter
2584 what. This reflects the way a "real world" verify callback would behave.
2587 *) GOST engine, supporting several GOST algorithms and public key formats.
2588 Kindly donated by Cryptocom.
2591 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2592 partitioned by DP are handled but no indirect CRL or reason partitioning
2593 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2594 selected via a scoring technique which handles IDP and AKID in CRLs.
2597 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2598 will ultimately be used for all verify operations: this will remove the
2599 X509_STORE dependency on certificate verification and allow alternative
2600 lookup methods. X509_STORE based implementations of these two callbacks.
2603 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2604 Modify get_crl() to find a valid (unexpired) CRL if possible.
2607 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2608 this would be called X509_CRL_cmp() but that name is already used by
2609 a function that just compares CRL issuer names. Cache several CRL
2610 extensions in X509_CRL structure and cache CRLDP in X509.
2613 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2614 this maps equivalent X509_NAME structures into a consistent structure.
2615 Name comparison can then be performed rapidly using memcmp().
2618 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2622 *) Allow digests to supply their own micalg string for S/MIME type using
2623 the ctrl EVP_MD_CTRL_MICALG.
2626 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2627 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2628 ctrl. It can then customise the structure before and/or after signing
2632 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2633 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2634 to free up any added signature OIDs.
2637 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2638 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2639 digest and cipher tables. New options added to openssl utility:
2640 list-message-digest-algorithms and list-cipher-algorithms.
2643 *) Change the array representation of binary polynomials: the list
2644 of degrees of non-zero coefficients is now terminated with -1.
2645 Previously it was terminated with 0, which was also part of the
2646 value; thus, the array representation was not applicable to
2647 polynomials where t^0 has coefficient zero. This change makes
2648 the array representation useful in a more general context.
2651 *) Various modifications and fixes to SSL/TLS cipher string
2652 handling. For ECC, the code now distinguishes between fixed ECDH
2653 with RSA certificates on the one hand and with ECDSA certificates
2654 on the other hand, since these are separate ciphersuites. The
2655 unused code for Fortezza ciphersuites has been removed.
2657 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2658 (not "ECDHE"). For consistency with the code for DH
2659 certificates, use of ECDH certificates is now considered ECDH
2660 authentication, not RSA or ECDSA authentication (the latter is
2661 merely the CA's signing algorithm and not actively used in the
2664 The temporary ciphersuite alias "ECCdraft" is no longer
2665 available, and ECC ciphersuites are no longer excluded from "ALL"
2666 and "DEFAULT". The following aliases now exist for RFC 4492
2667 ciphersuites, most of these by analogy with the DH case:
2669 kECDHr - ECDH cert, signed with RSA
2670 kECDHe - ECDH cert, signed with ECDSA
2671 kECDH - ECDH cert (signed with either RSA or ECDSA)
2672 kEECDH - ephemeral ECDH
2673 ECDH - ECDH cert or ephemeral ECDH
2679 AECDH - anonymous ECDH
2680 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2684 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2685 Use correct micalg parameters depending on digest(s) in signed message.
2688 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2689 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2692 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2693 an engine to register a method. Add ENGINE lookups for methods and
2694 functional reference processing.
2697 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2698 EVP_{Sign,Verify}* which allow an application to customise the signature
2702 *) New -resign option to smime utility. This adds one or more signers
2703 to an existing PKCS#7 signedData structure. Also -md option to use an
2704 alternative message digest algorithm for signing.
2707 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2708 create PKCS7 structures containing multiple signers. Update smime
2709 application to support multiple signers.
2712 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2716 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2717 Reorganize PBE internals to lookup from a static table using NIDs,
2718 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2719 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2720 PRF which will be automatically used with PBES2.
2723 *) Replace the algorithm specific calls to generate keys in "req" with the
2727 *) Update PKCS#7 enveloped data routines to use new API. This is now
2728 supported by any public key method supporting the encrypt operation. A
2729 ctrl is added to allow the public key algorithm to examine or modify
2730 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2734 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2735 a default digest type to use. In most cases this will be SHA1 but some
2736 algorithms (such as GOST) need to specify an alternative digest. The
2737 return value indicates how strong the prefernce is 1 means optional and
2738 2 is mandatory (that is it is the only supported type). Modify
2739 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2740 use the default md. Update openssl utilities to use the default digest
2741 type for signing if it is not explicitly indicated.
2744 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2745 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2746 signing method from the key type. This effectively removes the link
2747 between digests and public key types.
2750 *) Add an OID cross reference table and utility functions. Its purpose is to
2751 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2752 rsaEncryption. This will allow some of the algorithm specific hackery
2753 needed to use the correct OID to be removed.
2756 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2757 structures for PKCS7_sign(). They are now set up by the relevant public
2761 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2764 *) Add support for key derivation (agreement) in the API, DH method and
2768 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2769 public and private key formats. As a side effect these add additional
2770 command line functionality not previously available: DSA signatures can be
2771 generated and verified using pkeyutl and DH key support and generation in
2776 [Oliver Tappe <zooey@hirschkaefer.de>]
2778 *) New make target "install_html_docs" installs HTML renditions of the
2780 [Oliver Tappe <zooey@hirschkaefer.de>]
2782 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
2783 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2784 support key and parameter generation and add initial key generation
2785 functionality for RSA.
2788 *) Add functions for main EVP_PKEY_method operations. The undocumented
2789 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2790 EVP_PKEY_{encrypt,decrypt}_old.
2793 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2794 key API, doesn't do much yet.
2797 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2798 public key algorithms. New option to openssl utility:
2799 "list-public-key-algorithms" to print out info.
2802 *) Implement the Supported Elliptic Curves Extension for
2803 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2806 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2807 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2810 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2811 utilities such as rsa, dsa, dsaparam etc except they process any key
2815 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2816 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2817 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2821 *) Initial support for pluggable public key ASN1.
2822 De-spaghettify the public key ASN1 handling. Move public and private
2823 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2824 algorithm specific handling to a single module within the relevant
2825 algorithm directory. Add functions to allow (near) opaque processing
2826 of public and private key structures.
2829 *) Implement the Supported Point Formats Extension for
2830 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2833 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2834 for the psk identity [hint] and the psk callback functions to the
2835 SSL_SESSION, SSL and SSL_CTX structure.
2838 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2842 SSL_CTX_use_psk_identity_hint
2843 SSL_get_psk_identity_hint
2844 SSL_get_psk_identity
2845 SSL_use_psk_identity_hint
2847 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2849 *) Add RFC 3161 compliant time stamp request creation, response generation
2850 and response verification functionality.
2851 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2853 *) Add initial support for TLS extensions, specifically for the server_name
2854 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2855 have new members for a host name. The SSL data structure has an
2856 additional member SSL_CTX *initial_ctx so that new sessions can be
2857 stored in that context to allow for session resumption, even after the
2858 SSL has been switched to a new SSL_CTX in reaction to a client's
2859 server_name extension.
2861 New functions (subject to change):
2863 SSL_get_servername()
2864 SSL_get_servername_type()
2867 New CTRL codes and macros (subject to change):
2869 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2870 - SSL_CTX_set_tlsext_servername_callback()
2871 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2872 - SSL_CTX_set_tlsext_servername_arg()
2873 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2875 openssl s_client has a new '-servername ...' option.
2877 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2878 '-key2 ...', '-servername_fatal' (subject to change). This allows
2879 testing the HostName extension for a specific single host name ('-cert'
2880 and '-key' remain fallbacks for handshakes without HostName
2881 negotiation). If the unrecogninzed_name alert has to be sent, this by
2882 default is a warning; it becomes fatal with the '-servername_fatal'
2885 [Peter Sylvester, Remy Allais, Christophe Renou]
2887 *) Whirlpool hash implementation is added.
2890 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2891 bn(64,32). Because of instruction set limitations it doesn't have
2892 any negative impact on performance. This was done mostly in order
2893 to make it possible to share assembler modules, such as bn_mul_mont
2894 implementations, between 32- and 64-bit builds without hassle.
2897 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2898 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2902 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2903 dedicated Montgomery multiplication procedure, is introduced.
2904 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2905 "64-bit" performance on certain 32-bit targets.
2908 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2909 in SSL structures. New SSL ctrl to set maximum send fragment size.
2910 Save memory by seeting the I/O buffer sizes dynamically instead of
2911 using the maximum available value.
2914 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2915 in addition to the text details.
2918 *) Very, very preliminary EXPERIMENTAL support for printing of general
2919 ASN1 structures. This currently produces rather ugly output and doesn't
2920 handle several customised structures at all.
2923 *) Integrated support for PVK file format and some related formats such
2924 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2925 these in the 'rsa' and 'dsa' utilities.
2928 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2931 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2932 place for the (very old) "NETSCAPE" format certificates which are now
2933 handled using new ASN1 code equivalents.
2936 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2937 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2938 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2941 *) Modify CRL distribution points extension code to print out previously
2942 unsupported fields. Enhance extension setting code to allow setting of
2946 *) Add print and set support for Issuing Distribution Point CRL extension.
2949 *) Change 'Configure' script to enable Camellia by default.
2952 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2954 *) When rejecting SSL/TLS records due to an incorrect version number, never
2955 update s->server with a new major version number. As of
2956 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2957 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2958 the previous behavior could result in a read attempt at NULL when
2959 receiving specific incorrect SSL/TLS records once record payload
2960 protection is active. (CVE-2010-0740)
2961 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2963 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2964 could be crashed if the relevant tables were not present (e.g. chrooted).
2965 [Tomas Hoger <thoger@redhat.com>]
2967 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2969 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2970 [Martin Olsson, Neel Mehta]
2972 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2973 accommodate for stack sorting, always a write lock!).
2976 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2977 excessive delays in the RAND_poll(): over a minute. As a workaround
2978 include a time check in the inner Heap32Next loop too.
2981 *) The code that handled flushing of data in SSL/TLS originally used the
2982 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2983 the problem outlined in PR#1949. The fix suggested there however can
2984 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2985 of Apache). So instead simplify the code to flush unconditionally.
2986 This should be fine since flushing with no data to flush is a no op.
2989 *) Handle TLS versions 2.0 and later properly and correctly use the
2990 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2991 off ancient servers have a habit of sticking around for a while...
2994 *) Modify compression code so it frees up structures without using the
2995 ex_data callbacks. This works around a problem where some applications
2996 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2997 restarting) then use compression (e.g. SSL with compression) later.
2998 This results in significant per-connection memory leaks and
2999 has caused some security issues including CVE-2008-1678 and
3003 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3004 change when encrypting or decrypting.
3007 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3008 connect and renegotiate with servers which do not support RI.
3009 Until RI is more widely deployed this option is enabled by default.
3012 *) Add "missing" ssl ctrls to clear options and mode.
3015 *) If client attempts to renegotiate and doesn't support RI respond with
3016 a no_renegotiation alert as required by RFC5746. Some renegotiating
3017 TLS clients will continue a connection gracefully when they receive
3018 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3019 waiting for a server hello which it will never receive. Now we treat a
3020 received no_renegotiation alert as a fatal error. This is because
3021 applications requesting a renegotiation might well expect it to succeed
3022 and would have no code in place to handle the server denying it so the
3023 only safe thing to do is to terminate the connection.
3026 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3027 peer supports secure renegotiation and 0 otherwise. Print out peer
3028 renegotiation support in s_client/s_server.
3031 *) Replace the highly broken and deprecated SPKAC certification method with
3032 the updated NID creation version. This should correctly handle UTF8.
3035 *) Implement RFC5746. Re-enable renegotiation but require the extension
3036 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3037 turns out to be a bad idea. It has been replaced by
3038 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3039 SSL_CTX_set_options(). This is really not recommended unless you
3040 know what you are doing.
3041 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3043 *) Fixes to stateless session resumption handling. Use initial_ctx when
3044 issuing and attempting to decrypt tickets in case it has changed during
3045 servername handling. Use a non-zero length session ID when attempting
3046 stateless session resumption: this makes it possible to determine if
3047 a resumption has occurred immediately after receiving server hello
3048 (several places in OpenSSL subtly assume this) instead of later in
3052 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3053 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3054 fixes for a few places where the return code is not checked
3056 [Julia Lawall <julia@diku.dk>]
3058 *) Add --strict-warnings option to Configure script to include devteam
3059 warnings in other configurations.
3062 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3063 makes it possible to install openssl libraries in locations which
3064 have names other than "lib", for example "/usr/lib64" which some
3066 [Steve Henson, based on patch from Jeremy Utley]
3068 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3069 X690 8.9.12 and can produce some misleading textual output of OIDs.
3070 [Steve Henson, reported by Dan Kaminsky]
3072 *) Delete MD2 from algorithm tables. This follows the recommendation in
3073 several standards that it is not used in new applications due to
3074 several cryptographic weaknesses. For binary compatibility reasons
3075 the MD2 API is still compiled in by default.
3078 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3082 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3083 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3085 [Guenter <lists@gknw.net>]
3087 *) Fix the server certificate chain building code to use X509_verify_cert(),
3088 it used to have an ad-hoc builder which was unable to cope with anything
3089 other than a simple chain.
3090 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3092 *) Don't check self signed certificate signatures in X509_verify_cert()
3093 by default (a flag can override this): it just wastes time without
3094 adding any security. As a useful side effect self signed root CAs
3095 with non-FIPS digests are now usable in FIPS mode.
3098 *) In dtls1_process_out_of_seq_message() the check if the current message
3099 is already buffered was missing. For every new message was memory
3100 allocated, allowing an attacker to perform an denial of service attack
3101 with sending out of seq handshake messages until there is no memory
3102 left. Additionally every future messege was buffered, even if the
3103 sequence number made no sense and would be part of another handshake.
3104 So only messages with sequence numbers less than 10 in advance will be
3105 buffered. (CVE-2009-1378)
3106 [Robin Seggelmann, discovered by Daniel Mentz]
3108 *) Records are buffered if they arrive with a future epoch to be
3109 processed after finishing the corresponding handshake. There is
3110 currently no limitation to this buffer allowing an attacker to perform
3111 a DOS attack with sending records with future epochs until there is no
3112 memory left. This patch adds the pqueue_size() function to detemine
3113 the size of a buffer and limits the record buffer to 100 entries.
3115 [Robin Seggelmann, discovered by Daniel Mentz]
3117 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3118 parent structure is freed. (CVE-2009-1379)
3121 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3122 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3125 [Ilya O. <vrghost@gmail.com>]
3127 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3129 *) Disable renegotiation completely - this fixes a severe security
3130 problem (CVE-2009-3555) at the cost of breaking all
3131 renegotiation. Renegotiation can be re-enabled by setting
3132 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3133 run-time. This is really not recommended unless you know what
3137 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3139 *) Don't set val to NULL when freeing up structures, it is freed up by
3140 underlying code. If sizeof(void *) > sizeof(long) this can result in
3141 zeroing past the valid field. (CVE-2009-0789)
3142 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3144 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3145 checked correctly. This would allow some invalid signed attributes to
3146 appear to verify correctly. (CVE-2009-0591)
3147 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3149 *) Reject UniversalString and BMPString types with invalid lengths. This
3150 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3151 a legal length. (CVE-2009-0590)
3154 *) Set S/MIME signing as the default purpose rather than setting it
3155 unconditionally. This allows applications to override it at the store
3159 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3160 to handle some structures.
3163 *) Improve efficiency of mem_gets: don't search whole buffer each time
3165 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3167 *) New -hex option for openssl rand.
3170 *) Print out UTF8String and NumericString when parsing ASN1.
3173 *) Support NumericString type for name components.
3176 *) Allow CC in the environment to override the automatically chosen
3177 compiler. Note that nothing is done to ensure flags work with the
3181 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3183 *) Properly check EVP_VerifyFinal() and similar return values
3185 [Ben Laurie, Bodo Moeller, Google Security Team]
3187 *) Enable TLS extensions by default.
3190 *) Allow the CHIL engine to be loaded, whether the application is
3191 multithreaded or not. (This does not release the developer from the
3192 obligation to set up the dynamic locking callbacks.)
3193 [Sander Temme <sander@temme.net>]
3195 *) Use correct exit code if there is an error in dgst command.
3196 [Steve Henson; problem pointed out by Roland Dirlewanger]
3198 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3199 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3202 *) Add experimental JPAKE support, including demo authentication in
3203 s_client and s_server.
3206 *) Set the comparison function in v3_addr_canonize().
3207 [Rob Austein <sra@hactrn.net>]
3209 *) Add support for XMPP STARTTLS in s_client.
3210 [Philip Paeps <philip@freebsd.org>]
3212 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3213 to ensure that even with this option, only ciphersuites in the
3214 server's preference list will be accepted. (Note that the option
3215 applies only when resuming a session, so the earlier behavior was
3216 just about the algorithm choice for symmetric cryptography.)
3219 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3221 *) Fix NULL pointer dereference if a DTLS server received
3222 ChangeCipherSpec as first record (CVE-2009-1386).
3225 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
3226 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3229 *) The fix in 0.9.8c that supposedly got rid of unsafe
3230 double-checked locking was incomplete for RSA blinding,
3231 addressing just one layer of what turns out to have been
3232 doubly unsafe triple-checked locking.
3234 So now fix this for real by retiring the MONT_HELPER macro
3235 in crypto/rsa/rsa_eay.c.
3237 [Bodo Moeller; problem pointed out by Marius Schilder]
3239 *) Various precautionary measures:
3241 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3243 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3244 (NB: This would require knowledge of the secret session ticket key
3245 to exploit, in which case you'd be SOL either way.)
3247 - Change bn_nist.c so that it will properly handle input BIGNUMs
3248 outside the expected range.
3250 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3253 [Neel Mehta, Bodo Moeller]
3255 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3256 the load fails. Useful for distros.
3257 [Ben Laurie and the FreeBSD team]
3259 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3262 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3265 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3267 This work was sponsored by Logica.
3270 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3271 keystores. Support for SSL/TLS client authentication too.
3272 Not compiled unless enable-capieng specified to Configure.
3274 This work was sponsored by Logica.
3277 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3278 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3279 attribute creation routines such as certifcate requests and PKCS#12
3283 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3285 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3286 handshake which could lead to a cilent crash as found using the
3287 Codenomicon TLS test suite (CVE-2008-1672)
3288 [Steve Henson, Mark Cox]
3290 *) Fix double free in TLS server name extensions which could lead to
3291 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3294 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3296 Clear the error queue to ensure that error entries left from
3297 older function calls do not interfere with the correct operation.
3298 [Lutz Jaenicke, Erik de Castro Lopo]
3300 *) Remove root CA certificates of commercial CAs:
3302 The OpenSSL project does not recommend any specific CA and does not
3303 have any policy with respect to including or excluding any CA.
3304 Therefore it does not make any sense to ship an arbitrary selection
3305 of root CA certificates with the OpenSSL software.
3308 *) RSA OAEP patches to fix two separate invalid memory reads.
3309 The first one involves inputs when 'lzero' is greater than
3310 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3311 before the beginning of from). The second one involves inputs where
3312 the 'db' section contains nothing but zeroes (there is a one-byte
3313 invalid read after the end of 'db').
3314 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3316 *) Partial backport from 0.9.9-dev:
3318 Introduce bn_mul_mont (dedicated Montgomery multiplication
3319 procedure) as a candidate for BIGNUM assembler implementation.
3320 While 0.9.9-dev uses assembler for various architectures, only
3321 x86_64 is available by default here in the 0.9.8 branch, and
3322 32-bit x86 is available through a compile-time setting.
3324 To try the 32-bit x86 assembler implementation, use Configure
3325 option "enable-montasm" (which exists only for this backport).
3327 As "enable-montasm" for 32-bit x86 disclaims code stability
3328 anyway, in this constellation we activate additional code
3329 backported from 0.9.9-dev for further performance improvements,
3330 namely BN_from_montgomery_word. (To enable this otherwise,
3331 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3333 [Andy Polyakov (backport partially by Bodo Moeller)]
3335 *) Add TLS session ticket callback. This allows an application to set
3336 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3337 values. This is useful for key rollover for example where several key
3338 sets may exist with different names.
3341 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3342 This was broken until now in 0.9.8 releases, such that the only way
3343 a registered ENGINE could be used (assuming it initialises
3344 successfully on the host) was to explicitly set it as the default
3345 for the relevant algorithms. This is in contradiction with 0.9.7
3346 behaviour and the documentation. With this fix, when an ENGINE is
3347 registered into a given algorithm's table of implementations, the
3348 'uptodate' flag is reset so that auto-discovery will be used next
3349 time a new context for that algorithm attempts to select an
3351 [Ian Lister (tweaked by Geoff Thorpe)]
3353 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3354 implemention in the following ways:
3356 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3359 Lack of BER streaming support means one pass streaming processing is
3360 only supported if data is detached: setting the streaming flag is
3361 ignored for embedded content.
3363 CMS support is disabled by default and must be explicitly enabled
3364 with the enable-cms configuration option.
3367 *) Update the GMP engine glue to do direct copies between BIGNUM and
3368 mpz_t when openssl and GMP use the same limb size. Otherwise the
3369 existing "conversion via a text string export" trick is still used.
3370 [Paul Sheer <paulsheer@gmail.com>]
3372 *) Zlib compression BIO. This is a filter BIO which compressed and
3373 uncompresses any data passed through it.
3376 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3377 RFC3394 compatible AES key wrapping.
3380 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3381 sets string data without copying. X509_ALGOR_set0() and
3382 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3383 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3384 from an X509_ATTRIBUTE structure optionally checking it occurs only
3385 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3389 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3390 to get the expected BN_FLG_CONSTTIME behavior.
3391 [Bodo Moeller (Google)]
3395 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3396 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3397 - added some more tests to do_tests.pl
3398 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3399 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3400 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3401 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3402 - various changes to netware.pl to enable gcc-cross builds on Win32
3404 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3405 - various changes to fix missing prototype warnings
3406 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3407 - added AES, WHIRLPOOL and CPUID assembler code to build files
3408 - added missing AES assembler make rules to mk1mf.pl
3409 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3410 [Guenter Knauf <eflash@gmx.net>]
3412 *) Implement certificate status request TLS extension defined in RFC3546.
3413 A client can set the appropriate parameters and receive the encoded
3414 OCSP response via a callback. A server can query the supplied parameters
3415 and set the encoded OCSP response in the callback. Add simplified examples
3416 to s_client and s_server.
3419 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3421 *) Fix various bugs:
3422 + Binary incompatibility of ssl_ctx_st structure
3423 + DTLS interoperation with non-compliant servers
3424 + Don't call get_session_cb() without proposed session
3425 + Fix ia64 assembler code
3426 [Andy Polyakov, Steve Henson]
3428 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3430 *) DTLS Handshake overhaul. There were longstanding issues with
3431 OpenSSL DTLS implementation, which were making it impossible for
3432 RFC 4347 compliant client to communicate with OpenSSL server.
3433 Unfortunately just fixing these incompatibilities would "cut off"
3434 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3435 server keeps tolerating non RFC compliant syntax. The opposite is
3436 not true, 0.9.8f client can not communicate with earlier server.
3437 This update even addresses CVE-2007-4995.
3440 *) Changes to avoid need for function casts in OpenSSL: some compilers
3441 (gcc 4.2 and later) reject their use.
3442 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3445 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3446 RFC4507bis. The encrypted ticket format is an encrypted encoded
3447 SSL_SESSION structure, that way new session features are automatically
3450 If a client application caches session in an SSL_SESSION structure
3451 support is transparent because tickets are now stored in the encoded
3454 The SSL_CTX structure automatically generates keys for ticket
3455 protection in servers so again support should be possible
3456 with no application modification.
3458 If a client or server wishes to disable RFC4507 support then the option
3459 SSL_OP_NO_TICKET can be set.
3461 Add a TLS extension debugging callback to allow the contents of any client
3462 or server extensions to be examined.
3464 This work was sponsored by Google.
3467 *) Add initial support for TLS extensions, specifically for the server_name
3468 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3469 have new members for a host name. The SSL data structure has an
3470 additional member SSL_CTX *initial_ctx so that new sessions can be
3471 stored in that context to allow for session resumption, even after the
3472 SSL has been switched to a new SSL_CTX in reaction to a client's
3473 server_name extension.
3475 New functions (subject to change):
3477 SSL_get_servername()
3478 SSL_get_servername_type()
3481 New CTRL codes and macros (subject to change):
3483 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3484 - SSL_CTX_set_tlsext_servername_callback()
3485 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3486 - SSL_CTX_set_tlsext_servername_arg()
3487 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3489 openssl s_client has a new '-servername ...' option.
3491 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3492 '-key2 ...', '-servername_fatal' (subject to change). This allows
3493 testing the HostName extension for a specific single host name ('-cert'
3494 and '-key' remain fallbacks for handshakes without HostName
3495 negotiation). If the unrecogninzed_name alert has to be sent, this by
3496 default is a warning; it becomes fatal with the '-servername_fatal'
3499 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3501 *) Add AES and SSE2 assembly language support to VC++ build.
3504 *) Mitigate attack on final subtraction in Montgomery reduction.
3507 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3508 (which previously caused an internal error).
3511 *) Squeeze another 10% out of IGE mode when in != out.
3514 *) AES IGE mode speedup.
3515 [Dean Gaudet (Google)]
3517 *) Add the Korean symmetric 128-bit cipher SEED (see
3518 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3519 add SEED ciphersuites from RFC 4162:
3521 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3522 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3523 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3524 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3526 To minimize changes between patchlevels in the OpenSSL 0.9.8
3527 series, SEED remains excluded from compilation unless OpenSSL
3528 is configured with 'enable-seed'.
3529 [KISA, Bodo Moeller]
3531 *) Mitigate branch prediction attacks, which can be practical if a
3532 single processor is shared, allowing a spy process to extract
3533 information. For detailed background information, see
3534 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3535 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3536 and Necessary Software Countermeasures"). The core of the change
3537 are new versions BN_div_no_branch() and
3538 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3539 respectively, which are slower, but avoid the security-relevant
3540 conditional branches. These are automatically called by BN_div()
3541 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3542 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3543 remove a conditional branch.
3545 BN_FLG_CONSTTIME is the new name for the previous
3546 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3547 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3548 in the exponent causes BN_mod_exp_mont() to use the alternative
3549 implementation in BN_mod_exp_mont_consttime().) The old name
3550 remains as a deprecated alias.
3552 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3553 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3554 constant-time implementations for more than just exponentiation.
3555 Here too the old name is kept as a deprecated alias.
3557 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3558 the BN_BLINDING structure gets an independent copy of the
3559 modulus. This means that the previous "BIGNUM *m" argument to
3560 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3561 essentially becomes "const BIGNUM *m", although we can't actually
3562 change this in the header file before 0.9.9. It allows
3563 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3564 enable BN_FLG_CONSTTIME.
3566 [Matthew D Wood (Intel Corp)]
3568 *) In the SSL/TLS server implementation, be strict about session ID
3569 context matching (which matters if an application uses a single
3570 external cache for different purposes). Previously,
3571 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3572 set. This did ensure strict client verification, but meant that,
3573 with applications using a single external cache for quite
3574 different requirements, clients could circumvent ciphersuite
3575 restrictions for a given session ID context by starting a session
3576 in a different context.
3579 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3580 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3581 authentication-only ciphersuites.
3584 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3585 not complete and could lead to a possible single byte overflow
3586 (CVE-2007-5135) [Ben Laurie]
3588 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3590 *) Since AES128 and AES256 (and similarly Camellia128 and
3591 Camellia256) share a single mask bit in the logic of
3592 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3593 kludge to work properly if AES128 is available and AES256 isn't
3594 (or if Camellia128 is available and Camellia256 isn't).
3597 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3598 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3599 When a point or a seed is encoded in a BIT STRING, we need to
3600 prevent the removal of trailing zero bits to get the proper DER
3601 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3602 of a NamedBitList, for which trailing 0 bits need to be removed.)
3605 *) Have SSL/TLS server implementation tolerate "mismatched" record
3606 protocol version while receiving ClientHello even if the
3607 ClientHello is fragmented. (The server can't insist on the
3608 particular protocol version it has chosen before the ServerHello
3609 message has informed the client about his choice.)
3612 *) Add RFC 3779 support.
3613 [Rob Austein for ARIN, Ben Laurie]
3615 *) Load error codes if they are not already present instead of using a
3616 static variable. This allows them to be cleanly unloaded and reloaded.
3617 Improve header file function name parsing.
3620 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3621 or CAPABILITY handshake as required by RFCs.
3624 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3626 *) Introduce limits to prevent malicious keys being able to
3627 cause a denial of service. (CVE-2006-2940)
3628 [Steve Henson, Bodo Moeller]
3630 *) Fix ASN.1 parsing of certain invalid structures that can result
3631 in a denial of service. (CVE-2006-2937) [Steve Henson]
3633 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3634 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3636 *) Fix SSL client code which could crash if connecting to a
3637 malicious SSLv2 server. (CVE-2006-4343)
3638 [Tavis Ormandy and Will Drewry, Google Security Team]
3640 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3641 match only those. Before that, "AES256-SHA" would be interpreted
3642 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3643 the same strength classification in 0.9.7h) as we currently only
3644 have a single AES bit in the ciphersuite description bitmap.
3645 That change, however, also applied to ciphersuite strings such as
3646 "RC4-MD5" that intentionally matched multiple ciphersuites --
3647 namely, SSL 2.0 ciphersuites in addition to the more common ones
3648 from SSL 3.0/TLS 1.0.
3650 So we change the selection algorithm again: Naming an explicit
3651 ciphersuite selects this one ciphersuite, and any other similar
3652 ciphersuite (same bitmap) from *other* protocol versions.
3653 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3654 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3656 Since SSL 2.0 does not have any ciphersuites for which the
3657 128/256 bit distinction would be relevant, this works for now.
3658 The proper fix will be to use different bits for AES128 and
3659 AES256, which would have avoided the problems from the beginning;
3660 however, bits are scarce, so we can only do this in a new release
3661 (not just a patchlevel) when we can change the SSL_CIPHER
3662 definition to split the single 'unsigned long mask' bitmap into
3663 multiple values to extend the available space.
3667 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3669 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3670 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3672 *) Add AES IGE and biIGE modes.
3675 *) Change the Unix randomness entropy gathering to use poll() when
3676 possible instead of select(), since the latter has some
3677 undesirable limitations.
3678 [Darryl Miles via Richard Levitte and Bodo Moeller]
3680 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3681 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3682 cannot be implicitly activated as part of, e.g., the "AES" alias.
3683 However, please upgrade to OpenSSL 0.9.9[-dev] for
3684 non-experimental use of the ECC ciphersuites to get TLS extension
3685 support, which is required for curve and point format negotiation
3686 to avoid potential handshake problems.
3689 *) Disable rogue ciphersuites:
3691 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3692 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3693 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3695 The latter two were purportedly from
3696 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3699 Also deactivate the remaining ciphersuites from
3700 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3701 unofficial, and the ID has long expired.
3704 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3705 dual-core machines) and other potential thread-safety issues.
3708 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3709 versions), which is now available for royalty-free use
3710 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3711 Also, add Camellia TLS ciphersuites from RFC 4132.
3713 To minimize changes between patchlevels in the OpenSSL 0.9.8
3714 series, Camellia remains excluded from compilation unless OpenSSL
3715 is configured with 'enable-camellia'.
3718 *) Disable the padding bug check when compression is in use. The padding
3719 bug check assumes the first packet is of even length, this is not
3720 necessarily true if compresssion is enabled and can result in false
3721 positives causing handshake failure. The actual bug test is ancient
3722 code so it is hoped that implementations will either have fixed it by
3723 now or any which still have the bug do not support compression.
3726 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3728 *) When applying a cipher rule check to see if string match is an explicit
3729 cipher suite and only match that one cipher suite if it is.
3732 *) Link in manifests for VC++ if needed.
3733 [Austin Ziegler <halostatue@gmail.com>]
3735 *) Update support for ECC-based TLS ciphersuites according to
3736 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3737 TLS extensions, which are supported starting with the 0.9.9
3738 branch, not in the OpenSSL 0.9.8 branch).
3741 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3742 opaque EVP_CIPHER_CTX handling.
3745 *) Fixes and enhancements to zlib compression code. We now only use
3746 "zlib1.dll" and use the default __cdecl calling convention on Win32
3747 to conform with the standards mentioned here:
3748 http://www.zlib.net/DLL_FAQ.txt
3749 Static zlib linking now works on Windows and the new --with-zlib-include
3750 --with-zlib-lib options to Configure can be used to supply the location
3751 of the headers and library. Gracefully handle case where zlib library
3755 *) Several fixes and enhancements to the OID generation code. The old code
3756 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3757 handle numbers larger than ULONG_MAX, truncated printing and had a
3758 non standard OBJ_obj2txt() behaviour.
3761 *) Add support for building of engines under engine/ as shared libraries
3762 under VC++ build system.
3765 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3766 Hopefully, we will not see any false combination of paths any more.
3769 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3771 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3772 (part of SSL_OP_ALL). This option used to disable the
3773 countermeasure against man-in-the-middle protocol-version
3774 rollback in the SSL 2.0 server implementation, which is a bad
3775 idea. (CVE-2005-2969)
3777 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3778 for Information Security, National Institute of Advanced Industrial
3779 Science and Technology [AIST], Japan)]
3781 *) Add two function to clear and return the verify parameter flags.
3784 *) Keep cipherlists sorted in the source instead of sorting them at
3785 runtime, thus removing the need for a lock.
3788 *) Avoid some small subgroup attacks in Diffie-Hellman.
3789 [Nick Mathewson and Ben Laurie]
3791 *) Add functions for well-known primes.
3794 *) Extended Windows CE support.
3795 [Satoshi Nakamura and Andy Polyakov]
3797 *) Initialize SSL_METHOD structures at compile time instead of during
3798 runtime, thus removing the need for a lock.
3801 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3802 attempting to decrypt each encrypted key in turn. Add support to
3806 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3808 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3811 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3814 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3815 key into the same file any more.
3818 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3821 *) Add -utf8 command line and config file option to 'ca'.
3822 [Stefan <stf@udoma.org]
3824 *) Removed the macro des_crypt(), as it seems to conflict with some
3825 libraries. Use DES_crypt().
3828 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3829 involves renaming the source and generated shared-libs for
3830 both. The engines will accept the corrected or legacy ids
3831 ('ncipher' and '4758_cca' respectively) when binding. NB,
3832 this only applies when building 'shared'.
3833 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3835 *) Add attribute functions to EVP_PKEY structure. Modify
3836 PKCS12_create() to recognize a CSP name attribute and
3837 use it. Make -CSP option work again in pkcs12 utility.
3840 *) Add new functionality to the bn blinding code:
3841 - automatic re-creation of the BN_BLINDING parameters after
3842 a fixed number of uses (currently 32)
3843 - add new function for parameter creation
3844 - introduce flags to control the update behaviour of the
3845 BN_BLINDING parameters
3846 - hide BN_BLINDING structure
3847 Add a second BN_BLINDING slot to the RSA structure to improve
3848 performance when a single RSA object is shared among several
3852 *) Add support for DTLS.
3853 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3855 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3856 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3859 *) Remove buggy and incompletet DH cert support from
3860 ssl/ssl_rsa.c and ssl/s3_both.c
3863 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3864 the apps/openssl applications.
3867 *) Compile clean with "-Wall -Wmissing-prototypes
3868 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3869 DEBUG_SAFESTACK must also be set.
3872 *) Change ./Configure so that certain algorithms can be disabled by default.
3873 The new counterpiece to "no-xxx" is "enable-xxx".
3875 The patented RC5 and MDC2 algorithms will now be disabled unless
3876 "enable-rc5" and "enable-mdc2", respectively, are specified.
3878 (IDEA remains enabled despite being patented. This is because IDEA
3879 is frequently required for interoperability, and there is no license
3880 fee for non-commercial use. As before, "no-idea" can be used to
3881 avoid this algorithm.)
3885 *) Add processing of proxy certificates (see RFC 3820). This work was
3886 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3887 EGEE (Enabling Grids for E-science in Europe).
3890 *) RC4 performance overhaul on modern architectures/implementations, such
3891 as Intel P4, IA-64 and AMD64.
3894 *) New utility extract-section.pl. This can be used specify an alternative
3895 section number in a pod file instead of having to treat each file as
3896 a separate case in Makefile. This can be done by adding two lines to the
3899 =for comment openssl_section:XXX
3901 The blank line is mandatory.
3905 *) New arguments -certform, -keyform and -pass for s_client and s_server
3906 to allow alternative format key and certificate files and passphrase
3910 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3911 update associated structures and add various utility functions.
3913 Add new policy related verify parameters, include policy checking in
3914 standard verify code. Enhance 'smime' application with extra parameters
3915 to support policy checking and print out.
3918 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3919 Nehemiah processors. These extensions support AES encryption in hardware
3920 as well as RNG (though RNG support is currently disabled).
3921 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3923 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3926 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3927 [Andy Polyakov and a number of other people]
3929 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3930 implementation contributed by IBM.
3931 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3933 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3934 exponent rather than 'unsigned long'. There is a corresponding change to
3935 the new 'rsa_keygen' element of the RSA_METHOD structure.
3936 [Jelte Jansen, Geoff Thorpe]
3938 *) Functionality for creating the initial serial number file is now
3939 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3941 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3942 number file to 1, which is bound to cause problems. To avoid
3943 the problems while respecting compatibility between different 0.9.7
3944 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3945 CA.pl for serial number initialization. With the new release 0.9.8,
3946 we can fix the problem directly in the 'ca' utility.)
3949 *) Reduced header interdepencies by declaring more opaque objects in
3950 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3951 give fewer recursive includes, which could break lazy source code - so
3952 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3953 developers should define this symbol when building and using openssl to
3954 ensure they track the recommended behaviour, interfaces, [etc], but
3955 backwards-compatible behaviour prevails when this isn't defined.
3958 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3961 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3962 This will generate a random key of the appropriate length based on the
3963 cipher context. The EVP_CIPHER can provide its own random key generation
3964 routine to support keys of a specific form. This is used in the des and
3965 3des routines to generate a key of the correct parity. Update S/MIME
3966 code to use new functions and hence generate correct parity DES keys.
3967 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3968 valid (weak or incorrect parity).
3971 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3972 as looking them up. This is useful when the verified structure may contain
3973 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3974 present unless the new PKCS7_NO_CRL flag is asserted.
3977 *) Extend ASN1 oid configuration module. It now additionally accepts the
3980 shortName = some long name, 1.2.3.4
3983 *) Reimplemented the BN_CTX implementation. There is now no more static
3984 limitation on the number of variables it can handle nor the depth of the
3985 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3986 information can now expand as required, and rather than having a single
3987 static array of bignums, BN_CTX now uses a linked-list of such arrays
3988 allowing it to expand on demand whilst maintaining the usefulness of
3989 BN_CTX's "bundling".
3992 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3993 to allow all RSA operations to function using a single BN_CTX.
3996 *) Preliminary support for certificate policy evaluation and checking. This
3997 is initially intended to pass the tests outlined in "Conformance Testing
3998 of Relying Party Client Certificate Path Processing Logic" v1.07.
4001 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4002 remained unused and not that useful. A variety of other little bignum
4003 tweaks and fixes have also been made continuing on from the audit (see
4007 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4008 associated ASN1, EVP and SSL functions and old ASN1 macros.
4011 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4012 and this should never fail. So the return value from the use of
4013 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4014 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4017 *) BN_CTX_get() should return zero-valued bignums, providing the same
4018 initialised value as BN_new().
4019 [Geoff Thorpe, suggested by Ulf Möller]
4021 *) Support for inhibitAnyPolicy certificate extension.
4024 *) An audit of the BIGNUM code is underway, for which debugging code is
4025 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4026 is considered valid when processing BIGNUMs, and causes execution to
4027 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4028 further steps are taken to deliberately pollute unused data in BIGNUM
4029 structures to try and expose faulty code further on. For now, openssl will
4030 (in its default mode of operation) continue to tolerate the inconsistent
4031 forms that it has tolerated in the past, but authors and packagers should
4032 consider trying openssl and their own applications when compiled with
4033 these debugging symbols defined. It will help highlight potential bugs in
4034 their own code, and will improve the test coverage for OpenSSL itself. At
4035 some point, these tighter rules will become openssl's default to improve
4036 maintainability, though the assert()s and other overheads will remain only
4037 in debugging configurations. See bn.h for more details.
4038 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4040 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4041 that can only be obtained through BN_CTX_new() (which implicitly
4042 initialises it). The presence of this function only made it possible
4043 to overwrite an existing structure (and cause memory leaks).
4046 *) Because of the callback-based approach for implementing LHASH as a
4047 template type, lh_insert() adds opaque objects to hash-tables and
4048 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4049 to clean up those corresponding objects before destroying the hash table
4050 (and losing the object pointers). So some over-zealous constifications in
4051 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4052 objects as "const" and the lh_doall[_arg] callback wrappers are not
4053 prototyped to have "const" restrictions on the object pointers they are
4054 given (and so aren't required to cast them away any more).
4057 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4058 (speed) prefers to use its own implementation. The two implementations
4059 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4060 its object type properly exposed (MS_TM) instead of casting to/from "char
4061 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4062 aren't necessarily the greatest nomenclatures - but this is what was used
4063 internally to the implementation so I've used that for now.
4066 *) Ensure that deprecated functions do not get compiled when
4067 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4068 the self-tests were still using deprecated key-generation functions so
4069 these have been updated also.
4072 *) Reorganise PKCS#7 code to separate the digest location functionality
4073 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
4074 New function PKCS7_set_digest() to set the digest type for PKCS#7
4075 digestedData type. Add additional code to correctly generate the
4076 digestedData type and add support for this type in PKCS7 initialization
4080 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4081 structure of type "other".
4084 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4085 sure the loop does correctly stop and breaking ("division by zero")
4086 modulus operations are not performed. The (pre-generated) prime
4087 table crypto/bn/bn_prime.h was already correct, but it could not be
4088 re-generated on some platforms because of the "division by zero"
4089 situation in the script.
4090 [Ralf S. Engelschall]
4092 *) Update support for ECC-based TLS ciphersuites according to
4093 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4094 SHA-1 now is only used for "small" curves (where the
4095 representation of a field element takes up to 24 bytes); for
4096 larger curves, the field element resulting from ECDH is directly
4097 used as premaster secret.
4098 [Douglas Stebila (Sun Microsystems Laboratories)]
4100 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4101 curve secp160r1 to the tests.
4102 [Douglas Stebila (Sun Microsystems Laboratories)]
4104 *) Add the possibility to load symbols globally with DSO.
4105 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4107 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4108 control of the error stack.
4111 *) Add support for STORE in ENGINE.
4114 *) Add the STORE type. The intention is to provide a common interface
4115 to certificate and key stores, be they simple file-based stores, or
4116 HSM-type store, or LDAP stores, or...
4117 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4120 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4121 pass a list of arguments to any function as well as provide a way
4122 for a function to pass data back to the caller.
4125 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4126 works like BUF_strdup() but can be used to duplicate a portion of
4127 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4131 *) Add the function sk_find_ex() which works like sk_find(), but will
4132 return an index to an element even if an exact match couldn't be
4133 found. The index is guaranteed to point at the element where the
4134 searched-for key would be inserted to preserve sorting order.
4137 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4138 takes an extra flags argument for optional functionality. Currently,
4139 the following flags are defined:
4141 OBJ_BSEARCH_VALUE_ON_NOMATCH
4142 This one gets OBJ_bsearch_ex() to return a pointer to the first
4143 element where the comparing function returns a negative or zero
4146 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4147 This one gets OBJ_bsearch_ex() to return a pointer to the first
4148 element where the comparing function returns zero. This is useful
4149 if there are more than one element where the comparing function
4153 *) Make it possible to create self-signed certificates with 'openssl ca'
4154 in such a way that the self-signed certificate becomes part of the
4155 CA database and uses the same mechanisms for serial number generation
4156 as all other certificate signing. The new flag '-selfsign' enables
4157 this functionality. Adapt CA.sh and CA.pl.in.
4160 *) Add functionality to check the public key of a certificate request
4161 against a given private. This is useful to check that a certificate
4162 request can be signed by that key (self-signing).
4165 *) Make it possible to have multiple active certificates with the same
4166 subject in the CA index file. This is done only if the keyword
4167 'unique_subject' is set to 'no' in the main CA section (default
4168 if 'CA_default') of the configuration file. The value is saved
4169 with the database itself in a separate index attribute file,
4170 named like the index file with '.attr' appended to the name.
4173 *) Generate muti valued AVAs using '+' notation in config files for
4177 *) Support for nameConstraints certificate extension.
4180 *) Support for policyConstraints certificate extension.
4183 *) Support for policyMappings certificate extension.
4186 *) Make sure the default DSA_METHOD implementation only uses its
4187 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4188 and change its own handlers to be NULL so as to remove unnecessary
4189 indirection. This lets alternative implementations fallback to the
4190 default implementation more easily.
4193 *) Support for directoryName in GeneralName related extensions
4197 *) Make it possible to link applications using Makefile.shared.
4198 Make that possible even when linking against static libraries!
4201 *) Support for single pass processing for S/MIME signing. This now
4202 means that S/MIME signing can be done from a pipe, in addition
4203 cleartext signing (multipart/signed type) is effectively streaming
4204 and the signed data does not need to be all held in memory.
4206 This is done with a new flag PKCS7_STREAM. When this flag is set
4207 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4208 is done after the data is output (and digests calculated) in
4209 SMIME_write_PKCS7().
4212 *) Add full support for -rpath/-R, both in shared libraries and
4213 applications, at least on the platforms where it's known how
4217 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4218 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4219 will now compute a table of multiples of the generator that
4220 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4221 faster (notably in the case of a single point multiplication,
4222 scalar * generator).
4223 [Nils Larsch, Bodo Moeller]
4225 *) IPv6 support for certificate extensions. The various extensions
4226 which use the IP:a.b.c.d can now take IPv6 addresses using the
4227 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4231 *) Added an ENGINE that implements RSA by performing private key
4232 exponentiations with the GMP library. The conversions to and from
4233 GMP's mpz_t format aren't optimised nor are any montgomery forms
4234 cached, and on x86 it appears OpenSSL's own performance has caught up.
4235 However there are likely to be other architectures where GMP could
4236 provide a boost. This ENGINE is not built in by default, but it can be
4237 specified at Configure time and should be accompanied by the necessary
4238 linker additions, eg;
4239 ./config -DOPENSSL_USE_GMP -lgmp
4242 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4243 testing availability of engines with "-t" - the old behaviour is
4244 produced by increasing the feature's verbosity with "-tt".
4247 *) ECDSA routines: under certain error conditions uninitialized BN objects
4248 could be freed. Solution: make sure initialization is performed early
4249 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4253 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4254 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4255 software implementations. For DSA and DH, parameter generation can
4256 also be overriden by providing the appropriate method callbacks.
4259 *) Change the "progress" mechanism used in key-generation and
4260 primality testing to functions that take a new BN_GENCB pointer in
4261 place of callback/argument pairs. The new API functions have "_ex"
4262 postfixes and the older functions are reimplemented as wrappers for
4263 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4264 declarations of the old functions to help (graceful) attempts to
4265 migrate to the new functions. Also, the new key-generation API
4266 functions operate on a caller-supplied key-structure and return
4267 success/failure rather than returning a key or NULL - this is to
4268 help make "keygen" another member function of RSA_METHOD etc.
4270 Example for using the new callback interface:
4272 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4276 BN_GENCB_set(&my_cb, my_callback, my_arg);
4278 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4279 /* For the meaning of a, b in calls to my_callback(), see the
4280 * documentation of the function that calls the callback.
4281 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4282 * my_callback should return 1 if it wants BN_is_prime_ex()
4283 * to continue, or 0 to stop.
4288 *) Change the ZLIB compression method to be stateful, and make it
4289 available to TLS with the number defined in
4290 draft-ietf-tls-compression-04.txt.
4293 *) Add the ASN.1 structures and functions for CertificatePair, which
4294 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4296 CertificatePair ::= SEQUENCE {
4297 forward [0] Certificate OPTIONAL,
4298 reverse [1] Certificate OPTIONAL,
4299 -- at least one of the pair shall be present -- }
4301 Also implement the PEM functions to read and write certificate
4302 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4304 This needed to be defined, mostly for the sake of the LDAP
4305 attribute crossCertificatePair, but may prove useful elsewhere as
4309 *) Make it possible to inhibit symlinking of shared libraries in
4310 Makefile.shared, for Cygwin's sake.
4313 *) Extend the BIGNUM API by creating a function
4314 void BN_set_negative(BIGNUM *a, int neg);
4315 and a macro that behave like
4316 int BN_is_negative(const BIGNUM *a);
4318 to avoid the need to access 'a->neg' directly in applications.
4321 *) Implement fast modular reduction for pseudo-Mersenne primes
4322 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4323 EC_GROUP_new_curve_GFp() will now automatically use this
4325 [Nils Larsch <nla@trustcenter.de>]
4327 *) Add new lock type (CRYPTO_LOCK_BN).
4330 *) Change the ENGINE framework to automatically load engines
4331 dynamically from specific directories unless they could be
4332 found to already be built in or loaded. Move all the
4333 current engines except for the cryptodev one to a new
4335 The engines in engines/ are built as shared libraries if
4336 the "shared" options was given to ./Configure or ./config.
4337 Otherwise, they are inserted in libcrypto.a.
4338 /usr/local/ssl/engines is the default directory for dynamic
4339 engines, but that can be overriden at configure time through
4340 the usual use of --prefix and/or --openssldir, and at run
4341 time with the environment variable OPENSSL_ENGINES.
4342 [Geoff Thorpe and Richard Levitte]
4344 *) Add Makefile.shared, a helper makefile to build shared
4345 libraries. Addapt Makefile.org.
4348 *) Add version info to Win32 DLLs.
4349 [Peter 'Luna' Runestig" <peter@runestig.com>]
4351 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4352 can be added using this API to created arbitrary PKCS#12
4353 files while avoiding the low level API.
4355 New options to PKCS12_create(), key or cert can be NULL and
4356 will then be omitted from the output file. The encryption
4357 algorithm NIDs can be set to -1 for no encryption, the mac
4358 iteration count can be set to 0 to omit the mac.
4360 Enhance pkcs12 utility by making the -nokeys and -nocerts
4361 options work when creating a PKCS#12 file. New option -nomac
4362 to omit the mac, NONE can be set for an encryption algorithm.
4363 New code is modified to use the enhanced PKCS12_create()
4364 instead of the low level API.
4367 *) Extend ASN1 encoder to support indefinite length constructed
4368 encoding. This can output sequences tags and octet strings in
4369 this form. Modify pk7_asn1.c to support indefinite length
4370 encoding. This is experimental and needs additional code to
4371 be useful, such as an ASN1 bio and some enhanced streaming
4374 Extend template encode functionality so that tagging is passed
4375 down to the template encoder.
4378 *) Let 'openssl req' fail if an argument to '-newkey' is not
4379 recognized instead of using RSA as a default.
4382 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4383 As these are not official, they are not included in "ALL";
4384 the "ECCdraft" ciphersuite group alias can be used to select them.
4385 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4387 *) Add ECDH engine support.
4388 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4390 *) Add ECDH in new directory crypto/ecdh/.
4391 [Douglas Stebila (Sun Microsystems Laboratories)]
4393 *) Let BN_rand_range() abort with an error after 100 iterations
4394 without success (which indicates a broken PRNG).
4397 *) Change BN_mod_sqrt() so that it verifies that the input value
4398 is really the square of the return value. (Previously,
4399 BN_mod_sqrt would show GIGO behaviour.)
4402 *) Add named elliptic curves over binary fields from X9.62, SECG,
4403 and WAP/WTLS; add OIDs that were still missing.
4405 [Sheueling Chang Shantz and Douglas Stebila
4406 (Sun Microsystems Laboratories)]
4408 *) Extend the EC library for elliptic curves over binary fields
4409 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4412 EC_GF2m_simple_method
4416 EC_GROUP_new_curve_GF2m
4417 EC_GROUP_set_curve_GF2m
4418 EC_GROUP_get_curve_GF2m
4419 EC_POINT_set_affine_coordinates_GF2m
4420 EC_POINT_get_affine_coordinates_GF2m
4421 EC_POINT_set_compressed_coordinates_GF2m
4423 Point compression for binary fields is disabled by default for
4424 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4427 As binary polynomials are represented as BIGNUMs, various members
4428 of the EC_GROUP and EC_POINT data structures can be shared
4429 between the implementations for prime fields and binary fields;
4430 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4431 are essentially identical to their ..._GFp counterparts.
4432 (For simplicity, the '..._GFp' prefix has been dropped from
4433 various internal method names.)
4435 An internal 'field_div' method (similar to 'field_mul' and
4436 'field_sqr') has been added; this is used only for binary fields.
4438 [Sheueling Chang Shantz and Douglas Stebila
4439 (Sun Microsystems Laboratories)]
4441 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4442 through methods ('mul', 'precompute_mult').
4444 The generic implementations (now internally called 'ec_wNAF_mul'
4445 and 'ec_wNAF_precomputed_mult') remain the default if these
4446 methods are undefined.
4448 [Sheueling Chang Shantz and Douglas Stebila
4449 (Sun Microsystems Laboratories)]
4451 *) New function EC_GROUP_get_degree, which is defined through
4452 EC_METHOD. For curves over prime fields, this returns the bit
4453 length of the modulus.
4455 [Sheueling Chang Shantz and Douglas Stebila
4456 (Sun Microsystems Laboratories)]
4458 *) New functions EC_GROUP_dup, EC_POINT_dup.
4459 (These simply call ..._new and ..._copy).
4461 [Sheueling Chang Shantz and Douglas Stebila
4462 (Sun Microsystems Laboratories)]
4464 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4465 Polynomials are represented as BIGNUMs (where the sign bit is not
4466 used) in the following functions [macros]:
4469 BN_GF2m_sub [= BN_GF2m_add]
4470 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4471 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4472 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4474 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4475 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4476 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4477 BN_GF2m_cmp [= BN_ucmp]
4479 (Note that only the 'mod' functions are actually for fields GF(2^m).
4480 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4482 For some functions, an the irreducible polynomial defining a
4483 field can be given as an 'unsigned int[]' with strictly
4484 decreasing elements giving the indices of those bits that are set;
4485 i.e., p[] represents the polynomial
4486 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4488 p[0] > p[1] > ... > p[k] = 0.
4489 This applies to the following functions:
4494 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4495 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4497 BN_GF2m_mod_sqrt_arr
4498 BN_GF2m_mod_solve_quad_arr
4502 Conversion can be performed by the following functions:
4507 bntest.c has additional tests for binary polynomial arithmetic.
4509 Two implementations for BN_GF2m_mod_div() are available.
4510 The default algorithm simply uses BN_GF2m_mod_inv() and
4511 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4512 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4513 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4515 [Sheueling Chang Shantz and Douglas Stebila
4516 (Sun Microsystems Laboratories)]
4518 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4519 functionality is disabled at compile-time.
4520 [Douglas Stebila <douglas.stebila@sun.com>]
4522 *) Change default behaviour of 'openssl asn1parse' so that more
4523 information is visible when viewing, e.g., a certificate:
4525 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4526 mode the content of non-printable OCTET STRINGs is output in a
4527 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4528 avoid the appearance of a printable string.
4529 [Nils Larsch <nla@trustcenter.de>]
4531 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4533 EC_GROUP_set_asn1_flag()
4534 EC_GROUP_get_asn1_flag()
4535 EC_GROUP_set_point_conversion_form()
4536 EC_GROUP_get_point_conversion_form()
4537 These control ASN1 encoding details:
4538 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4539 has been set to OPENSSL_EC_NAMED_CURVE.
4540 - Points are encoded in uncompressed form by default; options for
4541 asn1_for are as for point2oct, namely
4542 POINT_CONVERSION_COMPRESSED
4543 POINT_CONVERSION_UNCOMPRESSED
4544 POINT_CONVERSION_HYBRID
4546 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4549 EC_GROUP_get0_seed()
4550 EC_GROUP_get_seed_len()
4551 This is used only for ASN1 purposes (so far).
4552 [Nils Larsch <nla@trustcenter.de>]
4554 *) Add 'field_type' member to EC_METHOD, which holds the NID
4555 of the appropriate field type OID. The new function
4556 EC_METHOD_get_field_type() returns this value.
4557 [Nils Larsch <nla@trustcenter.de>]
4562 EC_POINT_point2hex()
4563 EC_POINT_hex2point()
4564 providing useful interfaces to EC_POINT_point2oct() and
4565 EC_POINT_oct2point().
4566 [Nils Larsch <nla@trustcenter.de>]
4568 *) Change internals of the EC library so that the functions
4569 EC_GROUP_set_generator()
4570 EC_GROUP_get_generator()
4571 EC_GROUP_get_order()
4572 EC_GROUP_get_cofactor()
4573 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4574 to methods, which would lead to unnecessary code duplication when
4575 adding different types of curves.
4576 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4578 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4579 arithmetic, and such that modified wNAFs are generated
4580 (which avoid length expansion in many cases).
4583 *) Add a function EC_GROUP_check_discriminant() (defined via
4584 EC_METHOD) that verifies that the curve discriminant is non-zero.
4586 Add a function EC_GROUP_check() that makes some sanity tests
4587 on a EC_GROUP, its generator and order. This includes
4588 EC_GROUP_check_discriminant().
4589 [Nils Larsch <nla@trustcenter.de>]
4591 *) Add ECDSA in new directory crypto/ecdsa/.
4593 Add applications 'openssl ecparam' and 'openssl ecdsa'
4594 (these are based on 'openssl dsaparam' and 'openssl dsa').
4596 ECDSA support is also included in various other files across the
4597 library. Most notably,
4598 - 'openssl req' now has a '-newkey ecdsa:file' option;
4599 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4600 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4601 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4602 them suitable for ECDSA where domain parameters must be
4603 extracted before the specific public key;
4604 - ECDSA engine support has been added.
4605 [Nils Larsch <nla@trustcenter.de>]
4607 *) Include some named elliptic curves, and add OIDs from X9.62,
4608 SECG, and WAP/WTLS. Each curve can be obtained from the new
4610 EC_GROUP_new_by_curve_name(),
4611 and the list of available named curves can be obtained with
4612 EC_get_builtin_curves().
4613 Also add a 'curve_name' member to EC_GROUP objects, which can be
4615 EC_GROUP_set_curve_name()
4616 EC_GROUP_get_curve_name()
4617 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4619 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4620 was actually never needed) and in BN_mul(). The removal in BN_mul()
4621 required a small change in bn_mul_part_recursive() and the addition
4622 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4623 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4624 bn_sub_words() and bn_add_words() except they take arrays with
4628 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4630 *) Cleanse PEM buffers before freeing them since they may contain
4632 [Benjamin Bennett <ben@psc.edu>]
4634 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4635 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4636 authentication-only ciphersuites.
4639 *) Since AES128 and AES256 share a single mask bit in the logic of
4640 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4641 kludge to work properly if AES128 is available and AES256 isn't.
4644 *) Expand security boundary to match 1.1.1 module.
4647 *) Remove redundant features: hash file source, editing of test vectors
4648 modify fipsld to use external fips_premain.c signature.
4651 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4652 run algorithm test programs.
4655 *) Make algorithm test programs more tolerant of whitespace.
4658 *) Have SSL/TLS server implementation tolerate "mismatched" record
4659 protocol version while receiving ClientHello even if the
4660 ClientHello is fragmented. (The server can't insist on the
4661 particular protocol version it has chosen before the ServerHello
4662 message has informed the client about his choice.)
4665 *) Load error codes if they are not already present instead of using a
4666 static variable. This allows them to be cleanly unloaded and reloaded.
4669 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4671 *) Introduce limits to prevent malicious keys being able to
4672 cause a denial of service. (CVE-2006-2940)
4673 [Steve Henson, Bodo Moeller]
4675 *) Fix ASN.1 parsing of certain invalid structures that can result
4676 in a denial of service. (CVE-2006-2937) [Steve Henson]
4678 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4679 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4681 *) Fix SSL client code which could crash if connecting to a
4682 malicious SSLv2 server. (CVE-2006-4343)
4683 [Tavis Ormandy and Will Drewry, Google Security Team]
4685 *) Change ciphersuite string processing so that an explicit
4686 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4687 will no longer include "AES128-SHA"), and any other similar
4688 ciphersuite (same bitmap) from *other* protocol versions (so that
4689 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4690 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4691 changes from 0.9.8b and 0.9.8d.
4694 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4696 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4697 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4699 *) Change the Unix randomness entropy gathering to use poll() when
4700 possible instead of select(), since the latter has some
4701 undesirable limitations.
4702 [Darryl Miles via Richard Levitte and Bodo Moeller]
4704 *) Disable rogue ciphersuites:
4706 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4707 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4708 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4710 The latter two were purportedly from
4711 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4714 Also deactive the remaining ciphersuites from
4715 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4716 unofficial, and the ID has long expired.
4719 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4720 dual-core machines) and other potential thread-safety issues.
4723 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4725 *) Adapt fipsld and the build system to link against the validated FIPS
4726 module in FIPS mode.
4729 *) Fixes for VC++ 2005 build under Windows.
4732 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4733 from a Windows bash shell such as MSYS. It is autodetected from the
4734 "config" script when run from a VC++ environment. Modify standard VC++
4735 build to use fipscanister.o from the GNU make build.
4738 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4740 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4741 The value now differs depending on if you build for FIPS or not.
4742 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4743 safely run with a non-FIPSed libcrypto, as it may crash because of
4744 the difference induced by this change.
4747 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4749 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4750 (part of SSL_OP_ALL). This option used to disable the
4751 countermeasure against man-in-the-middle protocol-version
4752 rollback in the SSL 2.0 server implementation, which is a bad
4753 idea. (CVE-2005-2969)
4755 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4756 for Information Security, National Institute of Advanced Industrial
4757 Science and Technology [AIST], Japan)]
4759 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4760 mainly for FIPS compliance and not fully integrated at this stage.
4763 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4764 the exponentiation using a fixed-length exponent. (Otherwise,
4765 the information leaked through timing could expose the secret key
4766 after many signatures; cf. Bleichenbacher's attack on DSA with
4770 *) Make a new fixed-window mod_exp implementation the default for
4771 RSA, DSA, and DH private-key operations so that the sequence of
4772 squares and multiplies and the memory access pattern are
4773 independent of the particular secret key. This will mitigate
4774 cache-timing and potential related attacks.
4776 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4777 and this is automatically used by BN_mod_exp_mont() if the new flag
4778 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4779 will use this BN flag for private exponents unless the flag
4780 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4781 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4783 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4785 *) Change the client implementation for SSLv23_method() and
4786 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4787 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4788 (Previously, the SSL 2.0 backwards compatible Client Hello
4789 message format would be used even with SSL_OP_NO_SSLv2.)
4792 *) Add support for smime-type MIME parameter in S/MIME messages which some
4796 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4797 a threadsafe manner. Modify rsa code to use new function and add calls
4798 to dsa and dh code (which had race conditions before).
4801 *) Include the fixed error library code in the C error file definitions
4802 instead of fixing them up at runtime. This keeps the error code
4803 structures constant.
4806 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4808 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4811 *) Fixes for newer kerberos headers. NB: the casts are needed because
4812 the 'length' field is signed on one version and unsigned on another
4813 with no (?) obvious way to tell the difference, without these VC++
4814 complains. Also the "definition" of FAR (blank) is no longer included
4815 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4816 some needed definitions.
4819 *) Undo Cygwin change.
4822 *) Added support for proxy certificates according to RFC 3820.
4823 Because they may be a security thread to unaware applications,
4824 they must be explicitely allowed in run-time. See
4825 docs/HOWTO/proxy_certificates.txt for further information.
4828 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4830 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4831 server and client random values. Previously
4832 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4833 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4835 This change has negligible security impact because:
4837 1. Server and client random values still have 24 bytes of pseudo random
4840 2. Server and client random values are sent in the clear in the initial
4843 3. The master secret is derived using the premaster secret (48 bytes in
4844 size for static RSA ciphersuites) as well as client server and random
4847 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4850 [Stephen Henson, reported by UK NISCC]
4852 *) Use Windows randomness collection on Cygwin.
4855 *) Fix hang in EGD/PRNGD query when communication socket is closed
4856 prematurely by EGD/PRNGD.
4857 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4859 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4862 *) Back-port of selected performance improvements from development
4863 branch, as well as improved support for PowerPC platforms.
4866 *) Add lots of checks for memory allocation failure, error codes to indicate
4867 failure and freeing up memory if a failure occurs.
4868 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4870 *) Add new -passin argument to dgst.
4873 *) Perform some character comparisons of different types in X509_NAME_cmp:
4874 this is needed for some certificates that reencode DNs into UTF8Strings
4875 (in violation of RFC3280) and can't or wont issue name rollover
4879 *) Make an explicit check during certificate validation to see that
4880 the CA setting in each certificate on the chain is correct. As a
4881 side effect always do the following basic checks on extensions,
4882 not just when there's an associated purpose to the check:
4884 - if there is an unhandled critical extension (unless the user
4885 has chosen to ignore this fault)
4886 - if the path length has been exceeded (if one is set at all)
4887 - that certain extensions fit the associated purpose (if one has
4891 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4893 *) Avoid a race condition when CRLs are checked in a multi threaded
4894 environment. This would happen due to the reordering of the revoked
4895 entries during signature checking and serial number lookup. Now the
4896 encoding is cached and the serial number sort performed under a lock.
4897 Add new STACK function sk_is_sorted().
4900 *) Add Delta CRL to the extension code.
4903 *) Various fixes to s3_pkt.c so alerts are sent properly.
4904 [David Holmes <d.holmes@f5.com>]
4906 *) Reduce the chances of duplicate issuer name and serial numbers (in
4907 violation of RFC3280) using the OpenSSL certificate creation utilities.
4908 This is done by creating a random 64 bit value for the initial serial
4909 number when a serial number file is created or when a self signed
4910 certificate is created using 'openssl req -x509'. The initial serial
4911 number file is created using 'openssl x509 -next_serial' in CA.pl
4912 rather than being initialized to 1.
4915 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4917 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4918 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4919 [Joe Orton, Steve Henson]
4921 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4923 [Joe Orton, Steve Henson]
4925 *) Make it possible to have multiple active certificates with the same
4926 subject in the CA index file. This is done only if the keyword
4927 'unique_subject' is set to 'no' in the main CA section (default
4928 if 'CA_default') of the configuration file. The value is saved
4929 with the database itself in a separate index attribute file,
4930 named like the index file with '.attr' appended to the name.
4933 *) X509 verify fixes. Disable broken certificate workarounds when
4934 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4935 keyUsage extension present. Don't accept CRLs with unhandled critical
4936 extensions: since verify currently doesn't process CRL extensions this
4937 rejects a CRL with *any* critical extensions. Add new verify error codes
4941 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4942 A clarification of RFC2560 will require the use of OCTET STRINGs and
4943 some implementations cannot handle the current raw format. Since OpenSSL
4944 copies and compares OCSP nonces as opaque blobs without any attempt at
4945 parsing them this should not create any compatibility issues.
4948 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4949 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4950 this HMAC (and other) operations are several times slower than OpenSSL
4954 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4955 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4957 *) Use the correct content when signing type "other".
4960 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4962 *) Fix various bugs revealed by running the NISCC test suite:
4964 Stop out of bounds reads in the ASN1 code when presented with
4965 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4967 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4969 If verify callback ignores invalid public key errors don't try to check
4970 certificate signature with the NULL public key.
4974 *) New -ignore_err option in ocsp application to stop the server
4975 exiting on the first error in a request.
4978 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4979 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4983 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4984 extra data after the compression methods not only for TLS 1.0
4985 but also for SSL 3.0 (as required by the specification).
4986 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4988 *) Change X509_certificate_type() to mark the key as exported/exportable
4989 when it's 512 *bits* long, not 512 bytes.
4992 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4993 blocks during encryption.
4996 *) Various fixes to base64 BIO and non blocking I/O. On write
4997 flushes were not handled properly if the BIO retried. On read
4998 data was not being buffered properly and had various logic bugs.
4999 This also affects blocking I/O when the data being decoded is a
5003 *) Various S/MIME bugfixes and compatibility changes:
5004 output correct application/pkcs7 MIME type if
5005 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5006 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5007 of files as .eml work). Correctly handle very long lines in MIME
5011 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5013 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5014 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5015 a protocol version number mismatch like a decryption error
5016 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5019 *) Turn on RSA blinding by default in the default implementation
5020 to avoid a timing attack. Applications that don't want it can call
5021 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5022 They would be ill-advised to do so in most cases.
5023 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5025 *) Change RSA blinding code so that it works when the PRNG is not
5026 seeded (in this case, the secret RSA exponent is abused as
5027 an unpredictable seed -- if it is not unpredictable, there
5028 is no point in blinding anyway). Make RSA blinding thread-safe
5029 by remembering the creator's thread ID in rsa->blinding and
5030 having all other threads use local one-time blinding factors
5031 (this requires more computation than sharing rsa->blinding, but
5032 avoids excessive locking; and if an RSA object is not shared
5033 between threads, blinding will still be very fast).
5036 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5037 ENGINE as defaults for all supported algorithms irrespective of
5038 the 'flags' parameter. 'flags' is now honoured, so applications
5039 should make sure they are passing it correctly.
5042 *) Target "mingw" now allows native Windows code to be generated in
5043 the Cygwin environment as well as with the MinGW compiler.
5046 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5048 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5049 via timing by performing a MAC computation even if incorrrect
5050 block cipher padding has been found. This is a countermeasure
5051 against active attacks where the attacker has to distinguish
5052 between bad padding and a MAC verification error. (CVE-2003-0078)
5054 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5055 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5056 Martin Vuagnoux (EPFL, Ilion)]
5058 *) Make the no-err option work as intended. The intention with no-err
5059 is not to have the whole error stack handling routines removed from
5060 libcrypto, it's only intended to remove all the function name and
5061 reason texts, thereby removing some of the footprint that may not
5062 be interesting if those errors aren't displayed anyway.
5064 NOTE: it's still possible for any application or module to have it's
5065 own set of error texts inserted. The routines are there, just not
5066 used by default when no-err is given.
5069 *) Add support for FreeBSD on IA64.
5070 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5072 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5073 Kerberos function mit_des_cbc_cksum(). Before this change,
5074 the value returned by DES_cbc_cksum() was like the one from
5075 mit_des_cbc_cksum(), except the bytes were swapped.
5076 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5078 *) Allow an application to disable the automatic SSL chain building.
5079 Before this a rather primitive chain build was always performed in
5080 ssl3_output_cert_chain(): an application had no way to send the
5081 correct chain if the automatic operation produced an incorrect result.
5083 Now the chain builder is disabled if either:
5085 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5087 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5089 The reasoning behind this is that an application would not want the
5090 auto chain building to take place if extra chain certificates are
5091 present and it might also want a means of sending no additional
5092 certificates (for example the chain has two certificates and the
5096 *) Add the possibility to build without the ENGINE framework.
5097 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5099 *) Under Win32 gmtime() can return NULL: check return value in
5100 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5103 *) DSA routines: under certain error conditions uninitialized BN objects
5104 could be freed. Solution: make sure initialization is performed early
5105 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5106 Nils Larsch <nla@trustcenter.de> via PR#459)
5109 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5110 checked on reconnect on the client side, therefore session resumption
5111 could still fail with a "ssl session id is different" error. This
5112 behaviour is masked when SSL_OP_ALL is used due to
5113 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5114 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5115 followup to PR #377.
5118 *) IA-32 assembler support enhancements: unified ELF targets, support
5119 for SCO/Caldera platforms, fix for Cygwin shared build.
5122 *) Add support for FreeBSD on sparc64. As a consequence, support for
5123 FreeBSD on non-x86 processors is separate from x86 processors on
5124 the config script, much like the NetBSD support.
5125 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5127 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5129 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5132 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5133 code (06) was taken as the first octet of the session ID and the last
5134 octet was ignored consequently. As a result SSLv2 client side session
5135 caching could not have worked due to the session ID mismatch between
5137 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5141 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5142 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5146 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5147 seems that in spite of existing for more than a year, many application
5148 author have done nothing to provide the necessary callbacks, which
5149 means that this particular engine will not work properly anywhere.
5150 This is a very unfortunate situation which forces us, in the name
5151 of usability, to give the hw_ncipher.c a static lock, which is part
5153 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5154 appear in 0.9.8 or later. We EXPECT application authors to have
5155 dealt properly with this when 0.9.8 is released (unless we actually
5156 make such changes in the libcrypto locking code that changes will
5157 have to be made anyway).
5160 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5161 octets have been read, EOF or an error occurs. Without this change
5162 some truncated ASN1 structures will not produce an error.
5165 *) Disable Heimdal support, since it hasn't been fully implemented.
5166 Still give the possibility to force the use of Heimdal, but with
5167 warnings and a request that patches get sent to openssl-dev.
5170 *) Add the VC-CE target, introduce the WINCE sysname, and add
5171 INSTALL.WCE and appropriate conditionals to make it build.
5172 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5174 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5175 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5176 edit numbers of the version.
5177 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5179 *) Introduce safe string copy and catenation functions
5180 (BUF_strlcpy() and BUF_strlcat()).
5181 [Ben Laurie (CHATS) and Richard Levitte]
5183 *) Avoid using fixed-size buffers for one-line DNs.
5184 [Ben Laurie (CHATS)]
5186 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5187 resizing buffers containing secrets, and use where appropriate.
5188 [Ben Laurie (CHATS)]
5190 *) Avoid using fixed size buffers for configuration file location.
5191 [Ben Laurie (CHATS)]
5193 *) Avoid filename truncation for various CA files.
5194 [Ben Laurie (CHATS)]
5196 *) Use sizeof in preference to magic numbers.
5197 [Ben Laurie (CHATS)]
5199 *) Avoid filename truncation in cert requests.
5200 [Ben Laurie (CHATS)]
5202 *) Add assertions to check for (supposedly impossible) buffer
5204 [Ben Laurie (CHATS)]
5206 *) Don't cache truncated DNS entries in the local cache (this could
5207 potentially lead to a spoofing attack).
5208 [Ben Laurie (CHATS)]
5210 *) Fix various buffers to be large enough for hex/decimal
5211 representations in a platform independent manner.
5212 [Ben Laurie (CHATS)]
5214 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5215 resizing buffers containing secrets, and use where appropriate.
5216 [Ben Laurie (CHATS)]
5218 *) Add BIO_indent() to avoid much slightly worrying code to do
5220 [Ben Laurie (CHATS)]
5222 *) Convert sprintf()/BIO_puts() to BIO_printf().
5223 [Ben Laurie (CHATS)]
5225 *) buffer_gets() could terminate with the buffer only half
5227 [Ben Laurie (CHATS)]
5229 *) Add assertions to prevent user-supplied crypto functions from
5230 overflowing internal buffers by having large block sizes, etc.
5231 [Ben Laurie (CHATS)]
5233 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5235 [Ben Laurie (CHATS)]
5237 *) Eliminate unused copy of key in RC4.
5238 [Ben Laurie (CHATS)]
5240 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5241 [Ben Laurie (CHATS)]
5243 *) Fix off-by-one error in EGD path.
5244 [Ben Laurie (CHATS)]
5246 *) If RANDFILE path is too long, ignore instead of truncating.
5247 [Ben Laurie (CHATS)]
5249 *) Eliminate unused and incorrectly sized X.509 structure
5251 [Ben Laurie (CHATS)]
5253 *) Eliminate unused and dangerous function knumber().
5254 [Ben Laurie (CHATS)]
5256 *) Eliminate unused and dangerous structure, KSSL_ERR.
5257 [Ben Laurie (CHATS)]
5259 *) Protect against overlong session ID context length in an encoded
5260 session object. Since these are local, this does not appear to be
5262 [Ben Laurie (CHATS)]
5264 *) Change from security patch (see 0.9.6e below) that did not affect
5265 the 0.9.6 release series:
5267 Remote buffer overflow in SSL3 protocol - an attacker could
5268 supply an oversized master key in Kerberos-enabled versions.
5270 [Ben Laurie (CHATS)]
5272 *) Change the SSL kerb5 codes to match RFC 2712.
5275 *) Make -nameopt work fully for req and add -reqopt switch.
5276 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5278 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5279 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5281 *) Make sure tests can be performed even if the corresponding algorithms
5282 have been removed entirely. This was also the last step to make
5283 OpenSSL compilable with DJGPP under all reasonable conditions.
5284 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5286 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5287 to allow version independent disabling of normally unselected ciphers,
5288 which may be activated as a side-effect of selecting a single cipher.
5290 (E.g., cipher list string "RSA" enables ciphersuites that are left
5291 out of "ALL" because they do not provide symmetric encryption.
5292 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5293 [Lutz Jaenicke, Bodo Moeller]
5295 *) Add appropriate support for separate platform-dependent build
5296 directories. The recommended way to make a platform-dependent
5297 build directory is the following (tested on Linux), maybe with
5300 # Place yourself outside of the OpenSSL source tree. In
5301 # this example, the environment variable OPENSSL_SOURCE
5302 # is assumed to contain the absolute OpenSSL source directory.
5303 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5304 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5305 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5306 mkdir -p `dirname $F`
5307 ln -s $OPENSSL_SOURCE/$F $F
5310 To be absolutely sure not to disturb the source tree, a "make clean"
5311 is a good thing. If it isn't successfull, don't worry about it,
5312 it probably means the source directory is very clean.
5315 *) Make sure any ENGINE control commands make local copies of string
5316 pointers passed to them whenever necessary. Otherwise it is possible
5317 the caller may have overwritten (or deallocated) the original string
5318 data when a later ENGINE operation tries to use the stored values.
5319 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5321 *) Improve diagnostics in file reading and command-line digests.
5322 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5324 *) Add AES modes CFB and OFB to the object database. Correct an
5325 error in AES-CFB decryption.
5328 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5329 allows existing EVP_CIPHER_CTX structures to be reused after
5330 calling EVP_*Final(). This behaviour is used by encryption
5331 BIOs and some applications. This has the side effect that
5332 applications must explicitly clean up cipher contexts with
5333 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5336 *) Check the values of dna and dnb in bn_mul_recursive before calling
5337 bn_mul_comba (a non zero value means the a or b arrays do not contain
5338 n2 elements) and fallback to bn_mul_normal if either is not zero.
5341 *) Fix escaping of non-ASCII characters when using the -subj option
5342 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5345 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5346 form for "surname", serialNumber has no short form.
5347 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5348 therefore remove "mail" short name for "internet 7".
5349 The OID for unique identifiers in X509 certificates is
5350 x500UniqueIdentifier, not uniqueIdentifier.
5351 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5354 *) Add an "init" command to the ENGINE config module and auto initialize
5355 ENGINEs. Without any "init" command the ENGINE will be initialized
5356 after all ctrl commands have been executed on it. If init=1 the
5357 ENGINE is initailized at that point (ctrls before that point are run
5358 on the uninitialized ENGINE and after on the initialized one). If
5359 init=0 then the ENGINE will not be iniatialized at all.
5362 *) Fix the 'app_verify_callback' interface so that the user-defined
5363 argument is actually passed to the callback: In the
5364 SSL_CTX_set_cert_verify_callback() prototype, the callback
5365 declaration has been changed from
5368 int (*cb)(X509_STORE_CTX *,void *);
5369 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5370 i=s->ctx->app_verify_callback(&ctx)
5371 has been changed into
5372 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5374 To update applications using SSL_CTX_set_cert_verify_callback(),
5375 a dummy argument can be added to their callback functions.
5376 [D. K. Smetters <smetters@parc.xerox.com>]
5378 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5379 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5381 *) Add and OPENSSL_LOAD_CONF define which will cause
5382 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5383 This allows older applications to transparently support certain
5384 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5385 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5386 load the config file and OPENSSL_add_all_algorithms_conf() which will
5387 always load it have also been added.
5390 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5391 Adjust NIDs and EVP layer.
5392 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5394 *) Config modules support in openssl utility.
5396 Most commands now load modules from the config file,
5397 though in a few (such as version) this isn't done
5398 because it couldn't be used for anything.
5400 In the case of ca and req the config file used is
5401 the same as the utility itself: that is the -config
5402 command line option can be used to specify an
5406 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5407 use "openssl_conf" if filename is NULL use default openssl config file.
5410 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5411 config section name. Add a new flag to tolerate a missing config file
5412 and move code to CONF_modules_load_file().
5415 *) Support for crypto accelerator cards from Accelerated Encryption
5416 Processing, www.aep.ie. (Use engine 'aep')
5417 The support was copied from 0.9.6c [engine] and adapted/corrected
5418 to work with the new engine framework.
5419 [AEP Inc. and Richard Levitte]
5421 *) Support for SureWare crypto accelerator cards from Baltimore
5422 Technologies. (Use engine 'sureware')
5423 The support was copied from 0.9.6c [engine] and adapted
5424 to work with the new engine framework.
5427 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5428 make the newer ENGINE framework commands for the CHIL engine work.
5429 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5431 *) Make it possible to produce shared libraries on ReliantUNIX.
5432 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5434 *) Add the configuration target debug-linux-ppro.
5435 Make 'openssl rsa' use the general key loading routines
5436 implemented in apps.c, and make those routines able to
5437 handle the key format FORMAT_NETSCAPE and the variant
5439 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5441 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5442 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5444 *) Add -keyform to rsautl, and document -engine.
5445 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5447 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5448 BIO_R_NO_SUCH_FILE error code rather than the generic
5449 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5452 *) Add new functions
5454 ERR_peek_last_error_line
5455 ERR_peek_last_error_line_data.
5456 These are similar to
5459 ERR_peek_error_line_data,
5460 but report on the latest error recorded rather than the first one
5461 still in the error queue.
5462 [Ben Laurie, Bodo Moeller]
5464 *) default_algorithms option in ENGINE config module. This allows things
5466 default_algorithms = ALL
5467 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5470 *) Prelminary ENGINE config module.
5473 *) New experimental application configuration code.
5476 *) Change the AES code to follow the same name structure as all other
5477 symmetric ciphers, and behave the same way. Move everything to
5478 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5479 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5481 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5482 [Ben Laurie and Theo de Raadt]
5484 *) Add option to output public keys in req command.
5485 [Massimiliano Pala madwolf@openca.org]
5487 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5488 (up to about 10% better than before for P-192 and P-224).
5491 *) New functions/macros
5493 SSL_CTX_set_msg_callback(ctx, cb)
5494 SSL_CTX_set_msg_callback_arg(ctx, arg)
5495 SSL_set_msg_callback(ssl, cb)
5496 SSL_set_msg_callback_arg(ssl, arg)
5498 to request calling a callback function
5500 void cb(int write_p, int version, int content_type,
5501 const void *buf, size_t len, SSL *ssl, void *arg)
5503 whenever a protocol message has been completely received
5504 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5505 protocol version according to which the SSL library interprets
5506 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5507 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5508 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5509 specification (change_cipher_spec(20), alert(21), handshake(22)).
5510 'buf' and 'len' point to the actual message, 'ssl' to the
5511 SSL object, and 'arg' is the application-defined value set by
5512 SSL[_CTX]_set_msg_callback_arg().
5514 'openssl s_client' and 'openssl s_server' have new '-msg' options
5515 to enable a callback that displays all protocol messages.
5518 *) Change the shared library support so shared libraries are built as
5519 soon as the corresponding static library is finished, and thereby get
5520 openssl and the test programs linked against the shared library.
5521 This still only happens when the keyword "shard" has been given to
5522 the configuration scripts.
5524 NOTE: shared library support is still an experimental thing, and
5525 backward binary compatibility is still not guaranteed.
5526 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5528 *) Add support for Subject Information Access extension.
5529 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5531 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5532 additional bytes when new memory had to be allocated, not just
5533 when reusing an existing buffer.
5536 *) New command line and configuration option 'utf8' for the req command.
5537 This allows field values to be specified as UTF8 strings.
5540 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5541 runs for the former and machine-readable output for the latter.
5544 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5545 of the e-mail address in the DN (i.e., it will go into a certificate
5546 extension only). The new configuration file option 'email_in_dn = no'
5547 has the same effect.
5548 [Massimiliano Pala madwolf@openca.org]
5550 *) Change all functions with names starting with des_ to be starting
5551 with DES_ instead. Add wrappers that are compatible with libdes,
5552 but are named _ossl_old_des_*. Finally, add macros that map the
5553 des_* symbols to the corresponding _ossl_old_des_* if libdes
5554 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5555 desired, the des_* symbols will be mapped to DES_*, with one
5558 Since we provide two compatibility mappings, the user needs to
5559 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5560 compatibility is desired. The default (i.e., when that macro
5561 isn't defined) is OpenSSL 0.9.6c compatibility.
5563 There are also macros that enable and disable the support of old
5564 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5565 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5566 are defined, the default will apply: to support the old des routines.
5568 In either case, one must include openssl/des.h to get the correct
5569 definitions. Do not try to just include openssl/des_old.h, that
5572 NOTE: This is a major break of an old API into a new one. Software
5573 authors are encouraged to switch to the DES_ style functions. Some
5574 time in the future, des_old.h and the libdes compatibility functions
5575 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5576 default), and then completely removed.
5579 *) Test for certificates which contain unsupported critical extensions.
5580 If such a certificate is found during a verify operation it is
5581 rejected by default: this behaviour can be overridden by either
5582 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5583 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5584 X509_supported_extension() has also been added which returns 1 if a
5585 particular extension is supported.
5588 *) Modify the behaviour of EVP cipher functions in similar way to digests
5589 to retain compatibility with existing code.
5592 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5593 compatibility with existing code. In particular the 'ctx' parameter does
5594 not have to be to be initialized before the call to EVP_DigestInit() and
5595 it is tidied up after a call to EVP_DigestFinal(). New function
5596 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5597 EVP_MD_CTX_copy() changed to not require the destination to be
5598 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5599 requires the destination to be valid.
5601 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5602 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5605 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5606 so that complete 'Handshake' protocol structures are kept in memory
5607 instead of overwriting 'msg_type' and 'length' with 'body' data.
5610 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5611 [Massimo Santin via Richard Levitte]
5613 *) Major restructuring to the underlying ENGINE code. This includes
5614 reduction of linker bloat, separation of pure "ENGINE" manipulation
5615 (initialisation, etc) from functionality dealing with implementations
5616 of specific crypto iterfaces. This change also introduces integrated
5617 support for symmetric ciphers and digest implementations - so ENGINEs
5618 can now accelerate these by providing EVP_CIPHER and EVP_MD
5619 implementations of their own. This is detailed in crypto/engine/README
5620 as it couldn't be adequately described here. However, there are a few
5621 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5622 were changed in the original introduction of ENGINE code have now
5623 reverted back - the hooking from this code to ENGINE is now a good
5624 deal more passive and at run-time, operations deal directly with
5625 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5626 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5627 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5628 they were not being used by the framework as there is no concept of a
5629 BIGNUM_METHOD and they could not be generalised to the new
5630 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5631 ENGINE_cpy() has been removed as it cannot be consistently defined in
5635 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5638 *) Change mkdef.pl to sort symbols that get the same entry number,
5639 and make sure the automatically generated functions ERR_load_*
5640 become part of libeay.num as well.
5643 *) New function SSL_renegotiate_pending(). This returns true once
5644 renegotiation has been requested (either SSL_renegotiate() call
5645 or HelloRequest/ClientHello receveived from the peer) and becomes
5646 false once a handshake has been completed.
5647 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5648 sends a HelloRequest, but does not ensure that a handshake takes
5649 place. SSL_renegotiate_pending() is useful for checking if the
5650 client has followed the request.)
5653 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5654 By default, clients may request session resumption even during
5655 renegotiation (if session ID contexts permit); with this option,
5656 session resumption is possible only in the first handshake.
5658 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5659 more bits available for options that should not be part of
5660 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5663 *) Add some demos for certificate and certificate request creation.
5666 *) Make maximum certificate chain size accepted from the peer application
5667 settable (SSL*_get/set_max_cert_list()), as proposed by
5668 "Douglas E. Engert" <deengert@anl.gov>.
5671 *) Add support for shared libraries for Unixware-7
5672 (Boyd Lynn Gerber <gerberb@zenez.com>).
5675 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5676 be done prior to destruction. Use this to unload error strings from
5677 ENGINEs that load their own error strings. NB: This adds two new API
5678 functions to "get" and "set" this destroy handler in an ENGINE.
5681 *) Alter all existing ENGINE implementations (except "openssl" and
5682 "openbsd") to dynamically instantiate their own error strings. This
5683 makes them more flexible to be built both as statically-linked ENGINEs
5684 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5685 Also, add stub code to each that makes building them as self-contained
5686 shared-libraries easier (see README.ENGINE).
5689 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5690 implementations into applications that are completely implemented in
5691 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5692 commands that can be used to configure what shared-library to load and
5693 to control aspects of the way it is handled. Also, made an update to
5694 the README.ENGINE file that brings its information up-to-date and
5695 provides some information and instructions on the "dynamic" ENGINE
5696 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5699 *) Make it possible to unload ranges of ERR strings with a new
5700 "ERR_unload_strings" function.
5703 *) Add a copy() function to EVP_MD.
5706 *) Make EVP_MD routines take a context pointer instead of just the
5707 md_data void pointer.
5710 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5711 that the digest can only process a single chunk of data
5712 (typically because it is provided by a piece of
5713 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5714 is only going to provide a single chunk of data, and hence the
5715 framework needn't accumulate the data for oneshot drivers.
5718 *) As with "ERR", make it possible to replace the underlying "ex_data"
5719 functions. This change also alters the storage and management of global
5720 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5721 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5722 index counters. The API functions that use this state have been changed
5723 to take a "class_index" rather than pointers to the class's local STACK
5724 and counter, and there is now an API function to dynamically create new
5725 classes. This centralisation allows us to (a) plug a lot of the
5726 thread-safety problems that existed, and (b) makes it possible to clean
5727 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5728 such data would previously have always leaked in application code and
5729 workarounds were in place to make the memory debugging turn a blind eye
5730 to it. Application code that doesn't use this new function will still
5731 leak as before, but their memory debugging output will announce it now
5732 rather than letting it slide.
5734 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5735 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5736 has a return value to indicate success or failure.
5739 *) Make it possible to replace the underlying "ERR" functions such that the
5740 global state (2 LHASH tables and 2 locks) is only used by the "default"
5741 implementation. This change also adds two functions to "get" and "set"
5742 the implementation prior to it being automatically set the first time
5743 any other ERR function takes place. Ie. an application can call "get",
5744 pass the return value to a module it has just loaded, and that module
5745 can call its own "set" function using that value. This means the
5746 module's "ERR" operations will use (and modify) the error state in the
5747 application and not in its own statically linked copy of OpenSSL code.
5750 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5751 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5752 the operation, and provides a more encapsulated way for external code
5753 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5754 to use these functions rather than manually incrementing the counts.
5756 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5759 *) Add EVP test program.
5762 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5765 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5766 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5767 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5768 These allow a CRL to be built without having to access X509_CRL fields
5769 directly. Modify 'ca' application to use new functions.
5772 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5773 bug workarounds. Rollback attack detection is a security feature.
5774 The problem will only arise on OpenSSL servers when TLSv1 is not
5775 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5776 Software authors not wanting to support TLSv1 will have special reasons
5777 for their choice and can explicitly enable this option.
5778 [Bodo Moeller, Lutz Jaenicke]
5780 *) Rationalise EVP so it can be extended: don't include a union of
5781 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5782 (similar to those existing for EVP_CIPHER_CTX).
5787 EVP_MD_CTX_init(&md); /* new function call */
5788 EVP_DigestInit(&md, EVP_sha1());
5789 EVP_DigestUpdate(&md, in, len);
5790 EVP_DigestFinal(&md, out, NULL);
5791 EVP_MD_CTX_cleanup(&md); /* new function call */
5795 *) Make DES key schedule conform to the usual scheme, as well as
5796 correcting its structure. This means that calls to DES functions
5797 now have to pass a pointer to a des_key_schedule instead of a
5798 plain des_key_schedule (which was actually always a pointer
5801 des_key_schedule ks;
5803 des_set_key_checked(..., &ks);
5804 des_ncbc_encrypt(..., &ks, ...);
5806 (Note that a later change renames 'des_...' into 'DES_...'.)
5809 *) Initial reduction of linker bloat: the use of some functions, such as
5810 PEM causes large amounts of unused functions to be linked in due to
5811 poor organisation. For example pem_all.c contains every PEM function
5812 which has a knock on effect of linking in large amounts of (unused)
5813 ASN1 code. Grouping together similar functions and splitting unrelated
5814 functions prevents this.
5817 *) Cleanup of EVP macros.
5820 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5821 correct _ecb suffix.
5824 *) Add initial OCSP responder support to ocsp application. The
5825 revocation information is handled using the text based index
5826 use by the ca application. The responder can either handle
5827 requests generated internally, supplied in files (for example
5828 via a CGI script) or using an internal minimal server.
5831 *) Add configuration choices to get zlib compression for TLS.
5834 *) Changes to Kerberos SSL for RFC 2712 compliance:
5835 1. Implemented real KerberosWrapper, instead of just using
5836 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5837 2. Implemented optional authenticator field of KerberosWrapper.
5839 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5840 and authenticator structs; see crypto/krb5/.
5842 Generalized Kerberos calls to support multiple Kerberos libraries.
5843 [Vern Staats <staatsvr@asc.hpc.mil>,
5844 Jeffrey Altman <jaltman@columbia.edu>
5845 via Richard Levitte]
5847 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5848 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5849 values for each of the key sizes rather than having just
5850 parameters (and 'speed' generating keys each time).
5853 *) Speed up EVP routines.
5856 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5857 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5858 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5859 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5861 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5862 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5863 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5866 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5868 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5871 *) Added the OS2-EMX target.
5872 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5874 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5875 to support NCONF routines in extension code. New function CONF_set_nconf()
5876 to allow functions which take an NCONF to also handle the old LHASH
5877 structure: this means that the old CONF compatible routines can be
5878 retained (in particular wrt extensions) without having to duplicate the
5879 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5882 *) Enhance the general user interface with mechanisms for inner control
5883 and with possibilities to have yes/no kind of prompts.
5886 *) Change all calls to low level digest routines in the library and
5887 applications to use EVP. Add missing calls to HMAC_cleanup() and
5888 don't assume HMAC_CTX can be copied using memcpy().
5889 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5891 *) Add the possibility to control engines through control names but with
5892 arbitrary arguments instead of just a string.
5893 Change the key loaders to take a UI_METHOD instead of a callback
5894 function pointer. NOTE: this breaks binary compatibility with earlier
5895 versions of OpenSSL [engine].
5896 Adapt the nCipher code for these new conditions and add a card insertion
5900 *) Enhance the general user interface with mechanisms to better support
5901 dialog box interfaces, application-defined prompts, the possibility
5902 to use defaults (for example default passwords from somewhere else)
5903 and interrupts/cancellations.
5906 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5907 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5910 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5911 tidy up some unnecessarily weird code in 'sk_new()').
5912 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5914 *) Change the key loading routines for ENGINEs to use the same kind
5915 callback (pem_password_cb) as all other routines that need this
5919 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5920 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5921 than this minimum value is recommended.
5924 *) New random seeder for OpenVMS, using the system process statistics
5925 that are easily reachable.
5928 *) Windows apparently can't transparently handle global
5929 variables defined in DLLs. Initialisations such as:
5931 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5933 wont compile. This is used by the any applications that need to
5934 declare their own ASN1 modules. This was fixed by adding the option
5935 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5936 needed for static libraries under Win32.
5939 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5940 setting of purpose and trust fields. New X509_STORE trust and
5941 purpose functions and tidy up setting in other SSL functions.
5944 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5945 structure. These are inherited by X509_STORE_CTX when it is
5946 initialised. This allows various defaults to be set in the
5947 X509_STORE structure (such as flags for CRL checking and custom
5948 purpose or trust settings) for functions which only use X509_STORE_CTX
5949 internally such as S/MIME.
5951 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5952 trust settings if they are not set in X509_STORE. This allows X509_STORE
5953 purposes and trust (in S/MIME for example) to override any set by default.
5955 Add command line options for CRL checking to smime, s_client and s_server
5959 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5960 are set then the CRL is looked up in the X509_STORE structure and
5961 its validity and signature checked, then if the certificate is found
5962 in the CRL the verify fails with a revoked error.
5964 Various new CRL related callbacks added to X509_STORE_CTX structure.
5966 Command line options added to 'verify' application to support this.
5968 This needs some additional work, such as being able to handle multiple
5969 CRLs with different times, extension based lookup (rather than just
5970 by subject name) and ultimately more complete V2 CRL extension
5974 *) Add a general user interface API (crypto/ui/). This is designed
5975 to replace things like des_read_password and friends (backward
5976 compatibility functions using this new API are provided).
5977 The purpose is to remove prompting functions from the DES code
5978 section as well as provide for prompting through dialog boxes in
5979 a window system and the like.
5982 *) Add "ex_data" support to ENGINE so implementations can add state at a
5983 per-structure level rather than having to store it globally.
5986 *) Make it possible for ENGINE structures to be copied when retrieved by
5987 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5988 This causes the "original" ENGINE structure to act like a template,
5989 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5990 operational state can be localised to each ENGINE structure, despite the
5991 fact they all share the same "methods". New ENGINE structures returned in
5992 this case have no functional references and the return value is the single
5993 structural reference. This matches the single structural reference returned
5994 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5998 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5999 needs to match any other type at all we need to manually clear the
6003 *) Changes to the "openssl engine" utility to include;
6004 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6005 about an ENGINE's available control commands.
6006 - executing control commands from command line arguments using the
6007 '-pre' and '-post' switches. '-post' is only used if '-t' is
6008 specified and the ENGINE is successfully initialised. The syntax for
6009 the individual commands are colon-separated, for example;
6010 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6013 *) New dynamic control command support for ENGINEs. ENGINEs can now
6014 declare their own commands (numbers), names (strings), descriptions,
6015 and input types for run-time discovery by calling applications. A
6016 subset of these commands are implicitly classed as "executable"
6017 depending on their input type, and only these can be invoked through
6018 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6019 can be based on user input, config files, etc). The distinction is
6020 that "executable" commands cannot return anything other than a boolean
6021 result and can only support numeric or string input, whereas some
6022 discoverable commands may only be for direct use through
6023 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6024 pointers, or other custom uses. The "executable" commands are to
6025 support parameterisations of ENGINE behaviour that can be
6026 unambiguously defined by ENGINEs and used consistently across any
6027 OpenSSL-based application. Commands have been added to all the
6028 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6029 control over shared-library paths without source code alterations.
6032 *) Changed all ENGINE implementations to dynamically allocate their
6033 ENGINEs rather than declaring them statically. Apart from this being
6034 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6035 this also allows the implementations to compile without using the
6036 internal engine_int.h header.
6039 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6040 'const' value. Any code that should be able to modify a RAND_METHOD
6041 should already have non-const pointers to it (ie. they should only
6042 modify their own ones).
6045 *) Made a variety of little tweaks to the ENGINE code.
6046 - "atalla" and "ubsec" string definitions were moved from header files
6047 to C code. "nuron" string definitions were placed in variables
6048 rather than hard-coded - allowing parameterisation of these values
6049 later on via ctrl() commands.
6050 - Removed unused "#if 0"'d code.
6051 - Fixed engine list iteration code so it uses ENGINE_free() to release
6052 structural references.
6053 - Constified the RAND_METHOD element of ENGINE structures.
6054 - Constified various get/set functions as appropriate and added
6055 missing functions (including a catch-all ENGINE_cpy that duplicates
6056 all ENGINE values onto a new ENGINE except reference counts/state).
6057 - Removed NULL parameter checks in get/set functions. Setting a method
6058 or function to NULL is a way of cancelling out a previously set
6059 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6060 and doesn't justify the extra error symbols and code.
6061 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6062 flags from engine_int.h to engine.h.
6063 - Changed prototypes for ENGINE handler functions (init(), finish(),
6064 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6067 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6068 to the algorithm using long division. The binary algorithm can be
6069 used only if the modulus is odd. On 32-bit systems, it is faster
6070 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6071 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6072 up to 450 bits. In 64-bit environments, the binary algorithm
6073 appears to be advantageous for much longer moduli; here we use it
6074 for moduli up to 2048 bits.
6077 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6078 could not support the combine flag in choice fields.
6081 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6082 extensions from a certificate request to the certificate.
6085 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6086 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6087 file: this allows the display of the certificate about to be
6088 signed to be customised, to allow certain fields to be included
6089 or excluded and extension details. The old system didn't display
6090 multicharacter strings properly, omitted fields not in the policy
6091 and couldn't display additional details such as extensions.
6094 *) Function EC_POINTs_mul for multiple scalar multiplication
6095 of an arbitrary number of elliptic curve points
6096 \sum scalars[i]*points[i],
6097 optionally including the generator defined for the EC_GROUP:
6098 scalar*generator + \sum scalars[i]*points[i].
6100 EC_POINT_mul is a simple wrapper function for the typical case
6101 that the point list has just one item (besides the optional
6105 *) First EC_METHODs for curves over GF(p):
6107 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6108 operations and provides various method functions that can also
6109 operate with faster implementations of modular arithmetic.
6111 EC_GFp_mont_method() reuses most functions that are part of
6112 EC_GFp_simple_method, but uses Montgomery arithmetic.
6114 [Bodo Moeller; point addition and point doubling
6115 implementation directly derived from source code provided by
6116 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6118 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6119 crypto/ec/ec_lib.c):
6121 Curves are EC_GROUP objects (with an optional group generator)
6122 based on EC_METHODs that are built into the library.
6124 Points are EC_POINT objects based on EC_GROUP objects.
6126 Most of the framework would be able to handle curves over arbitrary
6127 finite fields, but as there are no obvious types for fields other
6128 than GF(p), some functions are limited to that for now.
6131 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6132 that the file contains a complete HTTP response.
6135 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6136 change the def and num file printf format specifier from "%-40sXXX"
6137 to "%-39s XXX". The latter will always guarantee a space after the
6138 field while the former will cause them to run together if the field
6139 is 40 of more characters long.
6142 *) Constify the cipher and digest 'method' functions and structures
6143 and modify related functions to take constant EVP_MD and EVP_CIPHER
6147 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6148 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6151 *) Modify EVP_Digest*() routines so they now return values. Although the
6152 internal software routines can never fail additional hardware versions
6156 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6158 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6159 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6162 ERR_R_NESTED_ASN1_ERROR
6164 ERR_R_MISSING_ASN1_EOS
6165 were 4 .. 9, conflicting with
6166 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6168 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6169 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6171 Add new error code 'ERR_R_INTERNAL_ERROR'.
6174 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6178 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6179 sets the subject name for a new request or supersedes the
6180 subject name in a given request. Formats that can be parsed are
6181 'CN=Some Name, OU=myOU, C=IT'
6183 'CN=Some Name/OU=myOU/C=IT'.
6185 Add options '-batch' and '-verbose' to 'openssl req'.
6186 [Massimiliano Pala <madwolf@hackmasters.net>]
6188 *) Introduce the possibility to access global variables through
6189 functions on platform were that's the best way to handle exporting
6190 global variables in shared libraries. To enable this functionality,
6191 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6192 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6193 is normally done by Configure or something similar).
6195 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6196 in the source file (foo.c) like this:
6198 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6199 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6201 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6202 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6204 OPENSSL_DECLARE_GLOBAL(int,foo);
6205 #define foo OPENSSL_GLOBAL_REF(foo)
6206 OPENSSL_DECLARE_GLOBAL(double,bar);
6207 #define bar OPENSSL_GLOBAL_REF(bar)
6209 The #defines are very important, and therefore so is including the
6210 header file everywhere where the defined globals are used.
6212 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6213 of ASN.1 items, but that structure is a bit different.
6215 The largest change is in util/mkdef.pl which has been enhanced with
6216 better and easier to understand logic to choose which symbols should
6217 go into the Windows .def files as well as a number of fixes and code
6218 cleanup (among others, algorithm keywords are now sorted
6219 lexicographically to avoid constant rewrites).
6222 *) In BN_div() keep a copy of the sign of 'num' before writing the
6223 result to 'rm' because if rm==num the value will be overwritten
6224 and produce the wrong result if 'num' is negative: this caused
6225 problems with BN_mod() and BN_nnmod().
6228 *) Function OCSP_request_verify(). This checks the signature on an
6229 OCSP request and verifies the signer certificate. The signer
6230 certificate is just checked for a generic purpose and OCSP request
6234 *) Add OCSP_check_validity() function to check the validity of OCSP
6235 responses. OCSP responses are prepared in real time and may only
6236 be a few seconds old. Simply checking that the current time lies
6237 between thisUpdate and nextUpdate max reject otherwise valid responses
6238 caused by either OCSP responder or client clock inaccuracy. Instead
6239 we allow thisUpdate and nextUpdate to fall within a certain period of
6240 the current time. The age of the response can also optionally be
6241 checked. Two new options -validity_period and -status_age added to
6245 *) If signature or public key algorithm is unrecognized print out its
6246 OID rather that just UNKNOWN.
6249 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6250 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6251 ID to be generated from the issuer certificate alone which can then be
6252 passed to OCSP_id_issuer_cmp().
6255 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6256 ASN1 modules to export functions returning ASN1_ITEM pointers
6257 instead of the ASN1_ITEM structures themselves. This adds several
6258 new macros which allow the underlying ASN1 function/structure to
6259 be accessed transparently. As a result code should not use ASN1_ITEM
6260 references directly (such as &X509_it) but instead use the relevant
6261 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6262 use of the new ASN1 code on platforms where exporting structures
6263 is problematical (for example in shared libraries) but exporting
6264 functions returning pointers to structures is not.
6267 *) Add support for overriding the generation of SSL/TLS session IDs.
6268 These callbacks can be registered either in an SSL_CTX or per SSL.
6269 The purpose of this is to allow applications to control, if they wish,
6270 the arbitrary values chosen for use as session IDs, particularly as it
6271 can be useful for session caching in multiple-server environments. A
6272 command-line switch for testing this (and any client code that wishes
6273 to use such a feature) has been added to "s_server".
6274 [Geoff Thorpe, Lutz Jaenicke]
6276 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6277 of the form '#if defined(...) || defined(...) || ...' and
6278 '#if !defined(...) && !defined(...) && ...'. This also avoids
6279 the growing number of special cases it was previously handling.
6282 *) Make all configuration macros available for application by making
6283 sure they are available in opensslconf.h, by giving them names starting
6284 with "OPENSSL_" to avoid conflicts with other packages and by making
6285 sure e_os2.h will cover all platform-specific cases together with
6287 Additionally, it is now possible to define configuration/platform-
6288 specific names (called "system identities"). In the C code, these
6289 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6290 macro with the name beginning with "OPENSSL_SYS_", which is determined
6291 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6295 *) New option -set_serial to 'req' and 'x509' this allows the serial
6296 number to use to be specified on the command line. Previously self
6297 signed certificates were hard coded with serial number 0 and the
6298 CA options of 'x509' had to use a serial number in a file which was
6302 *) New options to 'ca' utility to support V2 CRL entry extensions.
6303 Currently CRL reason, invalidity date and hold instruction are
6304 supported. Add new CRL extensions to V3 code and some new objects.
6307 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6308 disable standard block padding (aka PKCS#5 padding) in the EVP
6309 API, which was previously mandatory. This means that the data is
6310 not padded in any way and so the total length much be a multiple
6311 of the block size, otherwise an error occurs.
6314 *) Initial (incomplete) OCSP SSL support.
6317 *) New function OCSP_parse_url(). This splits up a URL into its host,
6318 port and path components: primarily to parse OCSP URLs. New -url
6319 option to ocsp utility.
6322 *) New nonce behavior. The return value of OCSP_check_nonce() now
6323 reflects the various checks performed. Applications can decide
6324 whether to tolerate certain situations such as an absent nonce
6325 in a response when one was present in a request: the ocsp application
6326 just prints out a warning. New function OCSP_add1_basic_nonce()
6327 this is to allow responders to include a nonce in a response even if
6328 the request is nonce-less.
6331 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6332 skipped when using openssl x509 multiple times on a single input file,
6333 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6336 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6337 set string type: to handle setting ASN1_TIME structures. Fix ca
6338 utility to correctly initialize revocation date of CRLs.
6341 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6342 the clients preferred ciphersuites and rather use its own preferences.
6343 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6344 Internet Explorer by ensuring unchanged hash method during stepup.
6345 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6348 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6349 to aes and add a new 'exist' option to print out symbols that don't
6353 *) Additional options to ocsp utility to allow flags to be set and
6354 additional certificates supplied.
6357 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6358 OCSP client a number of certificate to only verify the response
6362 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6363 handle the new API. Currently only ECB, CBC modes supported. Add new
6366 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6367 Encryption Standard (AES) Ciphersuites for Transport Layer
6368 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6369 not enabled by default and were not part of the "ALL" ciphersuite
6370 alias because they were not yet official; they could be
6371 explicitly requested by specifying the "AESdraft" ciphersuite
6372 group alias. In the final release of OpenSSL 0.9.7, the group
6373 alias is called "AES" and is part of "ALL".)
6374 [Ben Laurie, Steve Henson, Bodo Moeller]
6376 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6377 request to response.
6380 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6381 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6382 extract information from a certificate request. OCSP_response_create()
6383 creates a response and optionally adds a basic response structure.
6384 OCSP_basic_add1_status() adds a complete single response to a basic
6385 response and returns the OCSP_SINGLERESP structure just added (to allow
6386 extensions to be included for example). OCSP_basic_add1_cert() adds a
6387 certificate to a basic response and OCSP_basic_sign() signs a basic
6388 response with various flags. New helper functions ASN1_TIME_check()
6389 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6390 (converts ASN1_TIME to GeneralizedTime).
6393 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6394 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6395 structure from a certificate. X509_pubkey_digest() digests the public_key
6396 contents: this is used in various key identifiers.
6399 *) Make sk_sort() tolerate a NULL argument.
6400 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6402 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6403 passed by the function are trusted implicitly. If any of them signed the
6404 response then it is assumed to be valid and is not verified.
6407 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6408 to data. This was previously part of the PKCS7 ASN1 code. This
6409 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6410 [Steve Henson, reported by Kenneth R. Robinette
6411 <support@securenetterm.com>]
6413 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6414 routines: without these tracing memory leaks is very painful.
6415 Fix leaks in PKCS12 and PKCS7 routines.
6418 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6419 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6420 effectively meant GeneralizedTime would never be used. Now it
6421 is initialised to -1 but X509_time_adj() now has to check the value
6422 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6423 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6424 [Steve Henson, reported by Kenneth R. Robinette
6425 <support@securenetterm.com>]
6427 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6428 result in a zero length in the ASN1_INTEGER structure which was
6429 not consistent with the structure when d2i_ASN1_INTEGER() was used
6430 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6431 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6432 where it did not print out a minus for negative ASN1_INTEGER.
6435 *) Add summary printout to ocsp utility. The various functions which
6436 convert status values to strings have been renamed to:
6437 OCSP_response_status_str(), OCSP_cert_status_str() and
6438 OCSP_crl_reason_str() and are no longer static. New options
6439 to verify nonce values and to disable verification. OCSP response
6440 printout format cleaned up.
6443 *) Add additional OCSP certificate checks. These are those specified
6444 in RFC2560. This consists of two separate checks: the CA of the
6445 certificate being checked must either be the OCSP signer certificate
6446 or the issuer of the OCSP signer certificate. In the latter case the
6447 OCSP signer certificate must contain the OCSP signing extended key
6448 usage. This check is performed by attempting to match the OCSP
6449 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6450 in the OCSP_CERTID structures of the response.
6453 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6454 and related routines. This uses the standard OpenSSL certificate
6455 verify routines to perform initial checks (just CA validity) and
6456 to obtain the certificate chain. Then additional checks will be
6457 performed on the chain. Currently the root CA is checked to see
6458 if it is explicitly trusted for OCSP signing. This is used to set
6459 a root CA as a global signing root: that is any certificate that
6460 chains to that CA is an acceptable OCSP signing certificate.
6463 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6464 extensions from a separate configuration file.
6465 As when reading extensions from the main configuration file,
6466 the '-extensions ...' option may be used for specifying the
6468 [Massimiliano Pala <madwolf@comune.modena.it>]
6470 *) New OCSP utility. Allows OCSP requests to be generated or
6471 read. The request can be sent to a responder and the output
6472 parsed, outputed or printed in text form. Not complete yet:
6473 still needs to check the OCSP response validity.
6476 *) New subcommands for 'openssl ca':
6477 'openssl ca -status <serial>' prints the status of the cert with
6478 the given serial number (according to the index file).
6479 'openssl ca -updatedb' updates the expiry status of certificates
6481 [Massimiliano Pala <madwolf@comune.modena.it>]
6483 *) New '-newreq-nodes' command option to CA.pl. This is like
6484 '-newreq', but calls 'openssl req' with the '-nodes' option
6485 so that the resulting key is not encrypted.
6486 [Damien Miller <djm@mindrot.org>]
6488 *) New configuration for the GNU Hurd.
6489 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6491 *) Initial code to implement OCSP basic response verify. This
6492 is currently incomplete. Currently just finds the signer's
6493 certificate and verifies the signature on the response.
6496 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6497 value of OPENSSLDIR. This is available via the new '-d' option
6498 to 'openssl version', and is also included in 'openssl version -a'.
6501 *) Allowing defining memory allocation callbacks that will be given
6502 file name and line number information in additional arguments
6503 (a const char* and an int). The basic functionality remains, as
6504 well as the original possibility to just replace malloc(),
6505 realloc() and free() by functions that do not know about these
6506 additional arguments. To register and find out the current
6507 settings for extended allocation functions, the following
6508 functions are provided:
6510 CRYPTO_set_mem_ex_functions
6511 CRYPTO_set_locked_mem_ex_functions
6512 CRYPTO_get_mem_ex_functions
6513 CRYPTO_get_locked_mem_ex_functions
6515 These work the same way as CRYPTO_set_mem_functions and friends.
6516 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6517 extended allocation function is enabled.
6518 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6519 a conventional allocation function is enabled.
6520 [Richard Levitte, Bodo Moeller]
6522 *) Finish off removing the remaining LHASH function pointer casts.
6523 There should no longer be any prototype-casting required when using
6524 the LHASH abstraction, and any casts that remain are "bugs". See
6525 the callback types and macros at the head of lhash.h for details
6526 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6529 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6530 If /dev/[u]random devices are not available or do not return enough
6531 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6533 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6534 /etc/entropy will be queried once each in this sequence, quering stops
6535 when enough entropy was collected without querying more sockets.
6538 *) Change the Unix RAND_poll() variant to be able to poll several
6539 random devices, as specified by DEVRANDOM, until a sufficient amount
6540 of data has been collected. We spend at most 10 ms on each file
6541 (select timeout) and read in non-blocking mode. DEVRANDOM now
6542 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6543 (previously it was just the string "/dev/urandom"), so on typical
6544 platforms the 10 ms delay will never occur.
6545 Also separate out the Unix variant to its own file, rand_unix.c.
6546 For VMS, there's a currently-empty rand_vms.c.
6549 *) Move OCSP client related routines to ocsp_cl.c. These
6550 provide utility functions which an application needing
6551 to issue a request to an OCSP responder and analyse the
6552 response will typically need: as opposed to those which an
6553 OCSP responder itself would need which will be added later.
6555 OCSP_request_sign() signs an OCSP request with an API similar
6556 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6557 response. OCSP_response_get1_basic() extracts basic response
6558 from response. OCSP_resp_find_status(): finds and extracts status
6559 information from an OCSP_CERTID structure (which will be created
6560 when the request structure is built). These are built from lower
6561 level functions which work on OCSP_SINGLERESP structures but
6562 wont normally be used unless the application wishes to examine
6563 extensions in the OCSP response for example.
6565 Replace nonce routines with a pair of functions.
6566 OCSP_request_add1_nonce() adds a nonce value and optionally
6567 generates a random value. OCSP_check_nonce() checks the
6568 validity of the nonce in an OCSP response.
6571 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6572 This doesn't copy the supplied OCSP_CERTID and avoids the
6573 need to free up the newly created id. Change return type
6574 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6575 This can then be used to add extensions to the request.
6576 Deleted OCSP_request_new(), since most of its functionality
6577 is now in OCSP_REQUEST_new() (and the case insensitive name
6578 clash) apart from the ability to set the request name which
6579 will be added elsewhere.
6582 *) Update OCSP API. Remove obsolete extensions argument from
6583 various functions. Extensions are now handled using the new
6584 OCSP extension code. New simple OCSP HTTP function which
6585 can be used to send requests and parse the response.
6588 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6589 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6590 uses the special reorder version of SET OF to sort the attributes
6591 and reorder them to match the encoded order. This resolves a long
6592 standing problem: a verify on a PKCS7 structure just after signing
6593 it used to fail because the attribute order did not match the
6594 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6595 it uses the received order. This is necessary to tolerate some broken
6596 software that does not order SET OF. This is handled by encoding
6597 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6598 to produce the required SET OF.
6601 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6602 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6603 files to get correct declarations of the ASN.1 item variables.
6606 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6607 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6608 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6609 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6610 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6611 ASN1_ITEM and no wrapper functions.
6614 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6615 replace the old function pointer based I/O routines. Change most of
6616 the *_d2i_bio() and *_d2i_fp() functions to use these.
6619 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6620 lines, recognice more "algorithms" that can be deselected, and make
6621 it complain about algorithm deselection that isn't recognised.
6624 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6625 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6626 to use new functions. Add NO_ASN1_OLD which can be set to remove
6627 some old style ASN1 functions: this can be used to determine if old
6628 code will still work when these eventually go away.
6631 *) New extension functions for OCSP structures, these follow the
6632 same conventions as certificates and CRLs.
6635 *) New function X509V3_add1_i2d(). This automatically encodes and
6636 adds an extension. Its behaviour can be customised with various
6637 flags to append, replace or delete. Various wrappers added for
6638 certifcates and CRLs.
6641 *) Fix to avoid calling the underlying ASN1 print routine when
6642 an extension cannot be parsed. Correct a typo in the
6643 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6646 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
6647 entries for variables.
6650 *) Add functionality to apps/openssl.c for detecting locking
6651 problems: As the program is single-threaded, all we have
6652 to do is register a locking callback using an array for
6653 storing which locks are currently held by the program.
6656 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6657 SSL_get_ex_data_X509_STORE_idx(), which is used in
6658 ssl_verify_cert_chain() and thus can be called at any time
6659 during TLS/SSL handshakes so that thread-safety is essential.
6660 Unfortunately, the ex_data design is not at all suited
6661 for multi-threaded use, so it probably should be abolished.
6664 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6665 [Broadcom, tweaked and integrated by Geoff Thorpe]
6667 *) Move common extension printing code to new function
6668 X509V3_print_extensions(). Reorganise OCSP print routines and
6669 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6672 *) New function X509_signature_print() to remove duplication in some
6676 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6677 set (this was treated exactly the same as SET OF previously). This
6678 is used to reorder the STACK representing the structure to match the
6679 encoding. This will be used to get round a problem where a PKCS7
6680 structure which was signed could not be verified because the STACK
6681 order did not reflect the encoded order.
6684 *) Reimplement the OCSP ASN1 module using the new code.
6687 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6688 for its ASN1 operations. The old style function pointers still exist
6689 for now but they will eventually go away.
6692 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6693 completely replaces the old ASN1 functionality with a table driven
6694 encoder and decoder which interprets an ASN1_ITEM structure describing
6695 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6696 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6697 has also been converted to the new form.
6700 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6701 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6702 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6703 for negative moduli.
6706 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6707 of not touching the result's sign bit.
6710 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6714 *) Changed the LHASH code to use prototypes for callbacks, and created
6715 macros to declare and implement thin (optionally static) functions
6716 that provide type-safety and avoid function pointer casting for the
6717 type-specific callbacks.
6720 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6722 [Veers Staats <staatsvr@asc.hpc.mil>,
6723 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6725 *) Reformat the FAQ so the different questions and answers can be divided
6726 in sections depending on the subject.
6729 *) Have the zlib compression code load ZLIB.DLL dynamically under
6733 *) New function BN_mod_sqrt for computing square roots modulo a prime
6734 (using the probabilistic Tonelli-Shanks algorithm unless
6735 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6736 be handled deterministically).
6737 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6739 *) Make BN_mod_inverse faster by explicitly handling small quotients
6740 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6741 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6744 *) New function BN_kronecker.
6747 *) Fix BN_gcd so that it works on negative inputs; the result is
6748 positive unless both parameters are zero.
6749 Previously something reasonably close to an infinite loop was
6750 possible because numbers could be growing instead of shrinking
6751 in the implementation of Euclid's algorithm.
6754 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6755 sign of the number in question.
6757 Fix BN_is_word(a,w) to work correctly for w == 0.
6759 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6760 because its test if the absolute value of 'a' equals 'w'.
6761 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6762 it exists mostly for use in the implementations of BN_is_zero(),
6763 BN_is_one(), and BN_is_word().
6766 *) New function BN_swap.
6769 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6770 the exponentiation functions are more likely to produce reasonable
6771 results on negative inputs.
6774 *) Change BN_mod_mul so that the result is always non-negative.
6775 Previously, it could be negative if one of the factors was negative;
6776 I don't think anyone really wanted that behaviour.
6779 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6780 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6781 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6782 and add new functions:
6791 BN_mod_lshift1_quick
6795 These functions always generate non-negative results.
6797 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6798 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6800 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6801 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6802 be reduced modulo m.
6803 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6806 The following entry accidentily appeared in the CHANGES file
6807 distributed with OpenSSL 0.9.7. The modifications described in
6808 it do *not* apply to OpenSSL 0.9.7.
6810 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6811 was actually never needed) and in BN_mul(). The removal in BN_mul()
6812 required a small change in bn_mul_part_recursive() and the addition
6813 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6814 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6815 bn_sub_words() and bn_add_words() except they take arrays with
6820 *) In 'openssl passwd', verify passwords read from the terminal
6821 unless the '-salt' option is used (which usually means that
6822 verification would just waste user's time since the resulting
6823 hash is going to be compared with some given password hash)
6824 or the new '-noverify' option is used.
6826 This is an incompatible change, but it does not affect
6827 non-interactive use of 'openssl passwd' (passwords on the command
6828 line, '-stdin' option, '-in ...' option) and thus should not
6832 *) Remove all references to RSAref, since there's no more need for it.
6835 *) Make DSO load along a path given through an environment variable
6836 (SHLIB_PATH) with shl_load().
6839 *) Constify the ENGINE code as a result of BIGNUM constification.
6840 Also constify the RSA code and most things related to it. In a
6841 few places, most notable in the depth of the ASN.1 code, ugly
6842 casts back to non-const were required (to be solved at a later
6846 *) Make it so the openssl application has all engines loaded by default.
6849 *) Constify the BIGNUM routines a little more.
6852 *) Add the following functions:
6854 ENGINE_load_cswift()
6856 ENGINE_load_atalla()
6858 ENGINE_load_builtin_engines()
6860 That way, an application can itself choose if external engines that
6861 are built-in in OpenSSL shall ever be used or not. The benefit is
6862 that applications won't have to be linked with libdl or other dso
6863 libraries unless it's really needed.
6865 Changed 'openssl engine' to load all engines on demand.
6866 Changed the engine header files to avoid the duplication of some
6867 declarations (they differed!).
6870 *) 'openssl engine' can now list capabilities.
6873 *) Better error reporting in 'openssl engine'.
6876 *) Never call load_dh_param(NULL) in s_server.
6879 *) Add engine application. It can currently list engines by name and
6880 identity, and test if they are actually available.
6883 *) Improve RPM specification file by forcing symbolic linking and making
6884 sure the installed documentation is also owned by root.root.
6885 [Damien Miller <djm@mindrot.org>]
6887 *) Give the OpenSSL applications more possibilities to make use of
6888 keys (public as well as private) handled by engines.
6891 *) Add OCSP code that comes from CertCo.
6894 *) Add VMS support for the Rijndael code.
6897 *) Added untested support for Nuron crypto accelerator.
6900 *) Add support for external cryptographic devices. This code was
6901 previously distributed separately as the "engine" branch.
6902 [Geoff Thorpe, Richard Levitte]
6904 *) Rework the filename-translation in the DSO code. It is now possible to
6905 have far greater control over how a "name" is turned into a filename
6906 depending on the operating environment and any oddities about the
6907 different shared library filenames on each system.
6910 *) Support threads on FreeBSD-elf in Configure.
6913 *) Fix for SHA1 assembly problem with MASM: it produces
6914 warnings about corrupt line number information when assembling
6915 with debugging information. This is caused by the overlapping
6917 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6920 NCONF_get_number() has no error checking at all. As a replacement,
6921 NCONF_get_number_e() is defined (_e for "error checking") and is
6922 promoted strongly. The old NCONF_get_number is kept around for
6923 binary backward compatibility.
6924 Make it possible for methods to load from something other than a BIO,
6925 by providing a function pointer that is given a name instead of a BIO.
6926 For example, this could be used to load configuration data from an
6930 *) Fix for non blocking accept BIOs. Added new I/O special reason
6931 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6932 with non blocking I/O was not possible because no retry code was
6933 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6937 *) Added the beginnings of Rijndael support.
6940 *) Fix for bug in DirectoryString mask setting. Add support for
6941 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6942 to allow certificate printing to more controllable, additional
6943 'certopt' option to 'x509' to allow new printing options to be
6947 *) Clean old EAY MD5 hack from e_os.h.
6950 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6952 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6953 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6954 [Joe Orton, Steve Henson]
6956 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6958 *) Fix additional bug revealed by the NISCC test suite:
6960 Stop bug triggering large recursion when presented with
6961 certain ASN.1 tags (CVE-2003-0851)
6964 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6966 *) Fix various bugs revealed by running the NISCC test suite:
6968 Stop out of bounds reads in the ASN1 code when presented with
6969 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6971 If verify callback ignores invalid public key errors don't try to check
6972 certificate signature with the NULL public key.
6976 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6977 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6981 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6982 extra data after the compression methods not only for TLS 1.0
6983 but also for SSL 3.0 (as required by the specification).
6984 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6986 *) Change X509_certificate_type() to mark the key as exported/exportable
6987 when it's 512 *bits* long, not 512 bytes.
6990 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6992 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6993 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6994 a protocol version number mismatch like a decryption error
6995 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6998 *) Turn on RSA blinding by default in the default implementation
6999 to avoid a timing attack. Applications that don't want it can call
7000 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7001 They would be ill-advised to do so in most cases.
7002 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7004 *) Change RSA blinding code so that it works when the PRNG is not
7005 seeded (in this case, the secret RSA exponent is abused as
7006 an unpredictable seed -- if it is not unpredictable, there
7007 is no point in blinding anyway). Make RSA blinding thread-safe
7008 by remembering the creator's thread ID in rsa->blinding and
7009 having all other threads use local one-time blinding factors
7010 (this requires more computation than sharing rsa->blinding, but
7011 avoids excessive locking; and if an RSA object is not shared
7012 between threads, blinding will still be very fast).
7015 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7017 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7018 via timing by performing a MAC computation even if incorrrect
7019 block cipher padding has been found. This is a countermeasure
7020 against active attacks where the attacker has to distinguish
7021 between bad padding and a MAC verification error. (CVE-2003-0078)
7023 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7024 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7025 Martin Vuagnoux (EPFL, Ilion)]
7027 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7029 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7030 memory from it's contents. This is done with a counter that will
7031 place alternating values in each byte. This can be used to solve
7032 two issues: 1) the removal of calls to memset() by highly optimizing
7033 compilers, and 2) cleansing with other values than 0, since those can
7034 be read through on certain media, for example a swap space on disk.
7037 *) Bugfix: client side session caching did not work with external caching,
7038 because the session->cipher setting was not restored when reloading
7039 from the external cache. This problem was masked, when
7040 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7041 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7044 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7045 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7046 [Zeev Lieber <zeev-l@yahoo.com>]
7048 *) Undo an undocumented change introduced in 0.9.6e which caused
7049 repeated calls to OpenSSL_add_all_ciphers() and
7050 OpenSSL_add_all_digests() to be ignored, even after calling
7054 *) Change the default configuration reader to deal with last line not
7055 being properly terminated.
7058 *) Change X509_NAME_cmp() so it applies the special rules on handling
7059 DN values that are of type PrintableString, as well as RDNs of type
7060 emailAddress where the value has the type ia5String.
7061 [stefank@valicert.com via Richard Levitte]
7063 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7064 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7065 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7066 the bitwise-OR of the two for use by the majority of applications
7067 wanting this behaviour, and update the docs. The documented
7068 behaviour and actual behaviour were inconsistent and had been
7069 changing anyway, so this is more a bug-fix than a behavioural
7071 [Geoff Thorpe, diagnosed by Nadav Har'El]
7073 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7074 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7077 *) Fix initialization code race conditions in
7078 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7079 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7080 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7081 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7082 ssl2_get_cipher_by_char(),
7083 ssl3_get_cipher_by_char().
7084 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7086 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7087 the cached sessions are flushed, as the remove_cb() might use ex_data
7088 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7089 (see [openssl.org #212]).
7090 [Geoff Thorpe, Lutz Jaenicke]
7092 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7093 length, instead of the encoding length to d2i_ASN1_OBJECT.
7096 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7098 *) [In 0.9.6g-engine release:]
7099 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7100 [Lynn Gazis <lgazis@rainbow.com>]
7102 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7104 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7105 and get fix the header length calculation.
7106 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7107 Alon Kantor <alonk@checkpoint.com> (and others),
7110 *) Use proper error handling instead of 'assertions' in buffer
7111 overflow checks added in 0.9.6e. This prevents DoS (the
7112 assertions could call abort()).
7113 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7115 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7117 *) Add various sanity checks to asn1_get_length() to reject
7118 the ASN1 length bytes if they exceed sizeof(long), will appear
7119 negative or the content length exceeds the length of the
7121 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7123 *) Fix cipher selection routines: ciphers without encryption had no flags
7124 for the cipher strength set and where therefore not handled correctly
7125 by the selection routines (PR #130).
7128 *) Fix EVP_dsa_sha macro.
7132 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7133 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7134 that was added in OpenSSL 0.9.6d.
7136 As the countermeasure turned out to be incompatible with some
7137 broken SSL implementations, the new option is part of SSL_OP_ALL.
7138 SSL_OP_ALL is usually employed when compatibility with weird SSL
7139 implementations is desired (e.g. '-bugs' option to 's_client' and
7140 's_server'), so the new option is automatically set in many
7144 *) Changes in security patch:
7146 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7147 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7148 Air Force Materiel Command, USAF, under agreement number
7151 *) Add various sanity checks to asn1_get_length() to reject
7152 the ASN1 length bytes if they exceed sizeof(long), will appear
7153 negative or the content length exceeds the length of the
7154 supplied buffer. (CVE-2002-0659)
7155 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7157 *) Assertions for various potential buffer overflows, not known to
7159 [Ben Laurie (CHATS)]
7161 *) Various temporary buffers to hold ASCII versions of integers were
7162 too small for 64 bit platforms. (CVE-2002-0655)
7163 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7165 *) Remote buffer overflow in SSL3 protocol - an attacker could
7166 supply an oversized session ID to a client. (CVE-2002-0656)
7167 [Ben Laurie (CHATS)]
7169 *) Remote buffer overflow in SSL2 protocol - an attacker could
7170 supply an oversized client master key. (CVE-2002-0656)
7171 [Ben Laurie (CHATS)]
7173 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7175 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7176 encoded as NULL) with id-dsa-with-sha1.
7177 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7179 *) Check various X509_...() return values in apps/req.c.
7180 [Nils Larsch <nla@trustcenter.de>]
7182 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7183 an end-of-file condition would erronously be flagged, when the CRLF
7184 was just at the end of a processed block. The bug was discovered when
7185 processing data through a buffering memory BIO handing the data to a
7186 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7187 <ptsekov@syntrex.com> and Nedelcho Stanev.
7190 *) Implement a countermeasure against a vulnerability recently found
7191 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7192 before application data chunks to avoid the use of known IVs
7193 with data potentially chosen by the attacker.
7196 *) Fix length checks in ssl3_get_client_hello().
7199 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7200 to prevent ssl3_read_internal() from incorrectly assuming that
7201 ssl3_read_bytes() found application data while handshake
7202 processing was enabled when in fact s->s3->in_read_app_data was
7203 merely automatically cleared during the initial handshake.
7204 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7206 *) Fix object definitions for Private and Enterprise: they were not
7207 recognized in their shortname (=lowercase) representation. Extend
7208 obj_dat.pl to issue an error when using undefined keywords instead
7209 of silently ignoring the problem (Svenning Sorensen
7210 <sss@sss.dnsalias.net>).
7213 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7214 generators, i.e. generators other than 2 and 5. (Previously, the
7215 code did not properly initialise the 'add' and 'rem' values to
7216 BN_generate_prime().)
7218 In the new general case, we do not insist that 'generator' is
7219 actually a primitive root: This requirement is rather pointless;
7220 a generator of the order-q subgroup is just as good, if not
7224 *) Map new X509 verification errors to alerts. Discovered and submitted by
7225 Tom Wu <tom@arcot.com>.
7228 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7229 returning non-zero before the data has been completely received
7230 when using non-blocking I/O.
7231 [Bodo Moeller; problem pointed out by John Hughes]
7233 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7234 [Ben Laurie, Lutz Jaenicke]
7236 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7237 Yoram Zahavi <YoramZ@gilian.com>).
7240 *) Add information about CygWin 1.3 and on, and preserve proper
7241 configuration for the versions before that.
7242 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7244 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7245 check whether we deal with a copy of a session and do not delete from
7246 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7247 <izhar@checkpoint.com>.
7250 *) Do not store session data into the internal session cache, if it
7251 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7252 flag is set). Proposed by Aslam <aslam@funk.com>.
7255 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7259 *) [In 0.9.6d-engine release:]
7260 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7261 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7263 *) Add the configuration target linux-s390x.
7264 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7266 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7267 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7268 variable as an indication that a ClientHello message has been
7269 received. As the flag value will be lost between multiple
7270 invocations of ssl3_accept when using non-blocking I/O, the
7271 function may not be aware that a handshake has actually taken
7272 place, thus preventing a new session from being added to the
7275 To avoid this problem, we now set s->new_session to 2 instead of
7276 using a local variable.
7277 [Lutz Jaenicke, Bodo Moeller]
7279 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7280 if the SSL_R_LENGTH_MISMATCH error is detected.
7281 [Geoff Thorpe, Bodo Moeller]
7283 *) New 'shared_ldflag' column in Configure platform table.
7286 *) Fix EVP_CIPHER_mode macro.
7287 ["Dan S. Camper" <dan@bti.net>]
7289 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7290 type, we must throw them away by setting rr->length to 0.
7291 [D P Chang <dpc@qualys.com>]
7293 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7295 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7296 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7297 worked incorrectly for those cases where range = 10..._2 and
7298 3*range is two bits longer than range.)
7301 *) Only add signing time to PKCS7 structures if it is not already
7305 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7306 OBJ_ld_ce should be OBJ_id_ce.
7307 Also some ip-pda OIDs in crypto/objects/objects.txt were
7308 incorrect (cf. RFC 3039).
7309 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7311 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7312 returns early because it has nothing to do.
7313 [Andy Schneider <andy.schneider@bjss.co.uk>]
7315 *) [In 0.9.6c-engine release:]
7316 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7317 [Andy Schneider <andy.schneider@bjss.co.uk>]
7319 *) [In 0.9.6c-engine release:]
7320 Add support for Cryptographic Appliance's keyserver technology.
7321 (Use engine 'keyclient')
7322 [Cryptographic Appliances and Geoff Thorpe]
7324 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7325 is called via tools/c89.sh because arguments have to be
7326 rearranged (all '-L' options must appear before the first object
7328 [Richard Shapiro <rshapiro@abinitio.com>]
7330 *) [In 0.9.6c-engine release:]
7331 Add support for Broadcom crypto accelerator cards, backported
7333 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7335 *) [In 0.9.6c-engine release:]
7336 Add support for SureWare crypto accelerator cards from
7337 Baltimore Technologies. (Use engine 'sureware')
7338 [Baltimore Technologies and Mark Cox]
7340 *) [In 0.9.6c-engine release:]
7341 Add support for crypto accelerator cards from Accelerated
7342 Encryption Processing, www.aep.ie. (Use engine 'aep')
7343 [AEP Inc. and Mark Cox]
7345 *) Add a configuration entry for gcc on UnixWare.
7346 [Gary Benson <gbenson@redhat.com>]
7348 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7349 messages are stored in a single piece (fixed-length part and
7350 variable-length part combined) and fix various bugs found on the way.
7353 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7354 instead. BIO_gethostbyname() does not know what timeouts are
7355 appropriate, so entries would stay in cache even when they have
7357 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7359 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7360 faced with a pathologically small ClientHello fragment that does
7361 not contain client_version: Instead of aborting with an error,
7362 simply choose the highest available protocol version (i.e.,
7363 TLS 1.0 unless it is disabled). In practice, ClientHello
7364 messages are never sent like this, but this change gives us
7365 strictly correct behaviour at least for TLS.
7368 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7369 never resets s->method to s->ctx->method when called from within
7370 one of the SSL handshake functions.
7371 [Bodo Moeller; problem pointed out by Niko Baric]
7373 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7374 (sent using the client's version number) if client_version is
7375 smaller than the protocol version in use. Also change
7376 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7377 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7378 the client will at least see that alert.
7381 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7385 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7386 client receives HelloRequest while in a handshake.
7387 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7389 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7390 should end in 'break', not 'goto end' which circuments various
7391 cleanups done in state SSL_ST_OK. But session related stuff
7392 must be disabled for SSL_ST_OK in the case that we just sent a
7395 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7396 before just sending a HelloRequest.
7397 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7399 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7400 reveal whether illegal block cipher padding was found or a MAC
7401 verification error occured. (Neither SSLerr() codes nor alerts
7402 are directly visible to potential attackers, but the information
7403 may leak via logfiles.)
7405 Similar changes are not required for the SSL 2.0 implementation
7406 because the number of padding bytes is sent in clear for SSL 2.0,
7407 and the extra bytes are just ignored. However ssl/s2_pkt.c
7408 failed to verify that the purported number of padding bytes is in
7412 *) Add OpenUNIX-8 support including shared libraries
7413 (Boyd Lynn Gerber <gerberb@zenez.com>).
7416 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7417 'wristwatch attack' using huge encoding parameters (cf.
7418 James H. Manger's CRYPTO 2001 paper). Note that the
7419 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7420 encoding parameters and hence was not vulnerable.
7423 *) BN_sqr() bug fix.
7424 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7426 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7427 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7428 followed by modular reduction.
7429 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7431 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7432 equivalent based on BN_pseudo_rand() instead of BN_rand().
7435 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7436 This function was broken, as the check for a new client hello message
7437 to handle SGC did not allow these large messages.
7438 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7441 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7444 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7445 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7448 *) Rework the configuration and shared library support for Tru64 Unix.
7449 The configuration part makes use of modern compiler features and
7450 still retains old compiler behavior for those that run older versions
7451 of the OS. The shared library support part includes a variant that
7452 uses the RPATH feature, and is available through the special
7453 configuration target "alpha-cc-rpath", which will never be selected
7455 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7457 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7458 with the same message size as in ssl3_get_certificate_request().
7459 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7460 messages might inadvertently be reject as too long.
7461 [Petr Lampa <lampa@fee.vutbr.cz>]
7463 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7466 *) Modified SSL library such that the verify_callback that has been set
7467 specificly for an SSL object with SSL_set_verify() is actually being
7468 used. Before the change, a verify_callback set with this function was
7469 ignored and the verify_callback() set in the SSL_CTX at the time of
7470 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7471 to allow the necessary settings.
7474 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7475 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7476 done automatically (in contradiction to the requirements of the C
7477 standard). This made problems when used from OpenSSH.
7480 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7481 dh->length and always used
7483 BN_rand_range(priv_key, dh->p).
7485 BN_rand_range() is not necessary for Diffie-Hellman, and this
7486 specific range makes Diffie-Hellman unnecessarily inefficient if
7487 dh->length (recommended exponent length) is much smaller than the
7488 length of dh->p. We could use BN_rand_range() if the order of
7489 the subgroup was stored in the DH structure, but we only have
7494 BN_rand(priv_key, l, ...)
7496 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7502 RSA_eay_public_encrypt
7503 RSA_eay_private_decrypt
7504 RSA_eay_private_encrypt (signing)
7505 RSA_eay_public_decrypt (signature verification)
7507 (default implementations for RSA_public_encrypt,
7508 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7509 always reject numbers >= n.
7512 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7513 to synchronize access to 'locking_thread'. This is necessary on
7514 systems where access to 'locking_thread' (an 'unsigned long'
7515 variable) is not atomic.
7518 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7519 *before* setting the 'crypto_lock_rand' flag. The previous code had
7520 a race condition if 0 is a valid thread ID.
7521 [Travis Vitek <vitek@roguewave.com>]
7523 *) Add support for shared libraries under Irix.
7524 [Albert Chin-A-Young <china@thewrittenword.com>]
7526 *) Add configuration option to build on Linux on both big-endian and
7528 [Ralf Baechle <ralf@uni-koblenz.de>]
7530 *) Add the possibility to create shared libraries on HP-UX.
7533 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7535 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7536 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7537 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7538 PRNG state recovery was possible based on the output of
7539 one PRNG request appropriately sized to gain knowledge on
7540 'md' followed by enough consecutive 1-byte PRNG requests
7541 to traverse all of 'state'.
7543 1. When updating 'md_local' (the current thread's copy of 'md')
7544 during PRNG output generation, hash all of the previous
7545 'md_local' value, not just the half used for PRNG output.
7547 2. Make the number of bytes from 'state' included into the hash
7548 independent from the number of PRNG bytes requested.
7550 The first measure alone would be sufficient to avoid
7551 Markku-Juhani's attack. (Actually it had never occurred
7552 to me that the half of 'md_local' used for chaining was the
7553 half from which PRNG output bytes were taken -- I had always
7554 assumed that the secret half would be used.) The second
7555 measure makes sure that additional data from 'state' is never
7556 mixed into 'md_local' in small portions; this heuristically
7557 further strengthens the PRNG.
7560 *) Fix crypto/bn/asm/mips3.s.
7563 *) When only the key is given to "enc", the IV is undefined. Print out
7564 an error message in this case.
7567 *) Handle special case when X509_NAME is empty in X509 printing routines.
7570 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7571 positive and less than q.
7574 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7575 used: it isn't thread safe and the add_lock_callback should handle
7577 [Paul Rose <Paul.Rose@bridge.com>]
7579 *) Verify that incoming data obeys the block size in
7580 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7584 [Ulf Möller, Bodo Möller]
7586 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7587 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7588 when fixing the server behaviour for backwards-compatible 'client
7589 hello' messages. (Note that the attack is impractical against
7590 SSL 3.0 and TLS 1.0 anyway because length and version checking
7591 means that the probability of guessing a valid ciphertext is
7592 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7595 Before 0.9.5, the countermeasure (hide the error by generating a
7596 random 'decryption result') did not work properly because
7597 ERR_clear_error() was missing, meaning that SSL_get_error() would
7598 detect the supposedly ignored error.
7600 Both problems are now fixed.
7603 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7604 (previously it was 1024).
7607 *) Fix for compatibility mode trust settings: ignore trust settings
7608 unless some valid trust or reject settings are present.
7611 *) Fix for blowfish EVP: its a variable length cipher.
7614 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7615 parameters in DSA public key structures and return an error in the
7616 DSA routines if parameters are absent.
7619 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7620 in the current directory if neither $RANDFILE nor $HOME was set.
7621 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7622 caused some confusion to Windows users who haven't defined $HOME.
7623 Thus RAND_file_name() is changed again: e_os.h can define a
7624 DEFAULT_HOME, which will be used if $HOME is not set.
7625 For Windows, we use "C:"; on other platforms, we still require
7626 environment variables.
7628 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7629 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7630 having multiple threads call RAND_poll() concurrently.
7633 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7634 combination of a flag and a thread ID variable.
7635 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7636 flag), *other* threads can enter ssleay_add_bytes without obeying
7637 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7638 that they do not hold after the first thread unsets add_do_not_lock).
7641 *) Change bctest again: '-x' expressions are not available in all
7645 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7647 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7648 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7650 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7651 the default extension for executables, if any. Also, make the perl
7652 scripts that use symlink() to test if it really exists and use "cp"
7653 if it doesn't. All this made OpenSSL compilable and installable in
7657 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7658 If SEQUENCE is length is indefinite just set c->slen to the total
7659 amount of data available.
7660 [Steve Henson, reported by shige@FreeBSD.org]
7661 [This change does not apply to 0.9.7.]
7663 *) Change bctest to avoid here-documents inside command substitution
7664 (workaround for FreeBSD /bin/sh bug).
7665 For compatibility with Ultrix, avoid shell functions (introduced
7666 in the bctest version that searches along $PATH).
7669 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7670 with des_encrypt() defined on some operating systems, like Solaris
7674 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7675 On the Importance of Eliminating Errors in Cryptographic
7676 Computations, J. Cryptology 14 (2001) 2, 101-119,
7677 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7680 *) MIPS assembler BIGNUM division bug fix.
7683 *) Disabled incorrect Alpha assembler code.
7686 *) Fix PKCS#7 decode routines so they correctly update the length
7687 after reading an EOC for the EXPLICIT tag.
7689 [This change does not apply to 0.9.7.]
7691 *) Fix bug in PKCS#12 key generation routines. This was triggered
7692 if a 3DES key was generated with a 0 initial byte. Include
7693 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7694 (but broken) behaviour.
7697 *) Enhance bctest to search for a working bc along $PATH and print
7699 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7701 *) Fix memory leaks in err.c: free err_data string if necessary;
7702 don't write to the wrong index in ERR_set_error_data.
7705 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7709 *) Replace rdtsc with _emit statements for VC++ version 5.
7710 [Jeremy Cooper <jeremy@baymoo.org>]
7712 *) Make it possible to reuse SSLv2 sessions.
7715 *) In copy_email() check for >= 0 as a return value for
7716 X509_NAME_get_index_by_NID() since 0 is a valid index.
7717 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7719 *) Avoid coredump with unsupported or invalid public keys by checking if
7720 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7721 PKCS7_verify() fails with non detached data.
7724 *) Don't use getenv in library functions when run as setuid/setgid.
7725 New function OPENSSL_issetugid().
7728 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7729 due to incorrect handling of multi-threading:
7731 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7733 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7735 3. Count how many times MemCheck_off() has been called so that
7736 nested use can be treated correctly. This also avoids
7737 inband-signalling in the previous code (which relied on the
7738 assumption that thread ID 0 is impossible).
7741 *) Add "-rand" option also to s_client and s_server.
7744 *) Fix CPU detection on Irix 6.x.
7745 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7746 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7748 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7751 [This change does not apply to 0.9.7.]
7753 *) Use the cached encoding of an X509_NAME structure rather than
7754 copying it. This is apparently the reason for the libsafe "errors"
7755 but the code is actually correct.
7758 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7759 Bleichenbacher's DSA attack.
7760 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7761 to be set and top=0 forces the highest bit to be set; top=-1 is new
7762 and leaves the highest bit random.
7763 [Ulf Moeller, Bodo Moeller]
7765 *) In the NCONF_...-based implementations for CONF_... queries
7766 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7767 a temporary CONF structure with the data component set to NULL
7768 (which gives segmentation faults in lh_retrieve).
7769 Instead, use NULL for the CONF pointer in CONF_get_string and
7770 CONF_get_number (which may use environment variables) and directly
7771 return NULL from CONF_get_section.
7774 *) Fix potential buffer overrun for EBCDIC.
7777 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7778 keyUsage if basicConstraints absent for a CA.
7781 *) Make SMIME_write_PKCS7() write mail header values with a format that
7782 is more generally accepted (no spaces before the semicolon), since
7783 some programs can't parse those values properly otherwise. Also make
7784 sure BIO's that break lines after each write do not create invalid
7788 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7789 macros previously used would not encode an empty SEQUENCE OF
7790 and break the signature.
7792 [This change does not apply to 0.9.7.]
7794 *) Zero the premaster secret after deriving the master secret in
7798 *) Add some EVP_add_digest_alias registrations (as found in
7799 OpenSSL_add_all_digests()) to SSL_library_init()
7800 aka OpenSSL_add_ssl_algorithms(). This provides improved
7801 compatibility with peers using X.509 certificates
7802 with unconventional AlgorithmIdentifier OIDs.
7805 *) Fix for Irix with NO_ASM.
7806 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7808 *) ./config script fixes.
7809 [Ulf Moeller, Richard Levitte]
7811 *) Fix 'openssl passwd -1'.
7814 *) Change PKCS12_key_gen_asc() so it can cope with non null
7815 terminated strings whose length is passed in the passlen
7816 parameter, for example from PEM callbacks. This was done
7817 by adding an extra length parameter to asc2uni().
7818 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7820 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7821 call failed, free the DSA structure.
7824 *) Fix to uni2asc() to cope with zero length Unicode strings.
7825 These are present in some PKCS#12 files.
7828 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7829 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7830 when writing a 32767 byte record.
7831 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7833 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7834 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7836 (RSA objects have a reference count access to which is protected
7837 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7838 so they are meant to be shared between threads.)
7839 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7840 "Reddie, Steven" <Steven.Reddie@ca.com>]
7842 *) Fix a deadlock in CRYPTO_mem_leaks().
7845 *) Use better test patterns in bntest.
7848 *) rand_win.c fix for Borland C.
7851 *) BN_rshift bugfix for n == 0.
7854 *) Add a 'bctest' script that checks for some known 'bc' bugs
7855 so that 'make test' does not abort just because 'bc' is broken.
7858 *) Store verify_result within SSL_SESSION also for client side to
7859 avoid potential security hole. (Re-used sessions on the client side
7860 always resulted in verify_result==X509_V_OK, not using the original
7861 result of the server certificate verification.)
7864 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7865 SSL3_RT_APPLICATION_DATA, return 0.
7866 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7870 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7871 releases, have been re-implemented by renaming the previous
7872 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7873 and ssl3_read_internal, respectively, and adding 'peek' parameters
7874 to them. The new ssl[23]_{read,peek} functions are calls to
7875 ssl[23]_read_internal with the 'peek' flag set appropriately.
7876 A 'peek' parameter has also been added to ssl3_read_bytes, which
7877 does the actual work for ssl3_read_internal.
7880 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7881 the method-specific "init()" handler. Also clean up ex_data after
7882 calling the method-specific "finish()" handler. Previously, this was
7883 happening the other way round.
7886 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7887 The previous value, 12, was not always sufficient for BN_mod_exp().
7890 *) Make sure that shared libraries get the internal name engine with
7891 the full version number and not just 0. This should mark the
7892 shared libraries as not backward compatible. Of course, this should
7893 be changed again when we can guarantee backward binary compatibility.
7896 *) Fix typo in get_cert_by_subject() in by_dir.c
7897 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7899 *) Rework the system to generate shared libraries:
7901 - Make note of the expected extension for the shared libraries and
7902 if there is a need for symbolic links from for example libcrypto.so.0
7903 to libcrypto.so.0.9.7. There is extended info in Configure for
7906 - Make as few rebuilds of the shared libraries as possible.
7908 - Still avoid linking the OpenSSL programs with the shared libraries.
7910 - When installing, install the shared libraries separately from the
7914 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7916 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7917 and not in SSL_clear because the latter is also used by the
7918 accept/connect functions; previously, the settings made by
7919 SSL_set_read_ahead would be lost during the handshake.
7920 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7922 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7923 Previously, it would create entries for disableed algorithms no
7927 *) Added several new manual pages for SSL_* function.
7930 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7932 *) In ssl23_get_client_hello, generate an error message when faced
7933 with an initial SSL 3.0/TLS record that is too small to contain the
7934 first two bytes of the ClientHello message, i.e. client_version.
7935 (Note that this is a pathologic case that probably has never happened
7936 in real life.) The previous approach was to use the version number
7937 from the record header as a substitute; but our protocol choice
7938 should not depend on that one because it is not authenticated
7939 by the Finished messages.
7942 *) More robust randomness gathering functions for Windows.
7943 [Jeffrey Altman <jaltman@columbia.edu>]
7945 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7946 not set then we don't setup the error code for issuer check errors
7947 to avoid possibly overwriting other errors which the callback does
7948 handle. If an application does set the flag then we assume it knows
7949 what it is doing and can handle the new informational codes
7953 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7954 a general "ANY" type, as such it should be able to decode anything
7955 including tagged types. However it didn't check the class so it would
7956 wrongly interpret tagged types in the same way as their universal
7957 counterpart and unknown types were just rejected. Changed so that the
7958 tagged and unknown types are handled in the same way as a SEQUENCE:
7959 that is the encoding is stored intact. There is also a new type
7960 "V_ASN1_OTHER" which is used when the class is not universal, in this
7961 case we have no idea what the actual type is so we just lump them all
7965 *) On VMS, stdout may very well lead to a file that is written to
7966 in a record-oriented fashion. That means that every write() will
7967 write a separate record, which will be read separately by the
7968 programs trying to read from it. This can be very confusing.
7970 The solution is to put a BIO filter in the way that will buffer
7971 text until a linefeed is reached, and then write everything a
7972 line at a time, so every record written will be an actual line,
7973 not chunks of lines and not (usually doesn't happen, but I've
7974 seen it once) several lines in one record. BIO_f_linebuffer() is
7977 Currently, it's a VMS-only method, because that's where it has
7978 been tested well enough.
7981 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7982 it can return incorrect results.
7983 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7984 but it was in 0.9.6-beta[12].)
7987 *) Disable the check for content being present when verifying detached
7988 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7989 include zero length content when signing messages.
7992 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7993 BIO_ctrl (for BIO pairs).
7996 *) Add DSO method for VMS.
7999 *) Bug fix: Montgomery multiplication could produce results with the
8003 *) Add RPM specification openssl.spec and modify it to build three
8004 packages. The default package contains applications, application
8005 documentation and run-time libraries. The devel package contains
8006 include files, static libraries and function documentation. The
8007 doc package contains the contents of the doc directory. The original
8008 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8011 *) Add a large number of documentation files for many SSL routines.
8012 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8014 *) Add a configuration entry for Sony News 4.
8015 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8017 *) Don't set the two most significant bits to one when generating a
8018 random number < q in the DSA library.
8021 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8022 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8023 the underlying transport is blocking) if a handshake took place.
8024 (The default behaviour is needed by applications such as s_client
8025 and s_server that use select() to determine when to use SSL_read;
8026 but for applications that know in advance when to expect data, it
8027 just makes things more complicated.)
8030 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8034 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8035 work better on such systems.
8036 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8038 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8039 Update PKCS12_parse() so it copies the friendlyName and the
8040 keyid to the certificates aux info.
8043 *) Fix bug in PKCS7_verify() which caused an infinite loop
8044 if there was more than one signature.
8045 [Sven Uszpelkat <su@celocom.de>]
8047 *) Major change in util/mkdef.pl to include extra information
8048 about each symbol, as well as presentig variables as well
8049 as functions. This change means that there's n more need
8050 to rebuild the .num files when some algorithms are excluded.
8053 *) Allow the verify time to be set by an application,
8054 rather than always using the current time.
8057 *) Phase 2 verify code reorganisation. The certificate
8058 verify code now looks up an issuer certificate by a
8059 number of criteria: subject name, authority key id
8060 and key usage. It also verifies self signed certificates
8061 by the same criteria. The main comparison function is
8062 X509_check_issued() which performs these checks.
8064 Lot of changes were necessary in order to support this
8065 without completely rewriting the lookup code.
8067 Authority and subject key identifier are now cached.
8069 The LHASH 'certs' is X509_STORE has now been replaced
8070 by a STACK_OF(X509_OBJECT). This is mainly because an
8071 LHASH can't store or retrieve multiple objects with
8072 the same hash value.
8074 As a result various functions (which were all internal
8075 use only) have changed to handle the new X509_STORE
8076 structure. This will break anything that messed round
8077 with X509_STORE internally.
8079 The functions X509_STORE_add_cert() now checks for an
8080 exact match, rather than just subject name.
8082 The X509_STORE API doesn't directly support the retrieval
8083 of multiple certificates matching a given criteria, however
8084 this can be worked round by performing a lookup first
8085 (which will fill the cache with candidate certificates)
8086 and then examining the cache for matches. This is probably
8087 the best we can do without throwing out X509_LOOKUP
8088 entirely (maybe later...).
8090 The X509_VERIFY_CTX structure has been enhanced considerably.
8092 All certificate lookup operations now go via a get_issuer()
8093 callback. Although this currently uses an X509_STORE it
8094 can be replaced by custom lookups. This is a simple way
8095 to bypass the X509_STORE hackery necessary to make this
8096 work and makes it possible to use more efficient techniques
8097 in future. A very simple version which uses a simple
8098 STACK for its trusted certificate store is also provided
8099 using X509_STORE_CTX_trusted_stack().
8101 The verify_cb() and verify() callbacks now have equivalents
8102 in the X509_STORE_CTX structure.
8104 X509_STORE_CTX also has a 'flags' field which can be used
8105 to customise the verify behaviour.
8108 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8109 excludes S/MIME capabilities.
8112 *) When a certificate request is read in keep a copy of the
8113 original encoding of the signed data and use it when outputing
8114 again. Signatures then use the original encoding rather than
8115 a decoded, encoded version which may cause problems if the
8116 request is improperly encoded.
8119 *) For consistency with other BIO_puts implementations, call
8120 buffer_write(b, ...) directly in buffer_puts instead of calling
8123 In BIO_puts, increment b->num_write as in BIO_write.
8124 [Peter.Sylvester@EdelWeb.fr]
8126 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8127 BN_zero, we may not return a BIGNUM with an array consisting of
8131 *) Avoid calling abort() from within the library when problems are
8132 detected, except if preprocessor symbols have been defined
8133 (such as REF_CHECK, BN_DEBUG etc.).
8136 *) New openssl application 'rsautl'. This utility can be
8137 used for low level RSA operations. DER public key
8138 BIO/fp routines also added.
8141 *) New Configure entry and patches for compiling on QNX 4.
8142 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8144 *) A demo state-machine implementation was sponsored by
8145 Nuron (http://www.nuron.com/) and is now available in
8146 demos/state_machine.
8149 *) New options added to the 'dgst' utility for signature
8150 generation and verification.
8153 *) Unrecognized PKCS#7 content types are now handled via a
8154 catch all ASN1_TYPE structure. This allows unsupported
8155 types to be stored as a "blob" and an application can
8156 encode and decode it manually.
8159 *) Fix various signed/unsigned issues to make a_strex.c
8161 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8163 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8164 length if passed a buffer. ASN1_INTEGER_to_BN failed
8165 if passed a NULL BN and its argument was negative.
8166 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8168 *) Modification to PKCS#7 encoding routines to output definite
8169 length encoding. Since currently the whole structures are in
8170 memory there's not real point in using indefinite length
8171 constructed encoding. However if OpenSSL is compiled with
8172 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8175 *) Added BIO_vprintf() and BIO_vsnprintf().
8178 *) Added more prefixes to parse for in the the strings written
8179 through a logging bio, to cover all the levels that are available
8180 through syslog. The prefixes are now:
8182 PANIC, EMERG, EMR => LOG_EMERG
8183 ALERT, ALR => LOG_ALERT
8184 CRIT, CRI => LOG_CRIT
8185 ERROR, ERR => LOG_ERR
8186 WARNING, WARN, WAR => LOG_WARNING
8187 NOTICE, NOTE, NOT => LOG_NOTICE
8188 INFO, INF => LOG_INFO
8189 DEBUG, DBG => LOG_DEBUG
8191 and as before, if none of those prefixes are present at the
8192 beginning of the string, LOG_ERR is chosen.
8194 On Win32, the LOG_* levels are mapped according to this:
8196 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8197 LOG_WARNING => EVENTLOG_WARNING_TYPE
8198 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8202 *) Made it possible to reconfigure with just the configuration
8203 argument "reconf" or "reconfigure". The command line arguments
8204 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8205 and are retrieved from there when reconfiguring.
8209 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8211 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8214 *) The obj_dat.pl script was messing up the sorting of object
8215 names. The reason was that it compared the quoted version
8216 of strings as a result "OCSP" > "OCSP Signing" because
8217 " > SPACE. Changed script to store unquoted versions of
8218 names and add quotes on output. It was also omitting some
8219 names from the lookup table if they were given a default
8220 value (that is if SN is missing it is given the same
8221 value as LN and vice versa), these are now added on the
8222 grounds that if an object has a name we should be able to
8223 look it up. Finally added warning output when duplicate
8224 short or long names are found.
8227 *) Changes needed for Tandem NSK.
8228 [Scott Uroff <scott@xypro.com>]
8230 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8231 RSA_padding_check_SSLv23(), special padding was never detected
8232 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8233 version rollback attacks was not effective.
8235 In s23_clnt.c, don't use special rollback-attack detection padding
8236 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8237 client; similarly, in s23_srvr.c, don't do the rollback check if
8238 SSL 2.0 is the only protocol enabled in the server.
8241 *) Make it possible to get hexdumps of unprintable data with 'openssl
8242 asn1parse'. By implication, the functions ASN1_parse_dump() and
8243 BIO_dump_indent() are added.
8246 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8247 these print out strings and name structures based on various
8248 flags including RFC2253 support and proper handling of
8249 multibyte characters. Added options to the 'x509' utility
8250 to allow the various flags to be set.
8253 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8254 Also change the functions X509_cmp_current_time() and
8255 X509_gmtime_adj() work with an ASN1_TIME structure,
8256 this will enable certificates using GeneralizedTime in validity
8257 dates to be checked.
8260 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8261 negative public key encodings) on by default,
8262 NO_NEG_PUBKEY_BUG can be set to disable it.
8265 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8266 content octets. An i2c_ASN1_OBJECT is unnecessary because
8267 the encoding can be trivially obtained from the structure.
8270 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8271 not read locks (CRYPTO_r_[un]lock).
8274 *) A first attempt at creating official support for shared
8275 libraries through configuration. I've kept it so the
8276 default is static libraries only, and the OpenSSL programs
8277 are always statically linked for now, but there are
8278 preparations for dynamic linking in place.
8279 This has been tested on Linux and Tru64.
8282 *) Randomness polling function for Win9x, as described in:
8283 Peter Gutmann, Software Generation of Practically Strong
8287 *) Fix so PRNG is seeded in req if using an already existing
8291 *) New options to smime application. -inform and -outform
8292 allow alternative formats for the S/MIME message including
8293 PEM and DER. The -content option allows the content to be
8294 specified separately. This should allow things like Netscape
8295 form signing output easier to verify.
8298 *) Fix the ASN1 encoding of tags using the 'long form'.
8301 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8302 STRING types. These convert content octets to and from the
8303 underlying type. The actual tag and length octets are
8304 already assumed to have been read in and checked. These
8305 are needed because all other string types have virtually
8306 identical handling apart from the tag. By having versions
8307 of the ASN1 functions that just operate on content octets
8308 IMPLICIT tagging can be handled properly. It also allows
8309 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8310 and ASN1_INTEGER are identical apart from the tag.
8313 *) Change the handling of OID objects as follows:
8315 - New object identifiers are inserted in objects.txt, following
8316 the syntax given in objects.README.
8317 - objects.pl is used to process obj_mac.num and create a new
8319 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8322 This is currently kind of a hack, and the perl code in objects.pl
8323 isn't very elegant, but it works as I intended. The simplest way
8324 to check that it worked correctly is to look in obj_dat.h and
8325 check the array nid_objs and make sure the objects haven't moved
8326 around (this is important!). Additions are OK, as well as
8327 consistent name changes.
8330 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8333 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8334 The given file adds to whatever has already been seeded into the
8335 random pool through the RANDFILE configuration file option or
8336 environment variable, or the default random state file.
8339 *) mkstack.pl now sorts each macro group into lexical order.
8340 Previously the output order depended on the order the files
8341 appeared in the directory, resulting in needless rewriting
8345 *) Patches to make OpenSSL compile under Win32 again. Mostly
8346 work arounds for the VC++ problem that it treats func() as
8347 func(void). Also stripped out the parts of mkdef.pl that
8348 added extra typesafe functions: these no longer exist.
8351 *) Reorganisation of the stack code. The macros are now all
8352 collected in safestack.h . Each macro is defined in terms of
8353 a "stack macro" of the form SKM_<name>(type, a, b). The
8354 DEBUG_SAFESTACK is now handled in terms of function casts,
8355 this has the advantage of retaining type safety without the
8356 use of additional functions. If DEBUG_SAFESTACK is not defined
8357 then the non typesafe macros are used instead. Also modified the
8358 mkstack.pl script to handle the new form. Needs testing to see
8359 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8360 the default if no major problems. Similar behaviour for ASN1_SET_OF
8361 and PKCS12_STACK_OF.
8364 *) When some versions of IIS use the 'NET' form of private key the
8365 key derivation algorithm is different. Normally MD5(password) is
8366 used as a 128 bit RC4 key. In the modified case
8367 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
8368 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8369 as the old Netscape_RSA functions except they have an additional
8370 'sgckey' parameter which uses the modified algorithm. Also added
8371 an -sgckey command line option to the rsa utility. Thanks to
8372 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8373 algorithm to openssl-dev.
8376 *) The evp_local.h macros were using 'c.##kname' which resulted in
8377 invalid expansion on some systems (SCO 5.0.5 for example).
8378 Corrected to 'c.kname'.
8379 [Phillip Porch <root@theporch.com>]
8381 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8382 a STACK of email addresses from a certificate or request, these look
8383 in the subject name and the subject alternative name extensions and
8384 omit any duplicate addresses.
8387 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8388 This makes DSA verification about 2 % faster.
8391 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8392 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8393 plus overhead for 1024 bit moduli).
8394 This makes exponentiations about 0.5 % faster for 1024 bit
8395 exponents (as measured by "openssl speed rsa2048").
8398 *) Rename memory handling macros to avoid conflicts with other
8400 Malloc => OPENSSL_malloc
8401 Malloc_locked => OPENSSL_malloc_locked
8402 Realloc => OPENSSL_realloc
8403 Free => OPENSSL_free
8406 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8407 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8410 *) CygWin32 support.
8411 [John Jarvie <jjarvie@newsguy.com>]
8413 *) The type-safe stack code has been rejigged. It is now only compiled
8414 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8415 by default all type-specific stack functions are "#define"d back to
8416 standard stack functions. This results in more streamlined output
8417 but retains the type-safety checking possibilities of the original
8421 *) The STACK code has been cleaned up, and certain type declarations
8422 that didn't make a lot of sense have been brought in line. This has
8423 also involved a cleanup of sorts in safestack.h to more correctly
8424 map type-safe stack functions onto their plain stack counterparts.
8425 This work has also resulted in a variety of "const"ifications of
8426 lots of the code, especially "_cmp" operations which should normally
8427 be prototyped with "const" parameters anyway.
8430 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8431 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8432 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8433 where all of 'md' is used each time the PRNG is used, but 'state'
8434 is used only indexed by a cyclic counter. As entropy may not be
8435 well distributed from the beginning, 'md' is important as a
8436 chaining variable. However, the output function chains only half
8437 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8438 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8439 in all of 'state' being rewritten, with the new values depending
8440 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8443 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8444 the handshake is continued after ssl_verify_cert_chain();
8445 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8446 can lead to 'unexplainable' connection aborts later.
8447 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8449 *) Major EVP API cipher revision.
8450 Add hooks for extra EVP features. This allows various cipher
8451 parameters to be set in the EVP interface. Support added for variable
8452 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8453 setting of RC2 and RC5 parameters.
8455 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8458 Remove lots of duplicated code from the EVP library. For example *every*
8459 cipher init() function handles the 'iv' in the same way according to the
8460 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8461 for CFB and OFB modes they zero ctx->num.
8463 New functionality allows removal of S/MIME code RC2 hack.
8465 Most of the routines have the same form and so can be declared in terms
8468 By shifting this to the top level EVP_CipherInit() it can be removed from
8469 all individual ciphers. If the cipher wants to handle IVs or keys
8470 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8473 Change lots of functions like EVP_EncryptUpdate() to now return a
8474 value: although software versions of the algorithms cannot fail
8475 any installed hardware versions can.
8478 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8479 this option is set, tolerate broken clients that send the negotiated
8480 protocol version number instead of the requested protocol version
8484 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8485 i.e. non-zero for export ciphersuites, zero otherwise.
8486 Previous versions had this flag inverted, inconsistent with
8487 rsa_tmp_cb (..._TMP_RSA_CB).
8488 [Bodo Moeller; problem reported by Amit Chopra]
8490 *) Add missing DSA library text string. Work around for some IIS
8491 key files with invalid SEQUENCE encoding.
8494 *) Add a document (doc/standards.txt) that list all kinds of standards
8495 and so on that are implemented in OpenSSL.
8498 *) Enhance c_rehash script. Old version would mishandle certificates
8499 with the same subject name hash and wouldn't handle CRLs at all.
8500 Added -fingerprint option to crl utility, to support new c_rehash
8504 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8507 *) Fix for SSL server purpose checking. Server checking was
8508 rejecting certificates which had extended key usage present
8509 but no ssl client purpose.
8510 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8512 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8513 is a little unclear about how a blank password is handled.
8514 Since the password in encoded as a BMPString with terminating
8515 double NULL a zero length password would end up as just the
8516 double NULL. However no password at all is different and is
8517 handled differently in the PKCS#12 key generation code. NS
8518 treats a blank password as zero length. MSIE treats it as no
8519 password on export: but it will try both on import. We now do
8520 the same: PKCS12_parse() tries zero length and no password if
8521 the password is set to "" or NULL (NULL is now a valid password:
8522 it wasn't before) as does the pkcs12 application.
8525 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8526 perror when PEM_read_bio_X509_REQ fails, the error message must
8527 be obtained from the error queue.
8530 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8531 it in ERR_remove_state if appropriate, and change ERR_get_state
8532 accordingly to avoid race conditions (this is necessary because
8533 thread_hash is no longer constant once set).
8536 *) Bugfix for linux-elf makefile.one.
8539 *) RSA_get_default_method() will now cause a default
8540 RSA_METHOD to be chosen if one doesn't exist already.
8541 Previously this was only set during a call to RSA_new()
8542 or RSA_new_method(NULL) meaning it was possible for
8543 RSA_get_default_method() to return NULL.
8546 *) Added native name translation to the existing DSO code
8547 that will convert (if the flag to do so is set) filenames
8548 that are sufficiently small and have no path information
8549 into a canonical native form. Eg. "blah" converted to
8550 "libblah.so" or "blah.dll" etc.
8553 *) New function ERR_error_string_n(e, buf, len) which is like
8554 ERR_error_string(e, buf), but writes at most 'len' bytes
8555 including the 0 terminator. For ERR_error_string_n, 'buf'
8557 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8559 *) CONF library reworked to become more general. A new CONF
8560 configuration file reader "class" is implemented as well as a
8561 new functions (NCONF_*, for "New CONF") to handle it. The now
8562 old CONF_* functions are still there, but are reimplemented to
8563 work in terms of the new functions. Also, a set of functions
8564 to handle the internal storage of the configuration data is
8565 provided to make it easier to write new configuration file
8566 reader "classes" (I can definitely see something reading a
8567 configuration file in XML format, for example), called _CONF_*,
8568 or "the configuration storage API"...
8570 The new configuration file reading functions are:
8572 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8573 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8575 NCONF_default, NCONF_WIN32
8577 NCONF_dump_fp, NCONF_dump_bio
8579 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8580 NCONF_new creates a new CONF object. This works in the same way
8581 as other interfaces in OpenSSL, like the BIO interface.
8582 NCONF_dump_* dump the internal storage of the configuration file,
8583 which is useful for debugging. All other functions take the same
8584 arguments as the old CONF_* functions wth the exception of the
8585 first that must be a `CONF *' instead of a `LHASH *'.
8587 To make it easer to use the new classes with the old CONF_* functions,
8588 the function CONF_set_default_method is provided.
8591 *) Add '-tls1' option to 'openssl ciphers', which was already
8592 mentioned in the documentation but had not been implemented.
8593 (This option is not yet really useful because even the additional
8594 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8597 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8598 OpenSSL-based applications) load shared libraries and bind to
8599 them in a portable way.
8600 [Geoff Thorpe, with contributions from Richard Levitte]
8602 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8604 *) Make sure _lrotl and _lrotr are only used with MSVC.
8606 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8607 (the default implementation of RAND_status).
8609 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8610 to '-clrext' (= clear extensions), as intended and documented.
8611 [Bodo Moeller; inconsistency pointed out by Michael Attili
8614 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8615 was larger than the MD block size.
8616 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8618 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8619 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8620 using the passed key: if the passed key was a private key the result
8621 of X509_print(), for example, would be to print out all the private key
8625 *) des_quad_cksum() byte order bug fix.
8626 [Ulf Möller, using the problem description in krb4-0.9.7, where
8627 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8629 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8631 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8633 *) For easily testing in shell scripts whether some command
8634 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8635 returns with exit code 0 iff no command of the given name is available.
8636 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8637 the output goes to stdout and nothing is printed to stderr.
8638 Additional arguments are always ignored.
8640 Since for each cipher there is a command of the same name,
8641 the 'no-cipher' compilation switches can be tested this way.
8643 ('openssl no-XXX' is not able to detect pseudo-commands such
8644 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8647 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8650 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8651 is set; it will be thrown away anyway because each handshake creates
8653 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8654 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8655 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
8656 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8659 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8660 'Q' and 'R' lose their special meanings (quit/renegotiate).
8661 This is part of what -quiet does; unlike -quiet, -ign_eof
8662 does not suppress any output.
8665 *) Add compatibility options to the purpose and trust code. The
8666 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8667 accepts a certificate or CA, this was the previous behaviour,
8668 with all the associated security issues.
8670 X509_TRUST_COMPAT is the old trust behaviour: only and
8671 automatically trust self signed roots in certificate store. A
8672 new trust setting X509_TRUST_DEFAULT is used to specify that
8673 a purpose has no associated trust setting and it should instead
8674 use the value in the default purpose.
8677 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8678 and fix a memory leak.
8681 *) In util/mkerr.pl (which implements 'make errors'), preserve
8682 reason strings from the previous version of the .c file, as
8683 the default to have only downcase letters (and digits) in
8684 automatically generated reasons codes is not always appropriate.
8687 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8688 using strerror. Previously, ERR_reason_error_string() returned
8689 library names as reason strings for SYSerr; but SYSerr is a special
8690 case where small numbers are errno values, not library numbers.
8693 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8694 converts DSA parameters into DH parameters. (When creating parameters,
8695 DSA_generate_parameters is used.)
8698 *) Include 'length' (recommended exponent length) in C code generated
8699 by 'openssl dhparam -C'.
8702 *) The second argument to set_label in perlasm was already being used
8703 so couldn't be used as a "file scope" flag. Moved to third argument
8707 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8708 instead of RAND_bytes for encryption IVs and salts.
8711 *) Include RAND_status() into RAND_METHOD instead of implementing
8712 it only for md_rand.c Otherwise replacing the PRNG by calling
8713 RAND_set_rand_method would be impossible.
8716 *) Don't let DSA_generate_key() enter an infinite loop if the random
8717 number generation fails.
8720 *) New 'rand' application for creating pseudo-random output.
8723 *) Added configuration support for Linux/IA64
8724 [Rolf Haberrecker <rolf@suse.de>]
8726 *) Assembler module support for Mingw32.
8729 *) Shared library support for HPUX (in shlib/).
8730 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8732 *) Shared library support for Solaris gcc.
8733 [Lutz Behnke <behnke@trustcenter.de>]
8735 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8737 *) PKCS7_encrypt() was adding text MIME headers twice because they
8738 were added manually and by SMIME_crlf_copy().
8741 *) In bntest.c don't call BN_rand with zero bits argument.
8742 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8744 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8745 case was implemented. This caused BN_div_recp() to fail occasionally.
8748 *) Add an optional second argument to the set_label() in the perl
8749 assembly language builder. If this argument exists and is set
8750 to 1 it signals that the assembler should use a symbol whose
8751 scope is the entire file, not just the current function. This
8752 is needed with MASM which uses the format label:: for this scope.
8753 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8755 *) Change the ASN1 types so they are typedefs by default. Before
8756 almost all types were #define'd to ASN1_STRING which was causing
8757 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8761 *) Change names of new functions to the new get1/get0 naming
8762 convention: After 'get1', the caller owns a reference count
8763 and has to call ..._free; 'get0' returns a pointer to some
8764 data structure without incrementing reference counters.
8765 (Some of the existing 'get' functions increment a reference
8766 counter, some don't.)
8767 Similarly, 'set1' and 'add1' functions increase reference
8768 counters or duplicate objects.
8771 *) Allow for the possibility of temp RSA key generation failure:
8772 the code used to assume it always worked and crashed on failure.
8775 *) Fix potential buffer overrun problem in BIO_printf().
8776 [Ulf Möller, using public domain code by Patrick Powell; problem
8777 pointed out by David Sacerdote <das33@cornell.edu>]
8779 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8780 RAND_egd() and RAND_status(). In the command line application,
8781 the EGD socket can be specified like a seed file using RANDFILE
8785 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8786 Some CAs (e.g. Verisign) distribute certificates in this form.
8789 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8790 list to exclude them. This means that no special compilation option
8791 is needed to use anonymous DH: it just needs to be included in the
8795 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8796 EVP_MD_type. The old functionality is available in a new macro called
8797 EVP_MD_md(). Change code that uses it and update docs.
8800 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8801 where the 'void *' argument is replaced by a function pointer argument.
8802 Previously 'void *' was abused to point to functions, which works on
8803 many platforms, but is not correct. As these functions are usually
8804 called by macros defined in OpenSSL header files, most source code
8805 should work without changes.
8808 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8809 sections with information on -D... compiler switches used for
8810 compiling the library so that applications can see them. To enable
8811 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8812 must be defined. E.g.,
8813 #define OPENSSL_ALGORITHM_DEFINES
8814 #include <openssl/opensslconf.h>
8815 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8816 [Richard Levitte, Ulf and Bodo Möller]
8818 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8822 *) Change the 'other' type in certificate aux info to a STACK_OF
8823 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8824 the required ASN1 format: arbitrary types determined by an OID.
8827 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8828 argument to 'req'. This is not because the function is newer or
8829 better than others it just uses the work 'NEW' in the certificate
8830 request header lines. Some software needs this.
8833 *) Reorganise password command line arguments: now passwords can be
8834 obtained from various sources. Delete the PEM_cb function and make
8835 it the default behaviour: i.e. if the callback is NULL and the
8836 usrdata argument is not NULL interpret it as a null terminated pass
8837 phrase. If usrdata and the callback are NULL then the pass phrase
8838 is prompted for as usual.
8841 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8842 the support is automatically enabled. The resulting binaries will
8843 autodetect the card and use it if present.
8844 [Ben Laurie and Compaq Inc.]
8846 *) Work around for Netscape hang bug. This sends certificate request
8847 and server done in one record. Since this is perfectly legal in the
8848 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8849 the bugs/SSLv3 entry for more info.
8852 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8855 *) Add -rand argument to smime and pkcs12 applications and read/write
8859 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8862 *) Add command line password options to the remaining applications.
8865 *) Bug fix for BN_div_recp() for numerators with an even number of
8869 *) More tests in bntest.c, and changed test_bn output.
8872 *) ./config recognizes MacOS X now.
8875 *) Bug fix for BN_div() when the first words of num and divsor are
8876 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8879 *) Add support for various broken PKCS#8 formats, and command line
8880 options to produce them.
8883 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8884 get temporary BIGNUMs from a BN_CTX.
8887 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8891 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8892 include a #define from the old name to the new. The original intent
8893 was that statically linked binaries could for example just call
8894 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8895 link with digests. This never worked becayse SSLeay_add_all_digests()
8896 and SSLeay_add_all_ciphers() were in the same source file so calling
8897 one would link with the other. They are now in separate source files.
8900 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8903 *) Use a less unusual form of the Miller-Rabin primality test (it used
8904 a binary algorithm for exponentiation integrated into the Miller-Rabin
8905 loop, our standard modexp algorithms are faster).
8908 *) Support for the EBCDIC character set completed.
8909 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8911 *) Source code cleanups: use const where appropriate, eliminate casts,
8912 use void * instead of char * in lhash.
8915 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8916 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8917 this the server could overwrite ephemeral keys that the client
8921 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8922 using 50 iterations of the Rabin-Miller test.
8924 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8925 iterations of the Rabin-Miller test as required by the appendix
8926 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8927 As BN_is_prime_fasttest includes trial division, DSA parameter
8928 generation becomes much faster.
8930 This implies a change for the callback functions in DSA_is_prime
8931 and DSA_generate_parameters: The callback function is called once
8932 for each positive witness in the Rabin-Miller test, not just
8933 occasionally in the inner loop; and the parameters to the
8934 callback function now provide an iteration count for the outer
8935 loop rather than for the current invocation of the inner loop.
8936 DSA_generate_parameters additionally can call the callback
8937 function with an 'iteration count' of -1, meaning that a
8938 candidate has passed the trial division test (when q is generated
8939 from an application-provided seed, trial division is skipped).
8942 *) New function BN_is_prime_fasttest that optionally does trial
8943 division before starting the Rabin-Miller test and has
8944 an additional BN_CTX * argument (whereas BN_is_prime always
8945 has to allocate at least one BN_CTX).
8946 'callback(1, -1, cb_arg)' is called when a number has passed the
8947 trial division stage.
8950 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8954 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8957 *) New function BN_pseudo_rand().
8960 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8961 bignum version of BN_from_montgomery() with the working code from
8962 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8966 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8967 made it impossible to use the same SSL_SESSION data structure in
8968 SSL2 clients in multiple threads.
8971 *) The return value of RAND_load_file() no longer counts bytes obtained
8972 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8973 to seed the PRNG (previously an explicit byte count was required).
8974 [Ulf Möller, Bodo Möller]
8976 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8977 used (char *) instead of (void *) and had casts all over the place.
8980 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8983 *) Retain source code compatibility for BN_prime_checks macro:
8984 BN_is_prime(..., BN_prime_checks, ...) now uses
8985 BN_prime_checks_for_size to determine the appropriate number of
8986 Rabin-Miller iterations.
8989 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8990 DH_CHECK_P_NOT_SAFE_PRIME.
8991 (Check if this is true? OpenPGP calls them "strong".)
8994 *) Merge the functionality of "dh" and "gendh" programs into a new program
8995 "dhparam". The old programs are retained for now but will handle DH keys
8996 (instead of parameters) in future.
8999 *) Make the ciphers, s_server and s_client programs check the return values
9000 when a new cipher list is set.
9003 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9004 ciphers. Before when the 56bit ciphers were enabled the sorting was
9007 The syntax for the cipher sorting has been extended to support sorting by
9008 cipher-strength (using the strength_bits hard coded in the tables).
9009 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9011 Fix a bug in the cipher-command parser: when supplying a cipher command
9012 string with an "undefined" symbol (neither command nor alphanumeric
9013 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9014 an error is flagged.
9016 Due to the strength-sorting extension, the code of the
9017 ssl_create_cipher_list() function was completely rearranged. I hope that
9018 the readability was also increased :-)
9019 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9021 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9022 for the first serial number and places 2 in the serial number file. This
9023 avoids problems when the root CA is created with serial number zero and
9024 the first user certificate has the same issuer name and serial number
9028 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9029 the new code. Add documentation for this stuff.
9032 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9033 X509_*() to X509at_*() on the grounds that they don't handle X509
9034 structures and behave in an analagous way to the X509v3 functions:
9035 they shouldn't be called directly but wrapper functions should be used
9038 So we also now have some wrapper functions that call the X509at functions
9039 when passed certificate requests. (TO DO: similar things can be done with
9040 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9041 things. Some of these need some d2i or i2d and print functionality
9042 because they handle more complex structures.)
9045 *) Add missing #ifndefs that caused missing symbols when building libssl
9046 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9047 NO_RSA in ssl/s2*.c.
9048 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9050 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9051 has a return value which indicates the quality of the random data
9052 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9053 error queue. New function RAND_pseudo_bytes() generates output that is
9054 guaranteed to be unique but not unpredictable. RAND_add is like
9055 RAND_seed, but takes an extra argument for an entropy estimate
9056 (RAND_seed always assumes full entropy).
9059 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9060 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9061 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9062 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9063 false-positive rate of at most 2^-80 for random input.
9066 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9069 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9070 in the 0.9.5 release), this returns the chain
9071 from an X509_CTX structure with a dup of the stack and all
9072 the X509 reference counts upped: so the stack will exist
9073 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9076 Also make SSL_SESSION_print() print out the verify return
9080 *) Add manpage for the pkcs12 command. Also change the default
9081 behaviour so MAC iteration counts are used unless the new
9082 -nomaciter option is used. This improves file security and
9083 only older versions of MSIE (4.0 for example) need it.
9086 *) Honor the no-xxx Configure options when creating .DEF files.
9089 *) Add PKCS#10 attributes to field table: challengePassword,
9090 unstructuredName and unstructuredAddress. These are taken from
9091 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9092 international characters are used.
9094 More changes to X509_ATTRIBUTE code: allow the setting of types
9095 based on strings. Remove the 'loc' parameter when adding
9096 attributes because these will be a SET OF encoding which is sorted
9100 *) Initial changes to the 'req' utility to allow request generation
9101 automation. This will allow an application to just generate a template
9102 file containing all the field values and have req construct the
9105 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9106 used all over the place including certificate requests and PKCS#7
9107 structures. They are currently handled manually where necessary with
9108 some primitive wrappers for PKCS#7. The new functions behave in a
9109 manner analogous to the X509 extension functions: they allow
9110 attributes to be looked up by NID and added.
9112 Later something similar to the X509V3 code would be desirable to
9113 automatically handle the encoding, decoding and printing of the
9114 more complex types. The string types like challengePassword can
9115 be handled by the string table functions.
9117 Also modified the multi byte string table handling. Now there is
9118 a 'global mask' which masks out certain types. The table itself
9119 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9120 is useful when for example there is only one permissible type
9121 (as in countryName) and using the mask might result in no valid
9125 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9126 SSL_get_peer_finished to allow applications to obtain the latest
9127 Finished messages sent to the peer or expected from the peer,
9128 respectively. (SSL_get_peer_finished is usually the Finished message
9129 actually received from the peer, otherwise the protocol will be aborted.)
9131 As the Finished message are message digests of the complete handshake
9132 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9133 be used for external authentication procedures when the authentication
9134 provided by SSL/TLS is not desired or is not enough.
9137 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9138 the host supports BWX extension and if Compaq C is present on the
9139 $PATH. Just exploiting of the BWX extension results in 20-30%
9140 performance kick for some algorithms, e.g. DES and RC4 to mention
9141 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9145 *) Add support for MS "fast SGC". This is arguably a violation of the
9146 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9147 weak crypto and after checking the certificate is SGC a second one
9148 with strong crypto. MS SGC stops the first handshake after receiving
9149 the server certificate message and sends a second client hello. Since
9150 a server will typically do all the time consuming operations before
9151 expecting any further messages from the client (server key exchange
9152 is the most expensive) there is little difference between the two.
9154 To get OpenSSL to support MS SGC we have to permit a second client
9155 hello message after we have sent server done. In addition we have to
9156 reset the MAC if we do get this second client hello.
9159 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9160 if a DER encoded private key is RSA or DSA traditional format. Changed
9161 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9162 format DER encoded private key. Newer code should use PKCS#8 format which
9163 has the key type encoded in the ASN1 structure. Added DER private key
9164 support to pkcs8 application.
9167 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9168 ciphersuites has been selected (as required by the SSL 3/TLS 1
9169 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9170 is set, we interpret this as a request to violate the specification
9171 (the worst that can happen is a handshake failure, and 'correct'
9172 behaviour would result in a handshake failure anyway).
9175 *) In SSL_CTX_add_session, take into account that there might be multiple
9176 SSL_SESSION structures with the same session ID (e.g. when two threads
9177 concurrently obtain them from an external cache).
9178 The internal cache can handle only one SSL_SESSION with a given ID,
9179 so if there's a conflict, we now throw out the old one to achieve
9183 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9184 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9185 some routines that use cipher OIDs: some ciphers do not have OIDs
9186 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9190 *) Simplify the trust setting structure and code. Now we just have
9191 two sequences of OIDs for trusted and rejected settings. These will
9192 typically have values the same as the extended key usage extension
9193 and any application specific purposes.
9195 The trust checking code now has a default behaviour: it will just
9196 check for an object with the same NID as the passed id. Functions can
9197 be provided to override either the default behaviour or the behaviour
9198 for a given id. SSL client, server and email already have functions
9199 in place for compatibility: they check the NID and also return "trusted"
9200 if the certificate is self signed.
9203 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9204 traditional format into an EVP_PKEY structure.
9207 *) Add a password callback function PEM_cb() which either prompts for
9208 a password if usr_data is NULL or otherwise assumes it is a null
9209 terminated password. Allow passwords to be passed on command line
9210 environment or config files in a few more utilities.
9213 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9214 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9215 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9216 Update documentation.
9219 *) Support for ASN1 "NULL" type. This could be handled before by using
9220 ASN1_TYPE but there wasn't any function that would try to read a NULL
9221 and produce an error if it couldn't. For compatibility we also have
9222 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9223 don't allocate anything because they don't need to.
9226 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9228 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9230 *) Rebuild of the memory allocation routines used by OpenSSL code and
9231 possibly others as well. The purpose is to make an interface that
9232 provide hooks so anyone can build a separate set of allocation and
9233 deallocation routines to be used by OpenSSL, for example memory
9234 pool implementations, or something else, which was previously hard
9235 since Malloc(), Realloc() and Free() were defined as macros having
9236 the values malloc, realloc and free, respectively (except for Win32
9237 compilations). The same is provided for memory debugging code.
9238 OpenSSL already comes with functionality to find memory leaks, but
9239 this gives people a chance to debug other memory problems.
9241 With these changes, a new set of functions and macros have appeared:
9243 CRYPTO_set_mem_debug_functions() [F]
9244 CRYPTO_get_mem_debug_functions() [F]
9245 CRYPTO_dbg_set_options() [F]
9246 CRYPTO_dbg_get_options() [F]
9247 CRYPTO_malloc_debug_init() [M]
9249 The memory debug functions are NULL by default, unless the library
9250 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9251 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9252 gives the standard debugging functions that come with OpenSSL) or
9253 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9254 provided by the library user) must be used. When the standard
9255 debugging functions are used, CRYPTO_dbg_set_options can be used to
9256 request additional information:
9257 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9258 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9260 Also, things like CRYPTO_set_mem_functions will always give the
9261 expected result (the new set of functions is used for allocation
9262 and deallocation) at all times, regardless of platform and compiler
9265 To finish it up, some functions that were never use in any other
9266 way than through macros have a new API and new semantic:
9269 CRYPTO_dbg_realloc()
9272 All macros of value have retained their old syntax.
9273 [Richard Levitte and Bodo Moeller]
9275 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9276 ordering of SMIMECapabilities wasn't in "strength order" and there
9277 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9281 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9282 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9283 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9285 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9286 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9287 functionality to handle multipart/signed properly) and a utility
9288 called 'smime' to call all this stuff. This is based on code I
9289 originally wrote for Celo who have kindly allowed it to be
9290 included in OpenSSL.
9293 *) Add variants des_set_key_checked and des_set_key_unchecked of
9294 des_set_key (aka des_key_sched). Global variable des_check_key
9295 decides which of these is called by des_set_key; this way
9296 des_check_key behaves as it always did, but applications and
9297 the library itself, which was buggy for des_check_key == 1,
9298 have a cleaner way to pick the version they need.
9301 *) New function PKCS12_newpass() which changes the password of a
9305 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9306 dynamic mix. In both cases the ids can be used as an index into the
9307 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9308 functions so they accept a list of the field values and the
9309 application doesn't need to directly manipulate the X509_TRUST
9313 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9317 *) Modify the way the V3 extension code looks up extensions. This now
9318 works in a similar way to the object code: we have some "standard"
9319 extensions in a static table which is searched with OBJ_bsearch()
9320 and the application can add dynamic ones if needed. The file
9321 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9322 updated whenever a new extension is added to the core code and kept
9323 in ext_nid order. There is a simple program 'tabtest.c' which checks
9324 this. New extensions are not added too often so this file can readily
9325 be maintained manually.
9327 There are two big advantages in doing things this way. The extensions
9328 can be looked up immediately and no longer need to be "added" using
9329 X509V3_add_standard_extensions(): this function now does nothing.
9330 [Side note: I get *lots* of email saying the extension code doesn't
9331 work because people forget to call this function]
9332 Also no dynamic allocation is done unless new extensions are added:
9333 so if we don't add custom extensions there is no need to call
9334 X509V3_EXT_cleanup().
9337 *) Modify enc utility's salting as follows: make salting the default. Add a
9338 magic header, so unsalted files fail gracefully instead of just decrypting
9339 to garbage. This is because not salting is a big security hole, so people
9340 should be discouraged from doing it.
9343 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9344 digest to be passed on the command line but it only used this
9345 parameter when signing a certificate. Modified so all relevant
9346 operations are affected by the digest parameter including the
9347 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9348 DSA key was used because it didn't fix the digest.
9351 *) Initial certificate chain verify code. Currently tests the untrusted
9352 certificates for consistency with the verify purpose (which is set
9353 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9355 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9356 this is because it will reject chains with invalid extensions whereas
9357 every previous version of OpenSSL and SSLeay made no checks at all.
9359 Trust code: checks the root CA for the relevant trust settings. Trust
9360 settings have an initial value consistent with the verify purpose: e.g.
9361 if the verify purpose is for SSL client use it expects the CA to be
9362 trusted for SSL client use. However the default value can be changed to
9363 permit custom trust settings: one example of this would be to only trust
9364 certificates from a specific "secure" set of CAs.
9366 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9367 which should be used for version portability: especially since the
9368 verify structure is likely to change more often now.
9370 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9371 to set them. If not set then assume SSL clients will verify SSL servers
9374 Two new options to the verify program: -untrusted allows a set of
9375 untrusted certificates to be passed in and -purpose which sets the
9376 intended purpose of the certificate. If a purpose is set then the
9377 new chain verify code is used to check extension consistency.
9380 *) Support for the authority information access extension.
9383 *) Modify RSA and DSA PEM read routines to transparently handle
9384 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9385 public keys in a format compatible with certificate
9386 SubjectPublicKeyInfo structures. Unfortunately there were already
9387 functions called *_PublicKey_* which used various odd formats so
9388 these are retained for compatibility: however the DSA variants were
9389 never in a public release so they have been deleted. Changed dsa/rsa
9390 utilities to handle the new format: note no releases ever handled public
9391 keys so we should be OK.
9393 The primary motivation for this change is to avoid the same fiasco
9394 that dogs private keys: there are several incompatible private key
9395 formats some of which are standard and some OpenSSL specific and
9396 require various evil hacks to allow partial transparent handling and
9397 even then it doesn't work with DER formats. Given the option anything
9398 other than PKCS#8 should be dumped: but the other formats have to
9399 stay in the name of compatibility.
9401 With public keys and the benefit of hindsight one standard format
9402 is used which works with EVP_PKEY, RSA or DSA structures: though
9403 it clearly returns an error if you try to read the wrong kind of key.
9405 Added a -pubkey option to the 'x509' utility to output the public key.
9406 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9407 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9408 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9409 that do the same as the EVP_PKEY_assign_*() except they up the
9410 reference count of the added key (they don't "swallow" the
9414 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9415 CRLs would fail if the file contained no certificates or no CRLs:
9416 added a new function to read in both types and return the number
9417 read: this means that if none are read it will be an error. The
9418 DER versions of the certificate and CRL reader would always fail
9419 because it isn't possible to mix certificates and CRLs in DER format
9420 without choking one or the other routine. Changed this to just read
9421 a certificate: this is the best we can do. Also modified the code
9422 in apps/verify.c to take notice of return codes: it was previously
9423 attempting to read in certificates from NULL pointers and ignoring
9424 any errors: this is one reason why the cert and CRL reader seemed
9425 to work. It doesn't check return codes from the default certificate
9426 routines: these may well fail if the certificates aren't installed.
9429 *) Code to support otherName option in GeneralName.
9432 *) First update to verify code. Change the verify utility
9433 so it warns if it is passed a self signed certificate:
9434 for consistency with the normal behaviour. X509_verify
9435 has been modified to it will now verify a self signed
9436 certificate if *exactly* the same certificate appears
9437 in the store: it was previously impossible to trust a
9438 single self signed certificate. This means that:
9439 openssl verify ss.pem
9440 now gives a warning about a self signed certificate but
9441 openssl verify -CAfile ss.pem ss.pem
9445 *) For servers, store verify_result in SSL_SESSION data structure
9446 (and add it to external session representation).
9447 This is needed when client certificate verifications fails,
9448 but an application-provided verification callback (set by
9449 SSL_CTX_set_cert_verify_callback) allows accepting the session
9450 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9451 but returns 1): When the session is reused, we have to set
9452 ssl->verify_result to the appropriate error code to avoid
9454 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9456 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9457 case in PKCS7_dataInit() where the signed PKCS7 structure
9458 didn't contain any existing data because it was being created.
9459 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9461 *) Add a salt to the key derivation routines in enc.c. This
9462 forms the first 8 bytes of the encrypted file. Also add a
9463 -S option to allow a salt to be input on the command line.
9466 *) New function X509_cmp(). Oddly enough there wasn't a function
9467 to compare two certificates. We do this by working out the SHA1
9468 hash and comparing that. X509_cmp() will be needed by the trust
9472 *) SSL_get1_session() is like SSL_get_session(), but increments
9473 the reference count in the SSL_SESSION returned.
9474 [Geoff Thorpe <geoff@eu.c2.net>]
9476 *) Fix for 'req': it was adding a null to request attributes.
9477 Also change the X509_LOOKUP and X509_INFO code to handle
9478 certificate auxiliary information.
9481 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9485 *) Add the possibility to add extra information to the memory leak
9486 detecting output, to form tracebacks, showing from where each
9487 allocation was originated: CRYPTO_push_info("constant string") adds
9488 the string plus current file name and line number to a per-thread
9489 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9490 is like calling CYRPTO_pop_info() until the stack is empty.
9491 Also updated memory leak detection code to be multi-thread-safe.
9494 *) Add options -text and -noout to pkcs7 utility and delete the
9495 encryption options which never did anything. Update docs.
9498 *) Add options to some of the utilities to allow the pass phrase
9499 to be included on either the command line (not recommended on
9500 OSes like Unix) or read from the environment. Update the
9501 manpages and fix a few bugs.
9504 *) Add a few manpages for some of the openssl commands.
9507 *) Fix the -revoke option in ca. It was freeing up memory twice,
9508 leaking and not finding already revoked certificates.
9511 *) Extensive changes to support certificate auxiliary information.
9512 This involves the use of X509_CERT_AUX structure and X509_AUX
9513 functions. An X509_AUX function such as PEM_read_X509_AUX()
9514 can still read in a certificate file in the usual way but it
9515 will also read in any additional "auxiliary information". By
9516 doing things this way a fair degree of compatibility can be
9517 retained: existing certificates can have this information added
9518 using the new 'x509' options.
9520 Current auxiliary information includes an "alias" and some trust
9521 settings. The trust settings will ultimately be used in enhanced
9522 certificate chain verification routines: currently a certificate
9523 can only be trusted if it is self signed and then it is trusted
9527 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9528 The problem was that one of the replacement routines had not been working
9529 since SSLeay releases. For now the offending routine has been replaced
9530 with non-optimised assembler. Even so, this now gives around 95%
9531 performance improvement for 1024 bit RSA signs.
9534 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9535 handling. Most clients have the effective key size in bits equal to
9536 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9537 A few however don't do this and instead use the size of the decrypted key
9538 to determine the RC2 key length and the AlgorithmIdentifier to determine
9539 the effective key length. In this case the effective key length can still
9540 be 40 bits but the key length can be 168 bits for example. This is fixed
9541 by manually forcing an RC2 key into the EVP_PKEY structure because the
9542 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9543 the key length and effective key length are equal.
9546 *) Add a bunch of functions that should simplify the creation of
9547 X509_NAME structures. Now you should be able to do:
9548 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9549 and have it automatically work out the correct field type and fill in
9550 the structures. The more adventurous can try:
9551 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9552 and it will (hopefully) work out the correct multibyte encoding.
9555 *) Change the 'req' utility to use the new field handling and multibyte
9556 copy routines. Before the DN field creation was handled in an ad hoc
9557 way in req, ca, and x509 which was rather broken and didn't support
9558 BMPStrings or UTF8Strings. Since some software doesn't implement
9559 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9560 using the dirstring_type option. See the new comment in the default
9561 openssl.cnf for more info.
9564 *) Make crypto/rand/md_rand.c more robust:
9565 - Assure unique random numbers after fork().
9566 - Make sure that concurrent threads access the global counter and
9567 md serializably so that we never lose entropy in them
9568 or use exactly the same state in multiple threads.
9569 Access to the large state is not always serializable because
9570 the additional locking could be a performance killer, and
9571 md should be large enough anyway.
9574 *) New file apps/app_rand.c with commonly needed functionality
9575 for handling the random seed file.
9577 Use the random seed file in some applications that previously did not:
9579 dsaparam -genkey (which also ignored its '-rand' option),
9582 x509 (when signing).
9583 Except on systems with /dev/urandom, it is crucial to have a random
9584 seed file at least for key creation, DSA signing, and for DH exchanges;
9585 for RSA signatures we could do without one.
9587 gendh and gendsa (unlike genrsa) used to read only the first byte
9588 of each file listed in the '-rand' option. The function as previously
9589 found in genrsa is now in app_rand.c and is used by all programs
9590 that support '-rand'.
9593 *) In RAND_write_file, use mode 0600 for creating files;
9594 don't just chmod when it may be too late.
9597 *) Report an error from X509_STORE_load_locations
9598 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9601 *) New function ASN1_mbstring_copy() this copies a string in either
9602 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9603 into an ASN1_STRING type. A mask of permissible types is passed
9604 and it chooses the "minimal" type to use or an error if not type
9608 *) Add function equivalents to the various macros in asn1.h. The old
9609 macros are retained with an M_ prefix. Code inside the library can
9610 use the M_ macros. External code (including the openssl utility)
9611 should *NOT* in order to be "shared library friendly".
9614 *) Add various functions that can check a certificate's extensions
9615 to see if it usable for various purposes such as SSL client,
9616 server or S/MIME and CAs of these types. This is currently
9617 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9618 verification. Also added a -purpose flag to x509 utility to
9619 print out all the purposes.
9622 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9626 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9627 for, obtain and decode and extension and obtain its critical flag.
9628 This allows all the necessary extension code to be handled in a
9629 single function call.
9632 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9633 platforms. See crypto/rc4/rc4_enc.c for further details.
9636 *) New -noout option to asn1parse. This causes no output to be produced
9637 its main use is when combined with -strparse and -out to extract data
9638 from a file (which may not be in ASN.1 format).
9641 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9642 when producing the local key id.
9643 [Richard Levitte <levitte@stacken.kth.se>]
9645 *) New option -dhparam in s_server. This allows a DH parameter file to be
9646 stated explicitly. If it is not stated then it tries the first server
9647 certificate file. The previous behaviour hard coded the filename
9651 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9652 a public key to be input or output. For example:
9653 openssl rsa -in key.pem -pubout -out pubkey.pem
9654 Also added necessary DSA public key functions to handle this.
9657 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9658 in the message. This was handled by allowing
9659 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9660 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9662 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9663 to the end of the strings whereas this didn't. This would cause problems
9664 if strings read with d2i_ASN1_bytes() were later modified.
9665 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9667 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9668 data and it contains EOF it will end up returning an error. This is
9669 caused by input 46 bytes long. The cause is due to the way base64
9670 BIOs find the start of base64 encoded data. They do this by trying a
9671 trial decode on each line until they find one that works. When they
9672 do a flag is set and it starts again knowing it can pass all the
9673 data directly through the decoder. Unfortunately it doesn't reset
9674 the context it uses. This means that if EOF is reached an attempt
9675 is made to pass two EOFs through the context and this causes the
9676 resulting error. This can also cause other problems as well. As is
9677 usual with these problems it takes *ages* to find and the fix is
9678 trivial: move one line.
9679 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9681 *) Ugly workaround to get s_client and s_server working under Windows. The
9682 old code wouldn't work because it needed to select() on sockets and the
9683 tty (for keypresses and to see if data could be written). Win32 only
9684 supports select() on sockets so we select() with a 1s timeout on the
9685 sockets and then see if any characters are waiting to be read, if none
9686 are present then we retry, we also assume we can always write data to
9687 the tty. This isn't nice because the code then blocks until we've
9688 received a complete line of data and it is effectively polling the
9689 keyboard at 1s intervals: however it's quite a bit better than not
9690 working at all :-) A dedicated Windows application might handle this
9691 with an event loop for example.
9694 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9695 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9696 will be called when RSA_sign() and RSA_verify() are used. This is useful
9697 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9698 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9699 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9700 This necessitated the support of an extra signature type NID_md5_sha1
9701 for SSL signatures and modifications to the SSL library to use it instead
9702 of calling RSA_public_decrypt() and RSA_private_encrypt().
9705 *) Add new -verify -CAfile and -CApath options to the crl program, these
9706 will lookup a CRL issuers certificate and verify the signature in a
9707 similar way to the verify program. Tidy up the crl program so it
9708 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9709 less strict. It will now permit CRL extensions even if it is not
9710 a V2 CRL: this will allow it to tolerate some broken CRLs.
9713 *) Initialize all non-automatic variables each time one of the openssl
9714 sub-programs is started (this is necessary as they may be started
9715 multiple times from the "OpenSSL>" prompt).
9716 [Lennart Bang, Bodo Moeller]
9718 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9719 removing all other RSA functionality (this is what NO_RSA does). This
9720 is so (for example) those in the US can disable those operations covered
9721 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9725 *) Non-copying interface to BIO pairs.
9726 (still largely untested)
9729 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9730 ASCII string. This was handled independently in various places before.
9733 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9734 UTF8 strings a character at a time.
9737 *) Use client_version from client hello to select the protocol
9738 (s23_srvr.c) and for RSA client key exchange verification
9739 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9742 *) Add various utility functions to handle SPKACs, these were previously
9743 handled by poking round in the structure internals. Added new function
9744 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9745 print, verify and generate SPKACs. Based on an original idea from
9746 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9749 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9752 *) Allow the config file extension section to be overwritten on the
9753 command line. Based on an original idea from Massimiliano Pala
9754 <madwolf@comune.modena.it>. The new option is called -extensions
9755 and can be applied to ca, req and x509. Also -reqexts to override
9756 the request extensions in req and -crlexts to override the crl extensions
9760 *) Add new feature to the SPKAC handling in ca. Now you can include
9761 the same field multiple times by preceding it by "XXXX." for example:
9764 this is the same syntax as used in the req config file.
9767 *) Allow certificate extensions to be added to certificate requests. These
9768 are specified in a 'req_extensions' option of the req section of the
9769 config file. They can be printed out with the -text option to req but
9770 are otherwise ignored at present.
9773 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9774 data read consists of only the final block it would not decrypted because
9775 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9776 A misplaced 'break' also meant the decrypted final block might not be
9777 copied until the next read.
9780 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9781 a few extra parameters to the DH structure: these will be useful if
9782 for example we want the value of 'q' or implement X9.42 DH.
9785 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9786 provides hooks that allow the default DSA functions or functions on a
9787 "per key" basis to be replaced. This allows hardware acceleration and
9788 hardware key storage to be handled without major modification to the
9789 library. Also added low level modexp hooks and CRYPTO_EX structure and
9790 associated functions.
9793 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9794 as "read only": it can't be written to and the buffer it points to will
9795 not be freed. Reading from a read only BIO is much more efficient than
9796 a normal memory BIO. This was added because there are several times when
9797 an area of memory needs to be read from a BIO. The previous method was
9798 to create a memory BIO and write the data to it, this results in two
9799 copies of the data and an O(n^2) reading algorithm. There is a new
9800 function BIO_new_mem_buf() which creates a read only memory BIO from
9801 an area of memory. Also modified the PKCS#7 routines to use read only
9805 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9806 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9807 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9808 but a retry condition occured while trying to read the rest.
9811 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9812 NID_pkcs7_encrypted by default: this was wrong since this should almost
9813 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9814 the encrypted data type: this is a more sensible place to put it and it
9815 allows the PKCS#12 code to be tidied up that duplicated this
9819 *) Changed obj_dat.pl script so it takes its input and output files on
9820 the command line. This should avoid shell escape redirection problems
9824 *) Initial support for certificate extension requests, these are included
9825 in things like Xenroll certificate requests. Included functions to allow
9826 extensions to be obtained and added.
9829 *) -crlf option to s_client and s_server for sending newlines as
9830 CRLF (as required by many protocols).
9833 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9835 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9836 [Ralf S. Engelschall]
9838 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9839 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9841 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9845 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9846 DH parameters/keys (q is lost during that conversion, but the resulting
9847 DH parameters contain its length).
9849 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9850 much faster than DH_generate_parameters (which creates parameters
9851 where p = 2*q + 1), and also the smaller q makes DH computations
9852 much more efficient (160-bit exponentiation instead of 1024-bit
9853 exponentiation); so this provides a convenient way to support DHE
9854 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9855 utter importance to use
9856 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9858 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9859 when such DH parameters are used, because otherwise small subgroup
9860 attacks may become possible!
9863 *) Avoid memory leak in i2d_DHparams.
9866 *) Allow the -k option to be used more than once in the enc program:
9867 this allows the same encrypted message to be read by multiple recipients.
9870 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9871 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9872 it will always use the numerical form of the OID, even if it has a short
9876 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9877 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9878 otherwise bn_mod_exp was called. In the case of hardware keys for example
9879 no private key components need be present and it might store extra data
9880 in the RSA structure, which cannot be accessed from bn_mod_exp.
9881 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9882 private key operations.
9885 *) Added support for SPARC Linux.
9888 *) pem_password_cb function type incompatibly changed from
9889 typedef int pem_password_cb(char *buf, int size, int rwflag);
9891 ....(char *buf, int size, int rwflag, void *userdata);
9892 so that applications can pass data to their callbacks:
9893 The PEM[_ASN1]_{read,write}... functions and macros now take an
9894 additional void * argument, which is just handed through whenever
9895 the password callback is called.
9896 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9898 New function SSL_CTX_set_default_passwd_cb_userdata.
9900 Compatibility note: As many C implementations push function arguments
9901 onto the stack in reverse order, the new library version is likely to
9902 interoperate with programs that have been compiled with the old
9903 pem_password_cb definition (PEM_whatever takes some data that
9904 happens to be on the stack as its last argument, and the callback
9905 just ignores this garbage); but there is no guarantee whatsoever that
9908 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9909 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9910 problems not only on Windows, but also on some Unix platforms.
9911 To avoid problematic command lines, these definitions are now in an
9912 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9913 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9916 *) MIPS III/IV assembler module is reimplemented.
9919 *) More DES library cleanups: remove references to srand/rand and
9920 delete an unused file.
9923 *) Add support for the the free Netwide assembler (NASM) under Win32,
9924 since not many people have MASM (ml) and it can be hard to obtain.
9925 This is currently experimental but it seems to work OK and pass all
9926 the tests. Check out INSTALL.W32 for info.
9929 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9930 without temporary keys kept an extra copy of the server key,
9931 and connections with temporary keys did not free everything in case
9935 *) New function RSA_check_key and new openssl rsa option -check
9936 for verifying the consistency of RSA keys.
9937 [Ulf Moeller, Bodo Moeller]
9939 *) Various changes to make Win32 compile work:
9940 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9941 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9942 comparison" warnings.
9943 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9946 *) Add a debugging option to PKCS#5 v2 key generation function: when
9947 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9948 derived keys are printed to stderr.
9951 *) Copy the flags in ASN1_STRING_dup().
9952 [Roman E. Pavlov <pre@mo.msk.ru>]
9954 *) The x509 application mishandled signing requests containing DSA
9955 keys when the signing key was also DSA and the parameters didn't match.
9957 It was supposed to omit the parameters when they matched the signing key:
9958 the verifying software was then supposed to automatically use the CA's
9959 parameters if they were absent from the end user certificate.
9961 Omitting parameters is no longer recommended. The test was also
9962 the wrong way round! This was probably due to unusual behaviour in
9963 EVP_cmp_parameters() which returns 1 if the parameters match.
9964 This meant that parameters were omitted when they *didn't* match and
9965 the certificate was useless. Certificates signed with 'ca' didn't have
9967 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9969 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9970 The interface is as follows:
9971 Applications can use
9972 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9973 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9974 "off" is now the default.
9975 The library internally uses
9976 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9977 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9978 to disable memory-checking temporarily.
9980 Some inconsistent states that previously were possible (and were
9981 even the default) are now avoided.
9983 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9984 with each memory chunk allocated; this is occasionally more helpful
9985 than just having a counter.
9987 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9989 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9993 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9994 which largely parallels "options", but is for changing API behaviour,
9995 whereas "options" are about protocol behaviour.
9996 Initial "mode" flags are:
9998 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9999 a single record has been written.
10000 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10001 retries use the same buffer location.
10002 (But all of the contents must be
10006 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10009 *) Fix problems with no-hmac etc.
10010 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10012 *) New functions RSA_get_default_method(), RSA_set_method() and
10013 RSA_get_method(). These allows replacement of RSA_METHODs without having
10014 to mess around with the internals of an RSA structure.
10017 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10018 Also really enable memory leak checks in openssl.c and in some
10020 [Chad C. Mulligan, Bodo Moeller]
10022 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10023 up the length of negative integers. This has now been simplified to just
10024 store the length when it is first determined and use it later, rather
10025 than trying to keep track of where data is copied and updating it to
10027 [Steve Henson, reported by Brien Wheeler
10028 <bwheeler@authentica-security.com>]
10030 *) Add a new function PKCS7_signatureVerify. This allows the verification
10031 of a PKCS#7 signature but with the signing certificate passed to the
10032 function itself. This contrasts with PKCS7_dataVerify which assumes the
10033 certificate is present in the PKCS#7 structure. This isn't always the
10034 case: certificates can be omitted from a PKCS#7 structure and be
10035 distributed by "out of band" means (such as a certificate database).
10038 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10039 function prototypes in pem.h, also change util/mkdef.pl to add the
10040 necessary function names.
10043 *) mk1mf.pl (used by Windows builds) did not properly read the
10044 options set by Configure in the top level Makefile, and Configure
10045 was not even able to write more than one option correctly.
10046 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10049 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10050 file to be loaded from a BIO or FILE pointer. The BIO version will
10051 for example allow memory BIOs to contain config info.
10054 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10055 Whoever hopes to achieve shared-library compatibility across versions
10056 must use this, not the compile-time macro.
10057 (Exercise 0.9.4: Which is the minimum library version required by
10059 Note: All this applies only to multi-threaded programs, others don't
10063 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10064 through a BIO pair triggered the default case, i.e.
10065 SSLerr(...,SSL_R_UNKNOWN_STATE).
10068 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10069 can use the SSL library even if none of the specific BIOs is
10073 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10074 for the encoded length.
10075 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10077 *) Add initial documentation of the X509V3 functions.
10080 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10081 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10082 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10083 secure PKCS#8 private key format with a high iteration count.
10086 *) Fix determination of Perl interpreter: A perl or perl5
10087 _directory_ in $PATH was also accepted as the interpreter.
10088 [Ralf S. Engelschall]
10090 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10091 wrong with it but it was very old and did things like calling
10092 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10093 unusual formatting.
10096 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10097 to use the new extension code.
10100 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10101 with macros. This should make it easier to change their form, add extra
10102 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10106 *) Add to configuration table a new entry that can specify an alternative
10107 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10108 according to Mark Crispin <MRC@Panda.COM>.
10112 *) DES CBC did not update the IV. Weird.
10115 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10116 Changing the behaviour of the former might break existing programs --
10117 where IV updating is needed, des_ncbc_encrypt can be used.
10120 *) When bntest is run from "make test" it drives bc to check its
10121 calculations, as well as internally checking them. If an internal check
10122 fails, it needs to cause bc to give a non-zero result or make test carries
10123 on without noticing the failure. Fixed.
10126 *) DES library cleanups.
10129 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10130 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10131 ciphers. NOTE: although the key derivation function has been verified
10132 against some published test vectors it has not been extensively tested
10133 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10137 *) Instead of "mkdir -p", which is not fully portable, use new
10138 Perl script "util/mkdir-p.pl".
10141 *) Rewrite the way password based encryption (PBE) is handled. It used to
10142 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10143 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10144 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10145 the 'parameter' field of the AlgorithmIdentifier is passed to the
10146 underlying key generation function so it must do its own ASN1 parsing.
10147 This has also changed the EVP_PBE_CipherInit() function which now has a
10148 'parameter' argument instead of literal salt and iteration count values
10149 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10152 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10153 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10154 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10155 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10156 value was just used as a "magic string" and not used directly its
10157 value doesn't matter.
10160 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10164 *) "linux-sparc64" configuration (ultrapenguin).
10165 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10166 "linux-sparc" configuration.
10167 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10169 *) config now generates no-xxx options for missing ciphers.
10172 *) Support the EBCDIC character set (work in progress).
10173 File ebcdic.c not yet included because it has a different license.
10174 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10176 *) Support BS2000/OSD-POSIX.
10177 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10179 *) Make callbacks for key generation use void * instead of char *.
10182 *) Make S/MIME samples compile (not yet tested).
10185 *) Additional typesafe stacks.
10188 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10192 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10194 *) New configuration variant "sco5-gcc".
10196 *) Updated some demos.
10197 [Sean O Riordain, Wade Scholine]
10199 *) Add missing BIO_free at exit of pkcs12 application.
10202 *) Fix memory leak in conf.c.
10205 *) Updates for Win32 to assembler version of MD5.
10208 *) Set #! path to perl in apps/der_chop to where we found it
10209 instead of using a fixed path.
10212 *) SHA library changes for irix64-mips4-cc.
10215 *) Improvements for VMS support.
10219 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10221 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10222 This also avoids the problems with SC4.2 and unpatched SC5.
10223 [Andy Polyakov <appro@fy.chalmers.se>]
10225 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10226 These are required because of the typesafe stack would otherwise break
10227 existing code. If old code used a structure member which used to be STACK
10228 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10229 sk_num or sk_value it would produce an error because the num, data members
10230 are not present in STACK_OF. Now it just produces a warning. sk_set
10231 replaces the old method of assigning a value to sk_value
10232 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10233 that does this will no longer work (and should use sk_set instead) but
10234 this could be regarded as a "questionable" behaviour anyway.
10237 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10238 correctly handle encrypted S/MIME data.
10241 *) Change type of various DES function arguments from des_cblock
10242 (which means, in function argument declarations, pointer to char)
10243 to des_cblock * (meaning pointer to array with 8 char elements),
10244 which allows the compiler to do more typechecking; it was like
10245 that back in SSLeay, but with lots of ugly casts.
10247 Introduce new type const_des_cblock.
10250 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10251 problems: find RecipientInfo structure that matches recipient certificate
10252 and initialise the ASN1 structures properly based on passed cipher.
10255 *) Belatedly make the BN tests actually check the results.
10258 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10259 to and from BNs: it was completely broken. New compilation option
10260 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10261 key elements as negative integers.
10264 *) Reorganize and speed up MD5.
10265 [Andy Polyakov <appro@fy.chalmers.se>]
10268 [Richard Levitte <richard@levitte.org>]
10270 *) New option -out to asn1parse to allow the parsed structure to be
10271 output to a file. This is most useful when combined with the -strparse
10272 option to examine the output of things like OCTET STRINGS.
10275 *) Make SSL library a little more fool-proof by not requiring any longer
10276 that SSL_set_{accept,connect}_state be called before
10277 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10278 in many applications because usually everything *appeared* to work as
10279 intended anyway -- now it really works as intended).
10282 *) Move openssl.cnf out of lib/.
10285 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10286 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10287 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10288 [Ralf S. Engelschall]
10290 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10291 handle PKCS#7 enveloped data properly.
10292 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10294 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10295 copying pointers. The cert_st handling is changed by this in
10296 various ways (and thus what used to be known as ctx->default_cert
10297 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10298 any longer when s->cert does not give us what we need).
10299 ssl_cert_instantiate becomes obsolete by this change.
10300 As soon as we've got the new code right (possibly it already is?),
10301 we have solved a couple of bugs of the earlier code where s->cert
10302 was used as if it could not have been shared with other SSL structures.
10304 Note that using the SSL API in certain dirty ways now will result
10305 in different behaviour than observed with earlier library versions:
10306 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10307 does not influence s as it used to.
10309 In order to clean up things more thoroughly, inside SSL_SESSION
10310 we don't use CERT any longer, but a new structure SESS_CERT
10311 that holds per-session data (if available); currently, this is
10312 the peer's certificate chain and, for clients, the server's certificate
10313 and temporary key. CERT holds only those values that can have
10314 meaningful defaults in an SSL_CTX.
10317 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10318 from the internal representation. Various PKCS#7 fixes: remove some
10319 evil casts and set the enc_dig_alg field properly based on the signing
10323 *) Allow PKCS#12 password to be set from the command line or the
10324 environment. Let 'ca' get its config file name from the environment
10325 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10329 *) Allow certificate policies extension to use an IA5STRING for the
10330 organization field. This is contrary to the PKIX definition but
10331 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10335 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10336 without disallowing inline assembler and the like for non-pedantic builds.
10339 *) Support Borland C++ builder.
10340 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10342 *) Support Mingw32.
10345 *) SHA-1 cleanups and performance enhancements.
10346 [Andy Polyakov <appro@fy.chalmers.se>]
10348 *) Sparc v8plus assembler for the bignum library.
10349 [Andy Polyakov <appro@fy.chalmers.se>]
10351 *) Accept any -xxx and +xxx compiler options in Configure.
10354 *) Update HPUX configuration.
10357 *) Add missing sk_<type>_unshift() function to safestack.h
10358 [Ralf S. Engelschall]
10360 *) New function SSL_CTX_use_certificate_chain_file that sets the
10361 "extra_cert"s in addition to the certificate. (This makes sense
10362 only for "PEM" format files, as chains as a whole are not
10366 *) Support verify_depth from the SSL API.
10367 x509_vfy.c had what can be considered an off-by-one-error:
10368 Its depth (which was not part of the external interface)
10369 was actually counting the number of certificates in a chain;
10370 now it really counts the depth.
10373 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10374 instead of X509err, which often resulted in confusing error
10375 messages since the error codes are not globally unique
10376 (e.g. an alleged error in ssl3_accept when a certificate
10377 didn't match the private key).
10379 *) New function SSL_CTX_set_session_id_context that allows to set a default
10380 value (so that you don't need SSL_set_session_id_context for each
10381 connection using the SSL_CTX).
10384 *) OAEP decoding bug fix.
10387 *) Support INSTALL_PREFIX for package builders, as proposed by
10391 *) New Configure options "threads" and "no-threads". For systems
10392 where the proper compiler options are known (currently Solaris
10393 and Linux), "threads" is the default.
10396 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10399 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10400 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10401 such as /usr/local/bin.
10404 *) "make linux-shared" to build shared libraries.
10405 [Niels Poppe <niels@netbox.org>]
10407 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10410 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10411 extension adding in x509 utility.
10414 *) Remove NOPROTO sections and error code comments.
10417 *) Partial rewrite of the DEF file generator to now parse the ANSI
10421 *) New Configure options --prefix=DIR and --openssldir=DIR.
10424 *) Complete rewrite of the error code script(s). It is all now handled
10425 by one script at the top level which handles error code gathering,
10426 header rewriting and C source file generation. It should be much better
10427 than the old method: it now uses a modified version of Ulf's parser to
10428 read the ANSI prototypes in all header files (thus the old K&R definitions
10429 aren't needed for error creation any more) and do a better job of
10430 translating function codes into names. The old 'ASN1 error code imbedded
10431 in a comment' is no longer necessary and it doesn't use .err files which
10432 have now been deleted. Also the error code call doesn't have to appear all
10433 on one line (which resulted in some large lines...).
10436 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10439 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10440 0 (which usually indicates a closed connection), but continue reading.
10443 *) Fix some race conditions.
10446 *) Add support for CRL distribution points extension. Add Certificate
10447 Policies and CRL distribution points documentation.
10450 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10453 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10454 8 of keying material. Merlin has also confirmed interop with this fix
10455 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10456 [Merlin Hughes <merlin@baltimore.ie>]
10458 *) Fix lots of warnings.
10459 [Richard Levitte <levitte@stacken.kth.se>]
10461 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10462 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10463 [Richard Levitte <levitte@stacken.kth.se>]
10465 *) Fix problems with sizeof(long) == 8.
10466 [Andy Polyakov <appro@fy.chalmers.se>]
10468 *) Change functions to ANSI C.
10471 *) Fix typos in error codes.
10472 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10474 *) Remove defunct assembler files from Configure.
10477 *) SPARC v8 assembler BIGNUM implementation.
10478 [Andy Polyakov <appro@fy.chalmers.se>]
10480 *) Support for Certificate Policies extension: both print and set.
10481 Various additions to support the r2i method this uses.
10484 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10485 return a const string when you are expecting an allocated buffer.
10488 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10489 types DirectoryString and DisplayText.
10492 *) Add code to allow r2i extensions to access the configuration database,
10493 add an LHASH database driver and add several ctx helper functions.
10496 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10497 fail when they extended the size of a BIGNUM.
10500 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10501 support typesafe stack.
10504 *) Fix typo in SSL_[gs]et_options().
10505 [Nils Frostberg <nils@medcom.se>]
10507 *) Delete various functions and files that belonged to the (now obsolete)
10508 old X509V3 handling code.
10511 *) New Configure option "rsaref".
10514 *) Don't auto-generate pem.h.
10517 *) Introduce type-safe ASN.1 SETs.
10520 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10521 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10523 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10524 that links with OpenSSL (well at least cause lots of warnings), but fear
10525 not: the conversion is trivial, and it eliminates loads of evil casts. A
10526 few STACKed things have been converted already. Feel free to convert more.
10527 In the fullness of time, I'll do away with the STACK type altogether.
10530 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10531 specified in <certfile> by updating the entry in the index.txt file.
10532 This way one no longer has to edit the index.txt file manually for
10533 revoking a certificate. The -revoke option does the gory details now.
10534 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10536 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10537 `-text' option at all and this way the `-noout -text' combination was
10538 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10539 [Ralf S. Engelschall]
10541 *) Make sure a corresponding plain text error message exists for the
10542 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10543 verify callback function determined that a certificate was revoked.
10544 [Ralf S. Engelschall]
10546 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10547 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10548 all available cipers including rc5, which was forgotten until now.
10549 In order to let the testing shell script know which algorithms
10550 are available, a new (up to now undocumented) command
10551 "openssl list-cipher-commands" is used.
10554 *) Bugfix: s_client occasionally would sleep in select() when
10555 it should have checked SSL_pending() first.
10558 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10559 the raw DSA values prior to ASN.1 encoding.
10562 *) Tweaks to Configure
10563 [Niels Poppe <niels@netbox.org>]
10565 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10569 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10572 *) New config option to avoid instructions that are illegal on the 80386.
10573 The default code is faster, but requires at least a 486.
10576 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10577 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10578 same as SSL2_VERSION anyway.
10581 *) New "-showcerts" option for s_client.
10584 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10585 application. Various cleanups and fixes.
10588 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10589 modify error routines to work internally. Add error codes and PBE init
10590 to library startup routines.
10593 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10594 packing functions to asn1 and evp. Changed function names and error
10595 codes along the way.
10598 *) PKCS12 integration: and so it begins... First of several patches to
10599 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10600 objects to objects.h
10603 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10604 and display support for Thawte strong extranet extension.
10607 *) Add LinuxPPC support.
10608 [Jeff Dubrule <igor@pobox.org>]
10610 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10611 bn_div_words in alpha.s.
10612 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10614 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10615 OAEP isn't supported when OpenSSL is built with RSAref.
10616 [Ulf Moeller <ulf@fitug.de>]
10618 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10619 so they no longer are missing under -DNOPROTO.
10620 [Soren S. Jorvang <soren@t.dk>]
10623 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10625 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10626 doesn't work when the session is reused. Coming soon!
10629 *) Fix a security hole, that allows sessions to be reused in the wrong
10630 context thus bypassing client cert protection! All software that uses
10631 client certs and session caches in multiple contexts NEEDS PATCHING to
10632 allow session reuse! A fuller solution is in the works.
10633 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10635 *) Some more source tree cleanups (removed obsolete files
10636 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10637 permission on "config" script to be executable) and a fix for the INSTALL
10639 [Ulf Moeller <ulf@fitug.de>]
10641 *) Remove some legacy and erroneous uses of malloc, free instead of
10643 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10645 *) Make rsa_oaep_test return non-zero on error.
10646 [Ulf Moeller <ulf@fitug.de>]
10648 *) Add support for native Solaris shared libraries. Configure
10649 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10650 if someone would make that last step automatic.
10651 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10653 *) ctx_size was not built with the right compiler during "make links". Fixed.
10656 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10657 except NULL ciphers". This means the default cipher list will no longer
10658 enable NULL ciphers. They need to be specifically enabled e.g. with
10659 the string "DEFAULT:eNULL".
10662 *) Fix to RSA private encryption routines: if p < q then it would
10663 occasionally produce an invalid result. This will only happen with
10664 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10667 *) Be less restrictive and allow also `perl util/perlpath.pl
10668 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10669 because this way one can also use an interpreter named `perl5' (which is
10670 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10671 installed as `perl').
10672 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10674 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10675 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10677 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10678 advapi32.lib to Win32 build and change the pem test comparision
10679 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10680 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10681 and crypto/des/ede_cbcm_enc.c.
10684 *) DES quad checksum was broken on big-endian architectures. Fixed.
10687 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10688 Win32 test batch file so it (might) work again. The Win32 test batch file
10689 is horrible: I feel ill....
10692 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10693 in e_os.h. Audit of header files to check ANSI and non ANSI
10694 sections: 10 functions were absent from non ANSI section and not exported
10695 from Windows DLLs. Fixed up libeay.num for new functions.
10698 *) Make `openssl version' output lines consistent.
10699 [Ralf S. Engelschall]
10701 *) Fix Win32 symbol export lists for BIO functions: Added
10702 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10703 to ms/libeay{16,32}.def.
10704 [Ralf S. Engelschall]
10706 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10707 fine under Unix and passes some trivial tests I've now added. But the
10708 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10709 added to make sure no one expects that this stuff really works in the
10710 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10711 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10713 [Ralf S. Engelschall]
10715 *) Fix the generation of two part addresses in perl.
10716 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10718 *) Add config entry for Linux on MIPS.
10719 [John Tobey <jtobey@channel1.com>]
10721 *) Make links whenever Configure is run, unless we are on Windoze.
10724 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10725 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10729 *) Add a useful kludge to allow package maintainers to specify compiler and
10730 other platforms details on the command line without having to patch the
10731 Configure script everytime: One now can use ``perl Configure
10732 <id>:<details>'', i.e. platform ids are allowed to have details appended
10733 to them (seperated by colons). This is treated as there would be a static
10734 pre-configured entry in Configure's %table under key <id> with value
10735 <details> and ``perl Configure <id>'' is called. So, when you want to
10736 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10737 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10738 now, which overrides the FreeBSD-elf entry on-the-fly.
10739 [Ralf S. Engelschall]
10741 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10744 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10745 on the `perl Configure ...' command line. This way one can compile
10746 OpenSSL libraries with Position Independent Code (PIC) which is needed
10747 for linking it into DSOs.
10748 [Ralf S. Engelschall]
10750 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10754 *) Cleaned up the LICENSE document: The official contact for any license
10755 questions now is the OpenSSL core team under openssl-core@openssl.org.
10756 And add a paragraph about the dual-license situation to make sure people
10757 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10758 to the OpenSSL toolkit.
10759 [Ralf S. Engelschall]
10761 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10762 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10763 Additonally cleaned up the `make links' target: Remove unnecessary
10764 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10765 to speed processing and no longer clutter the display with confusing
10766 stuff. Instead only the actually done links are displayed.
10767 [Ralf S. Engelschall]
10769 *) Permit null encryption ciphersuites, used for authentication only. It used
10770 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10771 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10775 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10776 signed attributes when verifying signatures (this would break them),
10777 the detached data encoding was wrong and public keys obtained using
10778 X509_get_pubkey() weren't freed.
10781 *) Add text documentation for the BUFFER functions. Also added a work around
10782 to a Win95 console bug. This was triggered by the password read stuff: the
10783 last character typed gets carried over to the next fread(). If you were
10784 generating a new cert request using 'req' for example then the last
10785 character of the passphrase would be CR which would then enter the first
10789 *) Added the new `Includes OpenSSL Cryptography Software' button as
10790 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10791 button and can be used by applications based on OpenSSL to show the
10792 relationship to the OpenSSL project.
10793 [Ralf S. Engelschall]
10795 *) Remove confusing variables in function signatures in files
10796 ssl/ssl_lib.c and ssl/ssl.h.
10797 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10799 *) Don't install bss_file.c under PREFIX/include/
10800 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10802 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10803 functions that return function pointers and has support for NT specific
10804 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10805 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10806 unsigned to signed types: this was killing the Win32 compile.
10809 *) Add new certificate file to stack functions,
10810 SSL_add_dir_cert_subjects_to_stack() and
10811 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10812 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10813 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10814 This means that Apache-SSL and similar packages don't have to mess around
10815 to add as many CAs as they want to the preferred list.
10818 *) Experiment with doxygen documentation. Currently only partially applied to
10820 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10821 openssl.doxy as the configuration file.
10824 *) Get rid of remaining C++-style comments which strict C compilers hate.
10825 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10827 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10828 compiled in by default: it has problems with large keys.
10831 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10832 DH private keys and/or callback functions which directly correspond to
10833 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10834 is needed for applications which have to configure certificates on a
10835 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10837 For the RSA certificate situation is makes no difference, but
10838 for the DSA certificate situation this fixes the "no shared cipher"
10839 problem where the OpenSSL cipher selection procedure failed because the
10840 temporary keys were not overtaken from the context and the API provided
10841 no way to reconfigure them.
10842 The new functions now let applications reconfigure the stuff and they
10843 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10844 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10845 non-public-API function ssl_cert_instantiate() is used as a helper
10846 function and also to reduce code redundancy inside ssl_rsa.c.
10847 [Ralf S. Engelschall]
10849 *) Move s_server -dcert and -dkey options out of the undocumented feature
10850 area because they are useful for the DSA situation and should be
10851 recognized by the users.
10852 [Ralf S. Engelschall]
10854 *) Fix the cipher decision scheme for export ciphers: the export bits are
10855 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10856 SSL_EXP_MASK. So, the original variable has to be used instead of the
10857 already masked variable.
10858 [Richard Levitte <levitte@stacken.kth.se>]
10860 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10861 [Richard Levitte <levitte@stacken.kth.se>]
10863 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10864 from `int' to `unsigned int' because it's a length and initialized by
10865 EVP_DigestFinal() which expects an `unsigned int *'.
10866 [Richard Levitte <levitte@stacken.kth.se>]
10868 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10869 script. Instead use the usual Shell->Perl transition trick.
10870 [Ralf S. Engelschall]
10872 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10873 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10874 -noout -modulus' as it's already the case for `openssl rsa -noout
10875 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10876 currently the public key is printed (a decision which was already done by
10877 `openssl dsa -modulus' in the past) which serves a similar purpose.
10878 Additionally the NO_RSA no longer completely removes the whole -modulus
10879 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10881 [Ralf S. Engelschall]
10883 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10884 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10885 [Arne Ansper <arne@ats.cyber.ee>]
10887 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10888 to be added. Now both 'req' and 'ca' can use new objects defined in the
10892 *) Add cool BIO that does syslog (or event log on NT).
10893 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10895 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10896 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10897 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10898 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10901 *) Add preliminary config info for new extension code.
10904 *) Make RSA_NO_PADDING really use no padding.
10905 [Ulf Moeller <ulf@fitug.de>]
10907 *) Generate errors when private/public key check is done.
10910 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10911 for some CRL extensions and new objects added.
10914 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10915 key usage extension and fuller support for authority key id.
10918 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10919 padding method for RSA, which is recommended for new applications in PKCS
10920 #1 v2.0 (RFC 2437, October 1998).
10921 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10922 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10923 against Bleichbacher's attack on RSA.
10924 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10927 *) Updates to the new SSL compression code
10928 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10930 *) Fix so that the version number in the master secret, when passed
10931 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10932 (because the server will not accept higher), that the version number
10933 is 0x03,0x01, not 0x03,0x00
10934 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10936 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10937 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10938 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10941 *) Support for RAW extensions where an arbitrary extension can be
10942 created by including its DER encoding. See apps/openssl.cnf for
10946 *) Make sure latest Perl versions don't interpret some generated C array
10947 code as Perl array code in the crypto/err/err_genc.pl script.
10948 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10950 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10951 not many people have the assembler. Various Win32 compilation fixes and
10952 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10953 build instructions.
10956 *) Modify configure script 'Configure' to automatically create crypto/date.h
10957 file under Win32 and also build pem.h from pem.org. New script
10958 util/mkfiles.pl to create the MINFO file on environments that can't do a
10959 'make files': perl util/mkfiles.pl >MINFO should work.
10962 *) Major rework of DES function declarations, in the pursuit of correctness
10963 and purity. As a result, many evil casts evaporated, and some weirdness,
10964 too. You may find this causes warnings in your code. Zapping your evil
10965 casts will probably fix them. Mostly.
10968 *) Fix for a typo in asn1.h. Bug fix to object creation script
10969 obj_dat.pl. It considered a zero in an object definition to mean
10970 "end of object": none of the objects in objects.h have any zeros
10971 so it wasn't spotted.
10972 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10974 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10975 Masking (CBCM). In the absence of test vectors, the best I have been able
10976 to do is check that the decrypt undoes the encrypt, so far. Send me test
10977 vectors if you have them.
10980 *) Correct calculation of key length for export ciphers (too much space was
10981 allocated for null ciphers). This has not been tested!
10984 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10985 message is now correct (it understands "crypto" and "ssl" on its
10986 command line). There is also now an "update" option. This will update
10987 the util/ssleay.num and util/libeay.num files with any new functions.
10989 perl util/mkdef.pl crypto ssl update
10990 it will update them.
10993 *) Overhauled the Perl interface (perl/*):
10994 - ported BN stuff to OpenSSL's different BN library
10995 - made the perl/ source tree CVS-aware
10996 - renamed the package from SSLeay to OpenSSL (the files still contain
10997 their history because I've copied them in the repository)
10998 - removed obsolete files (the test scripts will be replaced
10999 by better Test::Harness variants in the future)
11000 [Ralf S. Engelschall]
11002 *) First cut for a very conservative source tree cleanup:
11003 1. merge various obsolete readme texts into doc/ssleay.txt
11004 where we collect the old documents and readme texts.
11005 2. remove the first part of files where I'm already sure that we no
11006 longer need them because of three reasons: either they are just temporary
11007 files which were left by Eric or they are preserved original files where
11008 I've verified that the diff is also available in the CVS via "cvs diff
11009 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11010 the crypto/md/ stuff).
11011 [Ralf S. Engelschall]
11013 *) More extension code. Incomplete support for subject and issuer alt
11014 name, issuer and authority key id. Change the i2v function parameters
11015 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11016 what that's for :-) Fix to ASN1 macro which messed up
11017 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11020 *) Preliminary support for ENUMERATED type. This is largely copied from the
11024 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11025 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11027 *) Make sure `make rehash' target really finds the `openssl' program.
11028 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11030 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11031 like to hear about it if this slows down other processors.
11034 *) Add CygWin32 platform information to Configure script.
11035 [Alan Batie <batie@aahz.jf.intel.com>]
11037 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11038 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11040 *) New program nseq to manipulate netscape certificate sequences
11043 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11047 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11048 but the BN code had some problems that would cause failures when
11049 doing certificate verification and some other functions.
11050 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11052 *) Add ASN1 and PEM code to support netscape certificate sequences.
11055 *) Add ASN1 and PEM code to support netscape certificate sequences.
11058 *) Add several PKIX and private extended key usage OIDs.
11061 *) Modify the 'ca' program to handle the new extension code. Modify
11062 openssl.cnf for new extension format, add comments.
11065 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11066 and add a sample to openssl.cnf so req -x509 now adds appropriate
11070 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11071 error code, add initial support to X509_print() and x509 application.
11074 *) Takes a deep breath and start addding X509 V3 extension support code. Add
11075 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11076 stuff is currently isolated and isn't even compiled yet.
11079 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11080 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11081 Removed the versions check from X509 routines when loading extensions:
11082 this allows certain broken certificates that don't set the version
11083 properly to be processed.
11086 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11087 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11088 can still be regenerated with "make depend".
11091 *) Spelling mistake in C version of CAST-128.
11092 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11094 *) Changes to the error generation code. The perl script err-code.pl
11095 now reads in the old error codes and retains the old numbers, only
11096 adding new ones if necessary. It also only changes the .err files if new
11097 codes are added. The makefiles have been modified to only insert errors
11098 when needed (to avoid needlessly modifying header files). This is done
11099 by only inserting errors if the .err file is newer than the auto generated
11100 C file. To rebuild all the error codes from scratch (the old behaviour)
11101 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11102 or delete all the .err files.
11105 *) CAST-128 was incorrectly implemented for short keys. The C version has
11106 been fixed, but is untested. The assembler versions are also fixed, but
11107 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11108 to regenerate it if needed.
11109 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11110 Hagino <itojun@kame.net>]
11112 *) File was opened incorrectly in randfile.c.
11113 [Ulf Möller <ulf@fitug.de>]
11115 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11116 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11117 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11118 al: it's just almost always a UTCTime. Note this patch adds new error
11119 codes so do a "make errors" if there are problems.
11122 *) Correct Linux 1 recognition in config.
11123 [Ulf Möller <ulf@fitug.de>]
11125 *) Remove pointless MD5 hash when using DSA keys in ca.
11126 [Anonymous <nobody@replay.com>]
11128 *) Generate an error if given an empty string as a cert directory. Also
11129 generate an error if handed NULL (previously returned 0 to indicate an
11130 error, but didn't set one).
11131 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11133 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11136 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11137 parameters. This was causing a warning which killed off the Win32 compile.
11140 *) Remove C++ style comments from crypto/bn/bn_local.h.
11141 [Neil Costigan <neil.costigan@celocom.com>]
11143 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11144 based on a text string, looking up short and long names and finally
11145 "dot" format. The "dot" format stuff didn't work. Added new function
11146 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11147 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11148 OID is not part of the table.
11151 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11152 X509_LOOKUP_by_alias().
11155 *) Sort openssl functions by name.
11158 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11159 encryption from sample DSA keys (in case anyone is interested the password
11163 *) Make _all_ *_free functions accept a NULL pointer.
11164 [Frans Heymans <fheymans@isaserver.be>]
11166 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11168 [Anonymous <nobody@replay.com>]
11170 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11171 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11173 *) Don't blow it for numeric -newkey arguments to apps/req.
11174 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11176 *) Temp key "for export" tests were wrong in s3_srvr.c.
11177 [Anonymous <nobody@replay.com>]
11179 *) Add prototype for temp key callback functions
11180 SSL_CTX_set_tmp_{rsa,dh}_callback().
11183 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11184 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11187 *) X509_name_add_entry() freed the wrong thing after an error.
11188 [Arne Ansper <arne@ats.cyber.ee>]
11190 *) rsa_eay.c would attempt to free a NULL context.
11191 [Arne Ansper <arne@ats.cyber.ee>]
11193 *) BIO_s_socket() had a broken should_retry() on Windoze.
11194 [Arne Ansper <arne@ats.cyber.ee>]
11196 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11197 [Arne Ansper <arne@ats.cyber.ee>]
11199 *) Make sure the already existing X509_STORE->depth variable is initialized
11200 in X509_STORE_new(), but document the fact that this variable is still
11201 unused in the certificate verification process.
11202 [Ralf S. Engelschall]
11204 *) Fix the various library and apps files to free up pkeys obtained from
11205 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11208 *) Fix reference counting in X509_PUBKEY_get(). This makes
11209 demos/maurice/example2.c work, amongst others, probably.
11210 [Steve Henson and Ben Laurie]
11212 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11213 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11214 are no longer created. This way we have a single and consistent command
11215 line interface `openssl <command>', similar to `cvs <command>'.
11216 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11218 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11219 BIT STRING wrapper always have zero unused bits.
11222 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11225 *) Make the top-level INSTALL documentation easier to understand.
11228 *) Makefiles updated to exit if an error occurs in a sub-directory
11229 make (including if user presses ^C) [Paul Sutton]
11231 *) Make Montgomery context stuff explicit in RSA data structure.
11234 *) Fix build order of pem and err to allow for generated pem.h.
11237 *) Fix renumbering bug in X509_NAME_delete_entry().
11240 *) Enhanced the err-ins.pl script so it makes the error library number
11241 global and can add a library name. This is needed for external ASN1 and
11242 other error libraries.
11245 *) Fixed sk_insert which never worked properly.
11248 *) Fix ASN1 macros so they can handle indefinite length construted
11249 EXPLICIT tags. Some non standard certificates use these: they can now
11253 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11254 into a single doc/ssleay.txt bundle. This way the information is still
11255 preserved but no longer messes up this directory. Now it's new room for
11256 the new set of documenation files.
11257 [Ralf S. Engelschall]
11259 *) SETs were incorrectly DER encoded. This was a major pain, because they
11260 shared code with SEQUENCEs, which aren't coded the same. This means that
11261 almost everything to do with SETs or SEQUENCEs has either changed name or
11262 number of arguments.
11263 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11265 *) Fix test data to work with the above.
11268 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11269 was already fixed by Eric for 0.9.1 it seems.
11270 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11272 *) Autodetect FreeBSD3.
11275 *) Fix various bugs in Configure. This affects the following platforms:
11279 unixware-2.0-pentium
11283 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11284 before they are needed.
11287 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11291 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11293 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11294 changed SSLeay to OpenSSL in version strings.
11295 [Ralf S. Engelschall]
11297 *) Some fixups to the top-level documents.
11300 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11301 because the symlink to include/ was missing.
11302 [Ralf S. Engelschall]
11304 *) Incorporated the popular no-RSA/DSA-only patches
11305 which allow to compile a RSA-free SSLeay.
11306 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11308 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11309 when "ssleay" is still not found.
11310 [Ralf S. Engelschall]
11312 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11313 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11315 *) Updated the README file.
11316 [Ralf S. Engelschall]
11318 *) Added various .cvsignore files in the CVS repository subdirs
11319 to make a "cvs update" really silent.
11320 [Ralf S. Engelschall]
11322 *) Recompiled the error-definition header files and added
11323 missing symbols to the Win32 linker tables.
11324 [Ralf S. Engelschall]
11326 *) Cleaned up the top-level documents;
11327 o new files: CHANGES and LICENSE
11328 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11329 o merged COPYRIGHT into LICENSE
11330 o removed obsolete TODO file
11331 o renamed MICROSOFT to INSTALL.W32
11332 [Ralf S. Engelschall]
11334 *) Removed dummy files from the 0.9.1b source tree:
11335 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11336 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11337 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11338 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11339 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11340 [Ralf S. Engelschall]
11342 *) Added various platform portability fixes.
11345 *) The Genesis of the OpenSSL rpject:
11346 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11347 Young and Tim J. Hudson created while they were working for C2Net until
11349 [The OpenSSL Project]
11352 Changes between 0.9.0b and 0.9.1b [not released]
11354 *) Updated a few CA certificates under certs/
11357 *) Changed some BIGNUM api stuff.
11360 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11361 DGUX x86, Linux Alpha, etc.
11364 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11365 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11369 *) Add -strparse option to asn1pars program which parses nested
11371 [Dr Stephen Henson <shenson@bigfoot.com>]
11373 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11376 *) DSA fix for "ca" program.
11379 *) Added "-genkey" option to "dsaparam" program.
11382 *) Added RIPE MD160 (rmd160) message digest.
11385 *) Added -a (all) option to "ssleay version" command.
11388 *) Added PLATFORM define which is the id given to Configure.
11391 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11394 *) Extended the ASN.1 parser routines.
11397 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11400 *) Added a BN_CTX to the BN library.
11403 *) Fixed the weak key values in DES library
11406 *) Changed API in EVP library for cipher aliases.
11409 *) Added support for RC2/64bit cipher.
11412 *) Converted the lhash library to the crypto/mem.c functions.
11415 *) Added more recognized ASN.1 object ids.
11418 *) Added more RSA padding checks for SSL/TLS.
11421 *) Added BIO proxy/filter functionality.
11424 *) Added extra_certs to SSL_CTX which can be used
11425 send extra CA certificates to the client in the CA cert chain sending
11426 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11429 *) Now Fortezza is denied in the authentication phase because
11430 this is key exchange mechanism is not supported by SSLeay at all.
11433 *) Additional PKCS1 checks.
11436 *) Support the string "TLSv1" for all TLS v1 ciphers.
11439 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11440 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11443 *) Fixed a few memory leaks.
11446 *) Fixed various code and comment typos.
11449 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11450 bytes sent in the client random.
11451 [Edward Bishop <ebishop@spyglass.com>]