5 Changes between 1.0.2k and 1.0.2l [xx XXX xxxx]
9 Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
11 *) Truncated packet could crash via OOB read
13 If one side of an SSL/TLS path is running on a 32-bit host and a specific
14 cipher is being used, then a truncated packet can cause that host to
15 perform an out-of-bounds read, usually resulting in a crash.
17 This issue was reported to OpenSSL by Robert Święcki of Google.
21 *) BN_mod_exp may produce incorrect results on x86_64
23 There is a carry propagating bug in the x86_64 Montgomery squaring
24 procedure. No EC algorithms are affected. Analysis suggests that attacks
25 against RSA and DSA as a result of this defect would be very difficult to
26 perform and are not believed likely. Attacks against DH are considered just
27 feasible (although very difficult) because most of the work necessary to
28 deduce information about a private key may be performed offline. The amount
29 of resources required for such an attack would be very significant and
30 likely only accessible to a limited number of attackers. An attacker would
31 additionally need online access to an unpatched system using the target
32 private key in a scenario with persistent DH parameters and a private
33 key that is shared between multiple clients. For example this can occur by
34 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
35 similar to CVE-2015-3193 but must be treated as a separate problem.
37 This issue was reported to OpenSSL by the OSS-Fuzz project.
41 *) Montgomery multiplication may produce incorrect results
43 There is a carry propagating bug in the Broadwell-specific Montgomery
44 multiplication procedure that handles input lengths divisible by, but
45 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
46 and DH private keys are impossible. This is because the subroutine in
47 question is not used in operations with the private key itself and an input
48 of the attacker's direct choice. Otherwise the bug can manifest itself as
49 transient authentication and key negotiation failures or reproducible
50 erroneous outcome of public-key operations with specially crafted input.
51 Among EC algorithms only Brainpool P-512 curves are affected and one
52 presumably can attack ECDH key negotiation. Impact was not analyzed in
53 detail, because pre-requisites for attack are considered unlikely. Namely
54 multiple clients have to choose the curve in question and the server has to
55 share the private key among them, neither of which is default behaviour.
56 Even then only clients that chose the curve will be affected.
58 This issue was publicly reported as transient failures and was not
59 initially recognized as a security issue. Thanks to Richard Morgan for
60 providing reproducible case.
64 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
65 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
66 prevent issues where no progress is being made and the peer continually
67 sends unrecognised record types, using up resources processing them.
70 Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
72 *) Missing CRL sanity check
74 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
75 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
76 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
78 This issue only affects the OpenSSL 1.0.2i
82 Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
84 *) OCSP Status Request extension unbounded memory growth
86 A malicious client can send an excessively large OCSP Status Request
87 extension. If that client continually requests renegotiation, sending a
88 large OCSP Status Request extension each time, then there will be unbounded
89 memory growth on the server. This will eventually lead to a Denial Of
90 Service attack through memory exhaustion. Servers with a default
91 configuration are vulnerable even if they do not support OCSP. Builds using
92 the "no-ocsp" build time option are not affected.
94 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
98 *) In order to mitigate the SWEET32 attack, the DES ciphers were moved from
101 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
106 *) OOB write in MDC2_Update()
108 An overflow can occur in MDC2_Update() either if called directly or
109 through the EVP_DigestUpdate() function using MDC2. If an attacker
110 is able to supply very large amounts of input data after a previous
111 call to EVP_EncryptUpdate() with a partial block then a length check
112 can overflow resulting in a heap corruption.
114 The amount of data needed is comparable to SIZE_MAX which is impractical
117 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
121 *) Malformed SHA512 ticket DoS
123 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
124 DoS attack where a malformed ticket will result in an OOB read which will
127 The use of SHA512 in TLS session tickets is comparatively rare as it requires
128 a custom server callback and ticket lookup mechanism.
130 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
134 *) OOB write in BN_bn2dec()
136 The function BN_bn2dec() does not check the return value of BN_div_word().
137 This can cause an OOB write if an application uses this function with an
138 overly large BIGNUM. This could be a problem if an overly large certificate
139 or CRL is printed out from an untrusted source. TLS is not affected because
140 record limits will reject an oversized certificate before it is parsed.
142 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
146 *) OOB read in TS_OBJ_print_bio()
148 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
149 the total length the OID text representation would use and not the amount
150 of data written. This will result in OOB reads when large OIDs are
153 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
157 *) Pointer arithmetic undefined behaviour
159 Avoid some undefined pointer arithmetic
161 A common idiom in the codebase is to check limits in the following manner:
164 Where "p" points to some malloc'd data of SIZE bytes and
167 "len" here could be from some externally supplied data (e.g. from a TLS
170 The rules of C pointer arithmetic are such that "p + len" is only well
171 defined where len <= SIZE. Therefore the above idiom is actually
174 For example this could cause problems if some malloc implementation
175 provides an address for "p" such that "p + len" actually overflows for
176 values of len that are too big and therefore p + len < limit.
178 This issue was reported to OpenSSL by Guido Vranken
182 *) Constant time flag not preserved in DSA signing
184 Operations in the DSA signing algorithm should run in constant time in
185 order to avoid side channel attacks. A flaw in the OpenSSL DSA
186 implementation means that a non-constant time codepath is followed for
187 certain operations. This has been demonstrated through a cache-timing
188 attack to be sufficient for an attacker to recover the private DSA key.
190 This issue was reported by César Pereida (Aalto University), Billy Brumley
191 (Tampere University of Technology), and Yuval Yarom (The University of
196 *) DTLS buffered message DoS
198 In a DTLS connection where handshake messages are delivered out-of-order
199 those messages that OpenSSL is not yet ready to process will be buffered
200 for later use. Under certain circumstances, a flaw in the logic means that
201 those messages do not get removed from the buffer even though the handshake
202 has been completed. An attacker could force up to approx. 15 messages to
203 remain in the buffer when they are no longer required. These messages will
204 be cleared when the DTLS connection is closed. The default maximum size for
205 a message is 100k. Therefore the attacker could force an additional 1500k
206 to be consumed per connection. By opening many simulataneous connections an
207 attacker could cause a DoS attack through memory exhaustion.
209 This issue was reported to OpenSSL by Quan Luo.
213 *) DTLS replay protection DoS
215 A flaw in the DTLS replay attack protection mechanism means that records
216 that arrive for future epochs update the replay protection "window" before
217 the MAC for the record has been validated. This could be exploited by an
218 attacker by sending a record for the next epoch (which does not have to
219 decrypt or have a valid MAC), with a very large sequence number. This means
220 that all subsequent legitimate packets are dropped causing a denial of
221 service for a specific DTLS connection.
223 This issue was reported to OpenSSL by the OCAP audit team.
227 *) Certificate message OOB reads
229 In OpenSSL 1.0.2 and earlier some missing message length checks can result
230 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
231 theoretical DoS risk but this has not been observed in practice on common
234 The messages affected are client certificate, client certificate request
235 and server certificate. As a result the attack can only be performed
236 against a client or a server which enables client authentication.
238 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
242 Changes between 1.0.2g and 1.0.2h [3 May 2016]
244 *) Prevent padding oracle in AES-NI CBC MAC check
246 A MITM attacker can use a padding oracle attack to decrypt traffic
247 when the connection uses an AES CBC cipher and the server support
250 This issue was introduced as part of the fix for Lucky 13 padding
251 attack (CVE-2013-0169). The padding check was rewritten to be in
252 constant time by making sure that always the same bytes are read and
253 compared against either the MAC or padding bytes. But it no longer
254 checked that there was enough data to have both the MAC and padding
257 This issue was reported by Juraj Somorovsky using TLS-Attacker.
261 *) Fix EVP_EncodeUpdate overflow
263 An overflow can occur in the EVP_EncodeUpdate() function which is used for
264 Base64 encoding of binary data. If an attacker is able to supply very large
265 amounts of input data then a length check can overflow resulting in a heap
268 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
269 the PEM_write_bio* family of functions. These are mainly used within the
270 OpenSSL command line applications, so any application which processes data
271 from an untrusted source and outputs it as a PEM file should be considered
272 vulnerable to this issue. User applications that call these APIs directly
273 with large amounts of untrusted data may also be vulnerable.
275 This issue was reported by Guido Vranken.
279 *) Fix EVP_EncryptUpdate overflow
281 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
282 is able to supply very large amounts of input data after a previous call to
283 EVP_EncryptUpdate() with a partial block then a length check can overflow
284 resulting in a heap corruption. Following an analysis of all OpenSSL
285 internal usage of the EVP_EncryptUpdate() function all usage is one of two
286 forms. The first form is where the EVP_EncryptUpdate() call is known to be
287 the first called function after an EVP_EncryptInit(), and therefore that
288 specific call must be safe. The second form is where the length passed to
289 EVP_EncryptUpdate() can be seen from the code to be some small value and
290 therefore there is no possibility of an overflow. Since all instances are
291 one of these two forms, it is believed that there can be no overflows in
292 internal code due to this problem. It should be noted that
293 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
294 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
295 of these calls have also been analysed too and it is believed there are no
296 instances in internal usage where an overflow could occur.
298 This issue was reported by Guido Vranken.
302 *) Prevent ASN.1 BIO excessive memory allocation
304 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
305 a short invalid encoding can casuse allocation of large amounts of memory
306 potentially consuming excessive resources or exhausting memory.
308 Any application parsing untrusted data through d2i BIO functions is
309 affected. The memory based functions such as d2i_X509() are *not* affected.
310 Since the memory based functions are used by the TLS library, TLS
311 applications are not affected.
313 This issue was reported by Brian Carpenter.
319 ASN1 Strings that are over 1024 bytes can cause an overread in applications
320 using the X509_NAME_oneline() function on EBCDIC systems. This could result
321 in arbitrary stack data being returned in the buffer.
323 This issue was reported by Guido Vranken.
327 *) Modify behavior of ALPN to invoke callback after SNI/servername
328 callback, such that updates to the SSL_CTX affect ALPN.
331 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
335 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
336 methods are enabled and ssl2 is disabled the methods return NULL.
339 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
341 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
342 Builds that are not configured with "enable-weak-ssl-ciphers" will not
343 provide any "EXPORT" or "LOW" strength ciphers.
346 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
347 is by default disabled at build-time. Builds that are not configured with
348 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
349 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
350 will need to explicitly call either of:
352 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
354 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
356 as appropriate. Even if either of those is used, or the application
357 explicitly uses the version-specific SSLv2_method() or its client and
358 server variants, SSLv2 ciphers vulnerable to exhaustive search key
359 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
360 ciphers, and SSLv2 56-bit DES are no longer available.
364 *) Fix a double-free in DSA code
366 A double free bug was discovered when OpenSSL parses malformed DSA private
367 keys and could lead to a DoS attack or memory corruption for applications
368 that receive DSA private keys from untrusted sources. This scenario is
371 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
376 *) Disable SRP fake user seed to address a server memory leak.
378 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
380 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
381 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
382 was changed to ignore the "fake user" SRP seed, even if the seed
385 Users should use SRP_VBASE_get1_by_user instead. Note that in
386 SRP_VBASE_get1_by_user, caller must free the returned value. Note
387 also that even though configuring the SRP seed attempts to hide
388 invalid usernames by continuing the handshake with fake
389 credentials, this behaviour is not constant time and no strong
390 guarantees are made that the handshake is indistinguishable from
391 that of a valid user.
395 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
397 In the BN_hex2bn function the number of hex digits is calculated using an
398 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
399 large values of |i| this can result in |bn_expand| not allocating any
400 memory because |i * 4| is negative. This can leave the internal BIGNUM data
401 field as NULL leading to a subsequent NULL ptr deref. For very large values
402 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
403 In this case memory is allocated to the internal BIGNUM data field, but it
404 is insufficiently sized leading to heap corruption. A similar issue exists
405 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
406 is ever called by user applications with very large untrusted hex/dec data.
407 This is anticipated to be a rare occurrence.
409 All OpenSSL internal usage of these functions use data that is not expected
410 to be untrusted, e.g. config file data or application command line
411 arguments. If user developed applications generate config file data based
412 on untrusted data then it is possible that this could also lead to security
413 consequences. This is also anticipated to be rare.
415 This issue was reported to OpenSSL by Guido Vranken.
419 *) Fix memory issues in BIO_*printf functions
421 The internal |fmtstr| function used in processing a "%s" format string in
422 the BIO_*printf functions could overflow while calculating the length of a
423 string and cause an OOB read when printing very long strings.
425 Additionally the internal |doapr_outch| function can attempt to write to an
426 OOB memory location (at an offset from the NULL pointer) in the event of a
427 memory allocation failure. In 1.0.2 and below this could be caused where
428 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
429 could be in processing a very long "%s" format string. Memory leaks can
432 The first issue may mask the second issue dependent on compiler behaviour.
433 These problems could enable attacks where large amounts of untrusted data
434 is passed to the BIO_*printf functions. If applications use these functions
435 in this way then they could be vulnerable. OpenSSL itself uses these
436 functions when printing out human-readable dumps of ASN.1 data. Therefore
437 applications that print this data could be vulnerable if the data is from
438 untrusted sources. OpenSSL command line applications could also be
439 vulnerable where they print out ASN.1 data, or if untrusted data is passed
440 as command line arguments.
442 Libssl is not considered directly vulnerable. Additionally certificates etc
443 received via remote connections via libssl are also unlikely to be able to
444 trigger these issues because of message size limits enforced within libssl.
446 This issue was reported to OpenSSL Guido Vranken.
450 *) Side channel attack on modular exponentiation
452 A side-channel attack was found which makes use of cache-bank conflicts on
453 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
454 of RSA keys. The ability to exploit this issue is limited as it relies on
455 an attacker who has control of code in a thread running on the same
456 hyper-threaded core as the victim thread which is performing decryptions.
458 This issue was reported to OpenSSL by Yuval Yarom, The University of
459 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
460 Nadia Heninger, University of Pennsylvania with more information at
461 http://cachebleed.info.
465 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
466 if no keysize is specified with default_bits. This fixes an
467 omission in an earlier change that changed all RSA/DSA key generation
468 apps to use 2048 bits by default.
471 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
473 *) DH small subgroups
475 Historically OpenSSL only ever generated DH parameters based on "safe"
476 primes. More recently (in version 1.0.2) support was provided for
477 generating X9.42 style parameter files such as those required for RFC 5114
478 support. The primes used in such files may not be "safe". Where an
479 application is using DH configured with parameters based on primes that are
480 not "safe" then an attacker could use this fact to find a peer's private
481 DH exponent. This attack requires that the attacker complete multiple
482 handshakes in which the peer uses the same private DH exponent. For example
483 this could be used to discover a TLS server's private DH exponent if it's
484 reusing the private DH exponent or it's using a static DH ciphersuite.
486 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
487 TLS. It is not on by default. If the option is not set then the server
488 reuses the same private DH exponent for the life of the server process and
489 would be vulnerable to this attack. It is believed that many popular
490 applications do set this option and would therefore not be at risk.
492 The fix for this issue adds an additional check where a "q" parameter is
493 available (as is the case in X9.42 based parameters). This detects the
494 only known attack, and is the only possible defense for static DH
495 ciphersuites. This could have some performance impact.
497 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
498 default and cannot be disabled. This could have some performance impact.
500 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
504 *) SSLv2 doesn't block disabled ciphers
506 A malicious client can negotiate SSLv2 ciphers that have been disabled on
507 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
508 been disabled, provided that the SSLv2 protocol was not also disabled via
511 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
512 and Sebastian Schinzel.
516 *) Reject DH handshakes with parameters shorter than 1024 bits.
519 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
521 *) BN_mod_exp may produce incorrect results on x86_64
523 There is a carry propagating bug in the x86_64 Montgomery squaring
524 procedure. No EC algorithms are affected. Analysis suggests that attacks
525 against RSA and DSA as a result of this defect would be very difficult to
526 perform and are not believed likely. Attacks against DH are considered just
527 feasible (although very difficult) because most of the work necessary to
528 deduce information about a private key may be performed offline. The amount
529 of resources required for such an attack would be very significant and
530 likely only accessible to a limited number of attackers. An attacker would
531 additionally need online access to an unpatched system using the target
532 private key in a scenario with persistent DH parameters and a private
533 key that is shared between multiple clients. For example this can occur by
534 default in OpenSSL DHE based SSL/TLS ciphersuites.
536 This issue was reported to OpenSSL by Hanno Böck.
540 *) Certificate verify crash with missing PSS parameter
542 The signature verification routines will crash with a NULL pointer
543 dereference if presented with an ASN.1 signature using the RSA PSS
544 algorithm and absent mask generation function parameter. Since these
545 routines are used to verify certificate signature algorithms this can be
546 used to crash any certificate verification operation and exploited in a
547 DoS attack. Any application which performs certificate verification is
548 vulnerable including OpenSSL clients and servers which enable client
551 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
555 *) X509_ATTRIBUTE memory leak
557 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
558 memory. This structure is used by the PKCS#7 and CMS routines so any
559 application which reads PKCS#7 or CMS data from untrusted sources is
560 affected. SSL/TLS is not affected.
562 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
567 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
568 This changes the decoding behaviour for some invalid messages,
569 though the change is mostly in the more lenient direction, and
570 legacy behaviour is preserved as much as possible.
573 *) In DSA_generate_parameters_ex, if the provided seed is too short,
574 use a random seed, as already documented.
575 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
577 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
579 *) Alternate chains certificate forgery
581 During certificate verfification, OpenSSL will attempt to find an
582 alternative certificate chain if the first attempt to build such a chain
583 fails. An error in the implementation of this logic can mean that an
584 attacker could cause certain checks on untrusted certificates to be
585 bypassed, such as the CA flag, enabling them to use a valid leaf
586 certificate to act as a CA and "issue" an invalid certificate.
588 This issue was reported to OpenSSL by Adam Langley/David Benjamin
593 *) Race condition handling PSK identify hint
595 If PSK identity hints are received by a multi-threaded client then
596 the values are wrongly updated in the parent SSL_CTX structure. This can
597 result in a race condition potentially leading to a double free of the
602 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
604 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
605 incompatibility in the handling of HMAC. The previous ABI has now been
608 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
610 *) Malformed ECParameters causes infinite loop
612 When processing an ECParameters structure OpenSSL enters an infinite loop
613 if the curve specified is over a specially malformed binary polynomial
616 This can be used to perform denial of service against any
617 system which processes public keys, certificate requests or
618 certificates. This includes TLS clients and TLS servers with
619 client authentication enabled.
621 This issue was reported to OpenSSL by Joseph Barr-Pixton.
625 *) Exploitable out-of-bounds read in X509_cmp_time
627 X509_cmp_time does not properly check the length of the ASN1_TIME
628 string and can read a few bytes out of bounds. In addition,
629 X509_cmp_time accepts an arbitrary number of fractional seconds in the
632 An attacker can use this to craft malformed certificates and CRLs of
633 various sizes and potentially cause a segmentation fault, resulting in
634 a DoS on applications that verify certificates or CRLs. TLS clients
635 that verify CRLs are affected. TLS clients and servers with client
636 authentication enabled may be affected if they use custom verification
639 This issue was reported to OpenSSL by Robert Swiecki (Google), and
640 independently by Hanno Böck.
644 *) PKCS7 crash with missing EnvelopedContent
646 The PKCS#7 parsing code does not handle missing inner EncryptedContent
647 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
648 with missing content and trigger a NULL pointer dereference on parsing.
650 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
651 structures from untrusted sources are affected. OpenSSL clients and
652 servers are not affected.
654 This issue was reported to OpenSSL by Michal Zalewski (Google).
658 *) CMS verify infinite loop with unknown hash function
660 When verifying a signedData message the CMS code can enter an infinite loop
661 if presented with an unknown hash function OID. This can be used to perform
662 denial of service against any system which verifies signedData messages using
664 This issue was reported to OpenSSL by Johannes Bauer.
668 *) Race condition handling NewSessionTicket
670 If a NewSessionTicket is received by a multi-threaded client when attempting to
671 reuse a previous ticket then a race condition can occur potentially leading to
672 a double free of the ticket data.
676 *) Removed support for the two export grade static DH ciphersuites
677 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
678 were newly added (along with a number of other static DH ciphersuites) to
679 1.0.2. However the two export ones have *never* worked since they were
680 introduced. It seems strange in any case to be adding new export
681 ciphersuites, and given "logjam" it also does not seem correct to fix them.
684 *) Only support 256-bit or stronger elliptic curves with the
685 'ecdh_auto' setting (server) or by default (client). Of supported
686 curves, prefer P-256 (both).
689 *) Reject DH handshakes with parameters shorter than 768 bits.
690 [Kurt Roeckx and Emilia Kasper]
692 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
694 *) ClientHello sigalgs DoS fix
696 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
697 invalid signature algorithms extension a NULL pointer dereference will
698 occur. This can be exploited in a DoS attack against the server.
700 This issue was was reported to OpenSSL by David Ramos of Stanford
703 [Stephen Henson and Matt Caswell]
705 *) Multiblock corrupted pointer fix
707 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
708 feature only applies on 64 bit x86 architecture platforms that support AES
709 NI instructions. A defect in the implementation of "multiblock" can cause
710 OpenSSL's internal write buffer to become incorrectly set to NULL when
711 using non-blocking IO. Typically, when the user application is using a
712 socket BIO for writing, this will only result in a failed connection.
713 However if some other BIO is used then it is likely that a segmentation
714 fault will be triggered, thus enabling a potential DoS attack.
716 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
720 *) Segmentation fault in DTLSv1_listen fix
722 The DTLSv1_listen function is intended to be stateless and processes the
723 initial ClientHello from many peers. It is common for user code to loop
724 over the call to DTLSv1_listen until a valid ClientHello is received with
725 an associated cookie. A defect in the implementation of DTLSv1_listen means
726 that state is preserved in the SSL object from one invocation to the next
727 that can lead to a segmentation fault. Errors processing the initial
728 ClientHello can trigger this scenario. An example of such an error could be
729 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
732 This issue was reported to OpenSSL by Per Allansson.
736 *) Segmentation fault in ASN1_TYPE_cmp fix
738 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
739 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
740 certificate signature algorithm consistency this can be used to crash any
741 certificate verification operation and exploited in a DoS attack. Any
742 application which performs certificate verification is vulnerable including
743 OpenSSL clients and servers which enable client authentication.
747 *) Segmentation fault for invalid PSS parameters fix
749 The signature verification routines will crash with a NULL pointer
750 dereference if presented with an ASN.1 signature using the RSA PSS
751 algorithm and invalid parameters. Since these routines are used to verify
752 certificate signature algorithms this can be used to crash any
753 certificate verification operation and exploited in a DoS attack. Any
754 application which performs certificate verification is vulnerable including
755 OpenSSL clients and servers which enable client authentication.
757 This issue was was reported to OpenSSL by Brian Carpenter.
761 *) ASN.1 structure reuse memory corruption fix
763 Reusing a structure in ASN.1 parsing may allow an attacker to cause
764 memory corruption via an invalid write. Such reuse is and has been
765 strongly discouraged and is believed to be rare.
767 Applications that parse structures containing CHOICE or ANY DEFINED BY
768 components may be affected. Certificate parsing (d2i_X509 and related
769 functions) are however not affected. OpenSSL clients and servers are
774 *) PKCS7 NULL pointer dereferences fix
776 The PKCS#7 parsing code does not handle missing outer ContentInfo
777 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
778 missing content and trigger a NULL pointer dereference on parsing.
780 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
781 otherwise parse PKCS#7 structures from untrusted sources are
782 affected. OpenSSL clients and servers are not affected.
784 This issue was reported to OpenSSL by Michal Zalewski (Google).
788 *) DoS via reachable assert in SSLv2 servers fix
790 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
791 servers that both support SSLv2 and enable export cipher suites by sending
792 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
794 This issue was discovered by Sean Burford (Google) and Emilia Käsper
795 (OpenSSL development team).
799 *) Empty CKE with client auth and DHE fix
801 If client auth is used then a server can seg fault in the event of a DHE
802 ciphersuite being selected and a zero length ClientKeyExchange message
803 being sent by the client. This could be exploited in a DoS attack.
807 *) Handshake with unseeded PRNG fix
809 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
810 with an unseeded PRNG. The conditions are:
811 - The client is on a platform where the PRNG has not been seeded
812 automatically, and the user has not seeded manually
813 - A protocol specific client method version has been used (i.e. not
814 SSL_client_methodv23)
815 - A ciphersuite is used that does not require additional random data from
816 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
818 If the handshake succeeds then the client random that has been used will
819 have been generated from a PRNG with insufficient entropy and therefore the
820 output may be predictable.
822 For example using the following command with an unseeded openssl will
823 succeed on an unpatched platform:
825 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
829 *) Use After Free following d2i_ECPrivatekey error fix
831 A malformed EC private key file consumed via the d2i_ECPrivateKey function
832 could cause a use after free condition. This, in turn, could cause a double
833 free in several private key parsing functions (such as d2i_PrivateKey
834 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
835 for applications that receive EC private keys from untrusted
836 sources. This scenario is considered rare.
838 This issue was discovered by the BoringSSL project and fixed in their
843 *) X509_to_X509_REQ NULL pointer deref fix
845 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
846 the certificate key is invalid. This function is rarely used in practice.
848 This issue was discovered by Brian Carpenter.
852 *) Removed the export ciphers from the DEFAULT ciphers
855 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
857 *) Change RSA and DH/DSA key generation apps to generate 2048-bit
861 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
862 ARMv5 through ARMv8, as opposite to "locking" it to single one.
863 So far those who have to target multiple plaforms would compromise
864 and argue that binary targeting say ARMv5 would still execute on
865 ARMv8. "Universal" build resolves this compromise by providing
866 near-optimal performance even on newer platforms.
869 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
870 (other platforms pending).
871 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
873 *) Add support for the SignedCertificateTimestampList certificate and
874 OCSP response extensions from RFC6962.
877 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
878 for corner cases. (Certain input points at infinity could lead to
879 bogus results, with non-infinity inputs mapped to infinity too.)
882 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
883 This covers AES, SHA256/512 and GHASH. "Initial" means that most
884 common cases are optimized and there still is room for further
885 improvements. Vector Permutation AES for Altivec is also added.
888 *) Add support for little-endian ppc64 Linux target.
889 [Marcelo Cerri (IBM)]
891 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
892 SHA1, SHA256 and GHASH. "Initial" means that most common cases
893 are optimized and there still is room for further improvements.
894 Both 32- and 64-bit modes are supported.
895 [Andy Polyakov, Ard Biesheuvel (Linaro)]
897 *) Improved ARMv7 NEON support.
900 *) Support for SPARC Architecture 2011 crypto extensions, first
901 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
902 SHA256/512, MD5, GHASH and modular exponentiation.
903 [Andy Polyakov, David Miller]
905 *) Accelerated modular exponentiation for Intel processors, a.k.a.
907 [Shay Gueron & Vlad Krasnov (Intel Corp)]
909 *) Support for new and upcoming Intel processors, including AVX2,
910 BMI and SHA ISA extensions. This includes additional "stitched"
911 implementations, AESNI-SHA256 and GCM, and multi-buffer support
914 This work was sponsored by Intel Corp.
917 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
918 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
919 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
922 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
923 this fixes a limiation in previous versions of OpenSSL.
926 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
927 MGF1 digest and OAEP label.
930 *) Add EVP support for key wrapping algorithms, to avoid problems with
931 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
932 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
933 algorithms and include tests cases.
936 *) Add functions to allocate and set the fields of an ECDSA_METHOD
938 [Douglas E. Engert, Steve Henson]
940 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
941 difference in days and seconds between two tm or ASN1_TIME structures.
944 *) Add -rev test option to s_server to just reverse order of characters
945 received by client and send back to server. Also prints an abbreviated
946 summary of the connection parameters.
949 *) New option -brief for s_client and s_server to print out a brief summary
950 of connection parameters.
953 *) Add callbacks for arbitrary TLS extensions.
954 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
956 *) New option -crl_download in several openssl utilities to download CRLs
957 from CRLDP extension in certificates.
960 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
963 *) New function X509_CRL_diff to generate a delta CRL from the difference
964 of two full CRLs. Add support to "crl" utility.
967 *) New functions to set lookup_crls function and to retrieve
968 X509_STORE from X509_STORE_CTX.
971 *) Print out deprecated issuer and subject unique ID fields in
975 *) Extend OCSP I/O functions so they can be used for simple general purpose
976 HTTP as well as OCSP. New wrapper function which can be used to download
977 CRLs using the OCSP API.
980 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
983 *) SSL_CONF* functions. These provide a common framework for application
984 configuration using configuration files or command lines.
987 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
988 message callback and prints the results. Needs compile time option
989 "enable-ssl-trace". New options to s_client and s_server to enable
993 *) New ctrl and macro to retrieve supported points extensions.
994 Print out extension in s_server and s_client.
997 *) New functions to retrieve certificate signature and signature
1001 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1005 *) New Suite B modes for TLS code. These use and enforce the requirements
1006 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1007 only use Suite B curves. The Suite B modes can be set by using the
1008 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1011 *) New chain verification flags for Suite B levels of security. Check
1012 algorithms are acceptable when flags are set in X509_verify_cert.
1015 *) Make tls1_check_chain return a set of flags indicating checks passed
1016 by a certificate chain. Add additional tests to handle client
1017 certificates: checks for matching certificate type and issuer name
1021 *) If an attempt is made to use a signature algorithm not in the peer
1022 preference list abort the handshake. If client has no suitable
1023 signature algorithms in response to a certificate request do not
1024 use the certificate.
1027 *) If server EC tmp key is not in client preference list abort handshake.
1030 *) Add support for certificate stores in CERT structure. This makes it
1031 possible to have different stores per SSL structure or one store in
1032 the parent SSL_CTX. Include distint stores for certificate chain
1033 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1034 to build and store a certificate chain in CERT structure: returing
1035 an error if the chain cannot be built: this will allow applications
1036 to test if a chain is correctly configured.
1038 Note: if the CERT based stores are not set then the parent SSL_CTX
1039 store is used to retain compatibility with existing behaviour.
1043 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1044 mask based on the current session, check mask when sending client
1045 hello and checking the requested ciphersuite.
1048 *) New ctrls to retrieve and set certificate types in a certificate
1049 request message. Print out received values in s_client. If certificate
1050 types is not set with custom values set sensible values based on
1051 supported signature algorithms.
1054 *) Support for distinct client and server supported signature algorithms.
1057 *) Add certificate callback. If set this is called whenever a certificate
1058 is required by client or server. An application can decide which
1059 certificate chain to present based on arbitrary criteria: for example
1060 supported signature algorithms. Add very simple example to s_server.
1061 This fixes many of the problems and restrictions of the existing client
1062 certificate callback: for example you can now clear an existing
1063 certificate and specify the whole chain.
1066 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1067 the certificate can be used for (if anything). Set valid_flags field
1068 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1069 to have similar checks in it.
1071 Add new "cert_flags" field to CERT structure and include a "strict mode".
1072 This enforces some TLS certificate requirements (such as only permitting
1073 certificate signature algorithms contained in the supported algorithms
1074 extension) which some implementations ignore: this option should be used
1075 with caution as it could cause interoperability issues.
1078 *) Update and tidy signature algorithm extension processing. Work out
1079 shared signature algorithms based on preferences and peer algorithms
1080 and print them out in s_client and s_server. Abort handshake if no
1081 shared signature algorithms.
1084 *) Add new functions to allow customised supported signature algorithms
1085 for SSL and SSL_CTX structures. Add options to s_client and s_server
1089 *) New function SSL_certs_clear() to delete all references to certificates
1090 from an SSL structure. Before this once a certificate had been added
1091 it couldn't be removed.
1094 *) Integrate hostname, email address and IP address checking with certificate
1095 verification. New verify options supporting checking in opensl utility.
1098 *) Fixes and wildcard matching support to hostname and email checking
1099 functions. Add manual page.
1100 [Florian Weimer (Red Hat Product Security Team)]
1102 *) New functions to check a hostname email or IP address against a
1103 certificate. Add options x509 utility to print results of checks against
1107 *) Fix OCSP checking.
1108 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1110 *) Initial experimental support for explicitly trusted non-root CAs.
1111 OpenSSL still tries to build a complete chain to a root but if an
1112 intermediate CA has a trust setting included that is used. The first
1113 setting is used: whether to trust (e.g., -addtrust option to the x509
1117 *) Add -trusted_first option which attempts to find certificates in the
1118 trusted store even if an untrusted chain is also supplied.
1121 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1122 platform support for Linux and Android.
1125 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1128 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1129 When in FIPS mode the approved implementations are used as normal,
1130 when not in FIPS mode the internal unapproved versions are used instead.
1131 This means that the FIPS capable OpenSSL isn't forced to use the
1132 (often lower perfomance) FIPS implementations outside FIPS mode.
1135 *) Transparently support X9.42 DH parameters when calling
1136 PEM_read_bio_DHparameters. This means existing applications can handle
1137 the new parameter format automatically.
1140 *) Initial experimental support for X9.42 DH parameter format: mainly
1141 to support use of 'q' parameter for RFC5114 parameters.
1144 *) Add DH parameters from RFC5114 including test data to dhtest.
1147 *) Support for automatic EC temporary key parameter selection. If enabled
1148 the most preferred EC parameters are automatically used instead of
1149 hardcoded fixed parameters. Now a server just has to call:
1150 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1151 support ECDH and use the most appropriate parameters.
1154 *) Enhance and tidy EC curve and point format TLS extension code. Use
1155 static structures instead of allocation if default values are used.
1156 New ctrls to set curves we wish to support and to retrieve shared curves.
1157 Print out shared curves in s_server. New options to s_server and s_client
1158 to set list of supported curves.
1161 *) New ctrls to retrieve supported signature algorithms and
1162 supported curve values as an array of NIDs. Extend openssl utility
1163 to print out received values.
1166 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1167 between NIDs and the more common NIST names such as "P-256". Enhance
1168 ecparam utility and ECC method to recognise the NIST names for curves.
1171 *) Enhance SSL/TLS certificate chain handling to support different
1172 chains for each certificate instead of one chain in the parent SSL_CTX.
1175 *) Support for fixed DH ciphersuite client authentication: where both
1176 server and client use DH certificates with common parameters.
1179 *) Support for fixed DH ciphersuites: those requiring DH server
1183 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1185 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1186 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1187 X509_CINF_get_signature were reverted post internal team review.
1189 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1191 *) Build fixes for the Windows and OpenVMS platforms
1192 [Matt Caswell and Richard Levitte]
1194 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1196 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1197 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1198 dereference. This could lead to a Denial Of Service attack. Thanks to
1199 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1203 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1204 dtls1_buffer_record function under certain conditions. In particular this
1205 could occur if an attacker sent repeated DTLS records with the same
1206 sequence number but for the next epoch. The memory leak could be exploited
1207 by an attacker in a Denial of Service attack through memory exhaustion.
1208 Thanks to Chris Mueller for reporting this issue.
1212 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1213 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1214 method would be set to NULL which could later result in a NULL pointer
1215 dereference. Thanks to Frank Schmirler for reporting this issue.
1219 *) Abort handshake if server key exchange message is omitted for ephemeral
1222 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1223 reporting this issue.
1227 *) Remove non-export ephemeral RSA code on client and server. This code
1228 violated the TLS standard by allowing the use of temporary RSA keys in
1229 non-export ciphersuites and could be used by a server to effectively
1230 downgrade the RSA key length used to a value smaller than the server
1231 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1232 INRIA or reporting this issue.
1236 *) Fixed issue where DH client certificates are accepted without verification.
1237 An OpenSSL server will accept a DH certificate for client authentication
1238 without the certificate verify message. This effectively allows a client to
1239 authenticate without the use of a private key. This only affects servers
1240 which trust a client certificate authority which issues certificates
1241 containing DH keys: these are extremely rare and hardly ever encountered.
1242 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1247 *) Ensure that the session ID context of an SSL is updated when its
1248 SSL_CTX is updated via SSL_set_SSL_CTX.
1250 The session ID context is typically set from the parent SSL_CTX,
1251 and can vary with the CTX.
1254 *) Fix various certificate fingerprint issues.
1256 By using non-DER or invalid encodings outside the signed portion of a
1257 certificate the fingerprint can be changed without breaking the signature.
1258 Although no details of the signed portion of the certificate can be changed
1259 this can cause problems with some applications: e.g. those using the
1260 certificate fingerprint for blacklists.
1262 1. Reject signatures with non zero unused bits.
1264 If the BIT STRING containing the signature has non zero unused bits reject
1265 the signature. All current signature algorithms require zero unused bits.
1267 2. Check certificate algorithm consistency.
1269 Check the AlgorithmIdentifier inside TBS matches the one in the
1270 certificate signature. NB: this will result in signature failure
1271 errors for some broken certificates.
1273 Thanks to Konrad Kraszewski from Google for reporting this issue.
1275 3. Check DSA/ECDSA signatures use DER.
1277 Reencode DSA/ECDSA signatures and compare with the original received
1278 signature. Return an error if there is a mismatch.
1280 This will reject various cases including garbage after signature
1281 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1282 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1283 (negative or with leading zeroes).
1285 Further analysis was conducted and fixes were developed by Stephen Henson
1286 of the OpenSSL core team.
1291 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1292 results on some platforms, including x86_64. This bug occurs at random
1293 with a very low probability, and is not known to be exploitable in any
1294 way, though its exact impact is difficult to determine. Thanks to Pieter
1295 Wuille (Blockstream) who reported this issue and also suggested an initial
1296 fix. Further analysis was conducted by the OpenSSL development team and
1297 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1298 the OpenSSL core team.
1302 *) Do not resume sessions on the server if the negotiated protocol
1303 version does not match the session's version. Resuming with a different
1304 version, while not strictly forbidden by the RFC, is of questionable
1305 sanity and breaks all known clients.
1306 [David Benjamin, Emilia Käsper]
1308 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1309 early CCS messages during renegotiation. (Note that because
1310 renegotiation is encrypted, this early CCS was not exploitable.)
1313 *) Tighten client-side session ticket handling during renegotiation:
1314 ensure that the client only accepts a session ticket if the server sends
1315 the extension anew in the ServerHello. Previously, a TLS client would
1316 reuse the old extension state and thus accept a session ticket if one was
1317 announced in the initial ServerHello.
1319 Similarly, ensure that the client requires a session ticket if one
1320 was advertised in the ServerHello. Previously, a TLS client would
1321 ignore a missing NewSessionTicket message.
1324 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1326 *) SRTP Memory Leak.
1328 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1329 sends a carefully crafted handshake message, to cause OpenSSL to fail
1330 to free up to 64k of memory causing a memory leak. This could be
1331 exploited in a Denial Of Service attack. This issue affects OpenSSL
1332 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1333 whether SRTP is used or configured. Implementations of OpenSSL that
1334 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1336 The fix was developed by the OpenSSL team.
1340 *) Session Ticket Memory Leak.
1342 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1343 integrity of that ticket is first verified. In the event of a session
1344 ticket integrity check failing, OpenSSL will fail to free memory
1345 causing a memory leak. By sending a large number of invalid session
1346 tickets an attacker could exploit this issue in a Denial Of Service
1351 *) Build option no-ssl3 is incomplete.
1353 When OpenSSL is configured with "no-ssl3" as a build option, servers
1354 could accept and complete a SSL 3.0 handshake, and clients could be
1355 configured to send them.
1357 [Akamai and the OpenSSL team]
1359 *) Add support for TLS_FALLBACK_SCSV.
1360 Client applications doing fallback retries should call
1361 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1363 [Adam Langley, Bodo Moeller]
1365 *) Add additional DigestInfo checks.
1367 Reencode DigestInto in DER and check against the original when
1368 verifying RSA signature: this will reject any improperly encoded
1369 DigestInfo structures.
1371 Note: this is a precautionary measure and no attacks are currently known.
1375 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1377 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1378 SRP code can be overrun an internal buffer. Add sanity check that
1379 g, A, B < N to SRP code.
1381 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1382 Group for discovering this issue.
1386 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1387 TLS 1.0 instead of higher protocol versions when the ClientHello message
1388 is badly fragmented. This allows a man-in-the-middle attacker to force a
1389 downgrade to TLS 1.0 even if both the server and the client support a
1390 higher protocol version, by modifying the client's TLS records.
1392 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1393 researching this issue.
1397 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1398 to a denial of service attack. A malicious server can crash the client
1399 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1400 ciphersuite and sending carefully crafted handshake messages.
1402 Thanks to Felix Gröbert (Google) for discovering and researching this
1407 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1408 to leak memory. This can be exploited through a Denial of Service attack.
1409 Thanks to Adam Langley for discovering and researching this issue.
1413 *) An attacker can force openssl to consume large amounts of memory whilst
1414 processing DTLS handshake messages. This can be exploited through a
1415 Denial of Service attack.
1416 Thanks to Adam Langley for discovering and researching this issue.
1420 *) An attacker can force an error condition which causes openssl to crash
1421 whilst processing DTLS packets due to memory being freed twice. This
1422 can be exploited through a Denial of Service attack.
1423 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1428 *) If a multithreaded client connects to a malicious server using a resumed
1429 session and the server sends an ec point format extension it could write
1430 up to 255 bytes to freed memory.
1432 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1437 *) A malicious server can crash an OpenSSL client with a null pointer
1438 dereference (read) by specifying an SRP ciphersuite even though it was not
1439 properly negotiated with the client. This can be exploited through a
1440 Denial of Service attack.
1442 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1443 discovering and researching this issue.
1447 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1448 X509_name_oneline, X509_name_print_ex et al. to leak some information
1449 from the stack. Applications may be affected if they echo pretty printing
1450 output to the attacker.
1452 Thanks to Ivan Fratric (Google) for discovering this issue.
1454 [Emilia Käsper, and Steve Henson]
1456 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1457 for corner cases. (Certain input points at infinity could lead to
1458 bogus results, with non-infinity inputs mapped to infinity too.)
1461 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1463 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1464 handshake can force the use of weak keying material in OpenSSL
1465 SSL/TLS clients and servers.
1467 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1468 researching this issue. (CVE-2014-0224)
1469 [KIKUCHI Masashi, Steve Henson]
1471 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1472 OpenSSL DTLS client the code can be made to recurse eventually crashing
1475 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1477 [Imre Rad, Steve Henson]
1479 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1480 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1481 client or server. This is potentially exploitable to run arbitrary
1482 code on a vulnerable client or server.
1484 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1485 [Jüri Aedla, Steve Henson]
1487 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1488 are subject to a denial of service attack.
1490 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1491 this issue. (CVE-2014-3470)
1492 [Felix Gröbert, Ivan Fratric, Steve Henson]
1494 *) Harmonize version and its documentation. -f flag is used to display
1496 [mancha <mancha1@zoho.com>]
1498 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1499 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1500 [mancha <mancha1@zoho.com>]
1502 *) Fix some double frees. These are not thought to be exploitable.
1503 [mancha <mancha1@zoho.com>]
1505 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1507 *) A missing bounds check in the handling of the TLS heartbeat extension
1508 can be used to reveal up to 64k of memory to a connected client or
1511 Thanks for Neel Mehta of Google Security for discovering this bug and to
1512 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1513 preparing the fix (CVE-2014-0160)
1514 [Adam Langley, Bodo Moeller]
1516 *) Fix for the attack described in the paper "Recovering OpenSSL
1517 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1518 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1519 http://eprint.iacr.org/2014/140
1521 Thanks to Yuval Yarom and Naomi Benger for discovering this
1522 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1523 [Yuval Yarom and Naomi Benger]
1525 *) TLS pad extension: draft-agl-tls-padding-03
1527 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1528 TLS client Hello record length value would otherwise be > 255 and
1529 less that 512 pad with a dummy extension containing zeroes so it
1530 is at least 512 bytes long.
1532 [Adam Langley, Steve Henson]
1534 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1536 *) Fix for TLS record tampering bug. A carefully crafted invalid
1537 handshake could crash OpenSSL with a NULL pointer exception.
1538 Thanks to Anton Johansson for reporting this issues.
1541 *) Keep original DTLS digest and encryption contexts in retransmission
1542 structures so we can use the previous session parameters if they need
1543 to be resent. (CVE-2013-6450)
1546 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1547 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1548 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1549 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1550 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1551 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1552 [Rob Stradling, Adam Langley]
1554 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1556 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1557 supporting platforms or when small records were transferred.
1558 [Andy Polyakov, Steve Henson]
1560 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1562 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1564 This addresses the flaw in CBC record processing discovered by
1565 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1566 at: http://www.isg.rhul.ac.uk/tls/
1568 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1569 Security Group at Royal Holloway, University of London
1570 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1571 Emilia Käsper for the initial patch.
1573 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1575 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1576 ciphersuites which can be exploited in a denial of service attack.
1577 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1578 and detecting this bug and to Wolfgang Ettlinger
1579 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1583 *) Return an error when checking OCSP signatures when key is NULL.
1584 This fixes a DoS attack. (CVE-2013-0166)
1587 *) Make openssl verify return errors.
1588 [Chris Palmer <palmer@google.com> and Ben Laurie]
1590 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1591 the right response is stapled. Also change SSL_get_certificate()
1592 so it returns the certificate actually sent.
1593 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1594 [Rob Stradling <rob.stradling@comodo.com>]
1596 *) Fix possible deadlock when decoding public keys.
1599 *) Don't use TLS 1.0 record version number in initial client hello
1603 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1605 *) Sanity check record length before skipping explicit IV in TLS
1606 1.2, 1.1 and DTLS to fix DoS attack.
1608 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1609 fuzzing as a service testing platform.
1613 *) Initialise tkeylen properly when encrypting CMS messages.
1614 Thanks to Solar Designer of Openwall for reporting this issue.
1617 *) In FIPS mode don't try to use composite ciphers as they are not
1621 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1623 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1624 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1625 mean any application compiled against OpenSSL 1.0.0 headers setting
1626 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1627 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1628 0x10000000L Any application which was previously compiled against
1629 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1630 will need to be recompiled as a result. Letting be results in
1631 inability to disable specifically TLS 1.1 and in client context,
1632 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1635 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1636 disable just protocol X, but all protocols above X *if* there are
1637 protocols *below* X still enabled. In more practical terms it means
1638 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1639 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1640 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1644 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1646 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1647 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1648 in CRYPTO_realloc_clean.
1650 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1651 issue and to Adam Langley <agl@chromium.org> for fixing it.
1653 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1655 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1658 *) Workarounds for some broken servers that "hang" if a client hello
1659 record length exceeds 255 bytes.
1661 1. Do not use record version number > TLS 1.0 in initial client
1662 hello: some (but not all) hanging servers will now work.
1663 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1664 the number of ciphers sent in the client hello. This should be
1665 set to an even number, such as 50, for example by passing:
1666 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1667 Most broken servers should now work.
1668 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1669 TLS 1.2 client support entirely.
1672 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1675 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1677 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1678 STRING form instead of a DigestInfo.
1681 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1682 and the RSA_sign/RSA_verify functions. This was made more apparent when
1683 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1684 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1685 the correct format in RSA_verify so both forms transparently work.
1688 *) Some servers which support TLS 1.0 can choke if we initially indicate
1689 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1690 encrypted premaster secret. As a workaround use the maximum pemitted
1691 client version in client hello, this should keep such servers happy
1692 and still work with previous versions of OpenSSL.
1695 *) Add support for TLS/DTLS heartbeats.
1696 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1698 *) Add support for SCTP.
1699 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1701 *) Improved PRNG seeding for VOS.
1702 [Paul Green <Paul.Green@stratus.com>]
1704 *) Extensive assembler packs updates, most notably:
1706 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1707 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1708 - x86_64: bit-sliced AES implementation;
1709 - ARM: NEON support, contemporary platforms optimizations;
1710 - s390x: z196 support;
1711 - *: GHASH and GF(2^m) multiplication implementations;
1715 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1716 (removal of unnecessary code)
1717 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1719 *) Add TLS key material exporter from RFC 5705.
1722 *) Add DTLS-SRTP negotiation from RFC 5764.
1725 *) Add Next Protocol Negotiation,
1726 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1727 disabled with a no-npn flag to config or Configure. Code donated
1729 [Adam Langley <agl@google.com> and Ben Laurie]
1731 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1732 NIST-P256, NIST-P521, with constant-time single point multiplication on
1733 typical inputs. Compiler support for the nonstandard type __uint128_t is
1734 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1735 Code made available under Apache License version 2.0.
1737 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1738 line to include this in your build of OpenSSL, and run "make depend" (or
1739 "make update"). This enables the following EC_METHODs:
1741 EC_GFp_nistp224_method()
1742 EC_GFp_nistp256_method()
1743 EC_GFp_nistp521_method()
1745 EC_GROUP_new_by_curve_name() will automatically use these (while
1746 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1748 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1750 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1751 all platforms. Move ssize_t definition from e_os.h to the public
1752 header file e_os2.h as it now appears in public header file cms.h
1755 *) New -sigopt option to the ca, req and x509 utilities. Additional
1756 signature parameters can be passed using this option and in
1760 *) Add RSA PSS signing function. This will generate and set the
1761 appropriate AlgorithmIdentifiers for PSS based on those in the
1762 corresponding EVP_MD_CTX structure. No application support yet.
1765 *) Support for companion algorithm specific ASN1 signing routines.
1766 New function ASN1_item_sign_ctx() signs a pre-initialised
1767 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1768 the appropriate parameters.
1771 *) Add new algorithm specific ASN1 verification initialisation function
1772 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1773 handling will be the same no matter what EVP_PKEY_METHOD is used.
1774 Add a PSS handler to support verification of PSS signatures: checked
1775 against a number of sample certificates.
1778 *) Add signature printing for PSS. Add PSS OIDs.
1779 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1781 *) Add algorithm specific signature printing. An individual ASN1 method
1782 can now print out signatures instead of the standard hex dump.
1784 More complex signatures (e.g. PSS) can print out more meaningful
1785 information. Include DSA version that prints out the signature
1789 *) Password based recipient info support for CMS library: implementing
1793 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1794 neatly separates the code into cipher and PBE sections and is required
1795 for some algorithms that split PBES2 into separate pieces (such as
1796 password based CMS).
1799 *) Session-handling fixes:
1800 - Fix handling of connections that are resuming with a session ID,
1801 but also support Session Tickets.
1802 - Fix a bug that suppressed issuing of a new ticket if the client
1803 presented a ticket with an expired session.
1804 - Try to set the ticket lifetime hint to something reasonable.
1805 - Make tickets shorter by excluding irrelevant information.
1806 - On the client side, don't ignore renewed tickets.
1807 [Adam Langley, Bodo Moeller (Google)]
1809 *) Fix PSK session representation.
1812 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1814 This work was sponsored by Intel.
1817 *) Add GCM support to TLS library. Some custom code is needed to split
1818 the IV between the fixed (from PRF) and explicit (from TLS record)
1819 portions. This adds all GCM ciphersuites supported by RFC5288 and
1820 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
1821 add a special AESGCM string for GCM only.
1824 *) Expand range of ctrls for AES GCM. Permit setting invocation
1825 field on decrypt and retrieval of invocation field only on encrypt.
1828 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1829 As required by RFC5289 these ciphersuites cannot be used if for
1830 versions of TLS earlier than 1.2.
1833 *) For FIPS capable OpenSSL interpret a NULL default public key method
1834 as unset and return the appopriate default but do *not* set the default.
1835 This means we can return the appopriate method in applications that
1836 swicth between FIPS and non-FIPS modes.
1839 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1840 ENGINE is used then we cannot handle that in the FIPS module so we
1841 keep original code iff non-FIPS operations are allowed.
1844 *) Add -attime option to openssl utilities.
1845 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1847 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1850 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1851 FIPS EC methods unconditionally for now.
1854 *) New build option no-ec2m to disable characteristic 2 code.
1857 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1858 all cases can be covered as some introduce binary incompatibilities.
1861 *) Redirect RSA operations to FIPS module including keygen,
1862 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1865 *) Add similar low level API blocking to ciphers.
1868 *) Low level digest APIs are not approved in FIPS mode: any attempt
1869 to use these will cause a fatal error. Applications that *really* want
1870 to use them can use the private_* version instead.
1873 *) Redirect cipher operations to FIPS module for FIPS builds.
1876 *) Redirect digest operations to FIPS module for FIPS builds.
1879 *) Update build system to add "fips" flag which will link in fipscanister.o
1880 for static and shared library builds embedding a signature if needed.
1883 *) Output TLS supported curves in preference order instead of numerical
1884 order. This is currently hardcoded for the highest order curves first.
1885 This should be configurable so applications can judge speed vs strength.
1888 *) Add TLS v1.2 server support for client authentication.
1891 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1895 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1896 FIPS modules versions.
1899 *) Add TLS v1.2 client side support for client authentication. Keep cache
1900 of handshake records longer as we don't know the hash algorithm to use
1901 until after the certificate request message is received.
1904 *) Initial TLS v1.2 client support. Add a default signature algorithms
1905 extension including all the algorithms we support. Parse new signature
1906 format in client key exchange. Relax some ECC signing restrictions for
1907 TLS v1.2 as indicated in RFC5246.
1910 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1911 to new signature format when needed using client digest preference.
1912 All server ciphersuites should now work correctly in TLS v1.2. No client
1913 support yet and no support for client certificates.
1916 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1917 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1918 ciphersuites. At present only RSA key exchange ciphersuites work with
1919 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1920 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1921 and version checking.
1924 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1925 with this defined it will not be affected by any changes to ssl internal
1926 structures. Add several utility functions to allow openssl application
1927 to work with OPENSSL_NO_SSL_INTERN defined.
1931 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1933 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1936 *) Permit abbreviated handshakes when renegotiating using the function
1937 SSL_renegotiate_abbreviated().
1938 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1940 *) Add call to ENGINE_register_all_complete() to
1941 ENGINE_load_builtin_engines(), so some implementations get used
1942 automatically instead of needing explicit application support.
1945 *) Add support for TLS key exporter as described in RFC5705.
1946 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1948 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1949 a few changes are required:
1951 Add SSL_OP_NO_TLSv1_1 flag.
1952 Add TLSv1_1 methods.
1953 Update version checking logic to handle version 1.1.
1954 Add explicit IV handling (ported from DTLS code).
1955 Add command line options to s_client/s_server.
1958 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1960 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1961 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1962 content decryption and always return the same error. Note: this attack
1963 needs on average 2^20 messages so it only affects automated senders. The
1964 old behaviour can be reenabled in the CMS code by setting the
1965 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1966 an MMA defence is not necessary.
1967 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1968 this issue. (CVE-2012-0884)
1971 *) Fix CVE-2011-4619: make sure we really are receiving a
1972 client hello before rejecting multiple SGC restarts. Thanks to
1973 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1976 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1978 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1979 Thanks to Antonio Martin, Enterprise Secure Access Research and
1980 Development, Cisco Systems, Inc. for discovering this bug and
1981 preparing a fix. (CVE-2012-0050)
1984 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1986 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1987 of the Vaudenay padding oracle attack on CBC mode encryption
1988 which enables an efficient plaintext recovery attack against
1989 the OpenSSL implementation of DTLS. Their attack exploits timing
1990 differences arising during decryption processing. A research
1991 paper describing this attack can be found at:
1992 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1993 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1994 Security Group at Royal Holloway, University of London
1995 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1996 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1997 for preparing the fix. (CVE-2011-4108)
1998 [Robin Seggelmann, Michael Tuexen]
2000 *) Clear bytes used for block padding of SSL 3.0 records.
2002 [Adam Langley (Google)]
2004 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2005 Kadianakis <desnacked@gmail.com> for discovering this issue and
2006 Adam Langley for preparing the fix. (CVE-2011-4619)
2007 [Adam Langley (Google)]
2009 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2010 [Andrey Kulikov <amdeich@gmail.com>]
2012 *) Prevent malformed RFC3779 data triggering an assertion failure.
2013 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2014 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2015 [Rob Austein <sra@hactrn.net>]
2017 *) Improved PRNG seeding for VOS.
2018 [Paul Green <Paul.Green@stratus.com>]
2020 *) Fix ssl_ciph.c set-up race.
2021 [Adam Langley (Google)]
2023 *) Fix spurious failures in ecdsatest.c.
2024 [Emilia Käsper (Google)]
2026 *) Fix the BIO_f_buffer() implementation (which was mixing different
2027 interpretations of the '..._len' fields).
2028 [Adam Langley (Google)]
2030 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2031 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2032 threads won't reuse the same blinding coefficients.
2034 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2035 lock to call BN_BLINDING_invert_ex, and avoids one use of
2036 BN_BLINDING_update for each BN_BLINDING structure (previously,
2037 the last update always remained unused).
2038 [Emilia Käsper (Google)]
2040 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2041 [Bob Buckholz (Google)]
2043 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2045 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2046 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2047 [Kaspar Brand <ossl@velox.ch>]
2049 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2050 for multi-threaded use of ECDH. (CVE-2011-3210)
2051 [Adam Langley (Google)]
2053 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2056 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2057 signature public key algorithm by using OID xref utilities instead.
2058 Before this you could only use some ECC ciphersuites with SHA1 only.
2061 *) Add protection against ECDSA timing attacks as mentioned in the paper
2062 by Billy Bob Brumley and Nicola Tuveri, see:
2064 http://eprint.iacr.org/2011/232.pdf
2066 [Billy Bob Brumley and Nicola Tuveri]
2068 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2070 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2071 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2073 *) Fix bug in string printing code: if *any* escaping is enabled we must
2074 escape the escape character (backslash) or the resulting string is
2078 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2080 *) Disable code workaround for ancient and obsolete Netscape browsers
2081 and servers: an attacker can use it in a ciphersuite downgrade attack.
2082 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2085 *) Fixed J-PAKE implementation error, originally discovered by
2086 Sebastien Martini, further info and confirmation from Stefan
2087 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2090 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2092 *) Fix extension code to avoid race conditions which can result in a buffer
2093 overrun vulnerability: resumed sessions must not be modified as they can
2094 be shared by multiple threads. CVE-2010-3864
2097 *) Fix WIN32 build system to correctly link an ENGINE directory into
2101 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2103 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2105 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2107 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2109 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2110 context. The operation can be customised via the ctrl mechanism in
2111 case ENGINEs want to include additional functionality.
2114 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2117 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2118 output hashes compatible with older versions of OpenSSL.
2119 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2121 *) Fix compression algorithm handling: if resuming a session use the
2122 compression algorithm of the resumed session instead of determining
2123 it from client hello again. Don't allow server to change algorithm.
2126 *) Add load_crls() function to apps tidying load_certs() too. Add option
2127 to verify utility to allow additional CRLs to be included.
2130 *) Update OCSP request code to permit adding custom headers to the request:
2131 some responders need this.
2134 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2136 [Julia Lawall <julia@diku.dk>]
2138 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2139 needlessly dereferenced structures, used obsolete functions and
2140 didn't handle all updated verify codes correctly.
2143 *) Disable MD2 in the default configuration.
2146 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2147 indicate the initial BIO being pushed or popped. This makes it possible
2148 to determine whether the BIO is the one explicitly called or as a result
2149 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2150 it handles reference counts correctly and doesn't zero out the I/O bio
2151 when it is not being explicitly popped. WARNING: applications which
2152 included workarounds for the old buggy behaviour will need to be modified
2153 or they could free up already freed BIOs.
2156 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2157 renaming to all platforms (within the 0.9.8 branch, this was
2158 done conditionally on Netware platforms to avoid a name clash).
2159 [Guenter <lists@gknw.net>]
2161 *) Add ECDHE and PSK support to DTLS.
2162 [Michael Tuexen <tuexen@fh-muenster.de>]
2164 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2168 *) Add "missing" function EVP_MD_flags() (without this the only way to
2169 retrieve a digest flags is by accessing the structure directly. Update
2170 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2171 or cipher is registered as in the "from" argument. Print out all
2172 registered digests in the dgst usage message instead of manually
2173 attempting to work them out.
2176 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2177 this allows the use of compression and extensions. Change default cipher
2178 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2179 by default unless an application cipher string requests it.
2182 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2183 key ids to find matching certificates and keys but some PKCS#12 files
2184 don't follow the (somewhat unwritten) rules and this strategy fails.
2185 Now just gather all certificates together and the first private key
2186 then look for the first certificate that matches the key.
2189 *) Support use of registered digest and cipher names for dgst and cipher
2190 commands instead of having to add each one as a special case. So now
2197 openssl dgst -sha256 foo
2199 and this works for ENGINE based algorithms too.
2203 *) Update Gost ENGINE to support parameter files.
2204 [Victor B. Wagner <vitus@cryptocom.ru>]
2206 *) Support GeneralizedTime in ca utility.
2207 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2209 *) Enhance the hash format used for certificate directory links. The new
2210 form uses the canonical encoding (meaning equivalent names will work
2211 even if they aren't identical) and uses SHA1 instead of MD5. This form
2212 is incompatible with the older format and as a result c_rehash should
2213 be used to rebuild symbolic links.
2216 *) Make PKCS#8 the default write format for private keys, replacing the
2217 traditional format. This form is standardised, more secure and doesn't
2218 include an implicit MD5 dependency.
2221 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2222 committed to OpenSSL should pass this lot as a minimum.
2225 *) Add session ticket override functionality for use by EAP-FAST.
2226 [Jouni Malinen <j@w1.fi>]
2228 *) Modify HMAC functions to return a value. Since these can be implemented
2229 in an ENGINE errors can occur.
2232 *) Type-checked OBJ_bsearch_ex.
2235 *) Type-checked OBJ_bsearch. Also some constification necessitated
2236 by type-checking. Still to come: TXT_DB, bsearch(?),
2237 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2241 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2242 seconds to a tm structure directly, instead of going through OS
2243 specific date routines. This avoids any issues with OS routines such
2244 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2245 and X509_time_adj_ex() to cover the extended range. The existing
2246 X509_time_adj() is still usable and will no longer have any date issues.
2249 *) Delta CRL support. New use deltas option which will attempt to locate
2250 and search any appropriate delta CRLs available.
2252 This work was sponsored by Google.
2255 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2256 code and add additional score elements. Validate alternate CRL paths
2257 as part of the CRL checking and indicate a new error "CRL path validation
2258 error" in this case. Applications wanting additional details can use
2259 the verify callback and check the new "parent" field. If this is not
2260 NULL CRL path validation is taking place. Existing applications wont
2261 see this because it requires extended CRL support which is off by
2264 This work was sponsored by Google.
2267 *) Support for freshest CRL extension.
2269 This work was sponsored by Google.
2272 *) Initial indirect CRL support. Currently only supported in the CRLs
2273 passed directly and not via lookup. Process certificate issuer
2274 CRL entry extension and lookup CRL entries by bother issuer name
2275 and serial number. Check and process CRL issuer entry in IDP extension.
2277 This work was sponsored by Google.
2280 *) Add support for distinct certificate and CRL paths. The CRL issuer
2281 certificate is validated separately in this case. Only enabled if
2282 an extended CRL support flag is set: this flag will enable additional
2283 CRL functionality in future.
2285 This work was sponsored by Google.
2288 *) Add support for policy mappings extension.
2290 This work was sponsored by Google.
2293 *) Fixes to pathlength constraint, self issued certificate handling,
2294 policy processing to align with RFC3280 and PKITS tests.
2296 This work was sponsored by Google.
2299 *) Support for name constraints certificate extension. DN, email, DNS
2300 and URI types are currently supported.
2302 This work was sponsored by Google.
2305 *) To cater for systems that provide a pointer-based thread ID rather
2306 than numeric, deprecate the current numeric thread ID mechanism and
2307 replace it with a structure and associated callback type. This
2308 mechanism allows a numeric "hash" to be extracted from a thread ID in
2309 either case, and on platforms where pointers are larger than 'long',
2310 mixing is done to help ensure the numeric 'hash' is usable even if it
2311 can't be guaranteed unique. The default mechanism is to use "&errno"
2312 as a pointer-based thread ID to distinguish between threads.
2314 Applications that want to provide their own thread IDs should now use
2315 CRYPTO_THREADID_set_callback() to register a callback that will call
2316 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2318 Note that ERR_remove_state() is now deprecated, because it is tied
2319 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2320 to free the current thread's error state should be replaced by
2321 ERR_remove_thread_state(NULL).
2323 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2324 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2325 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2326 application was previously providing a numeric thread callback that
2327 was inappropriate for distinguishing threads, then uniqueness might
2328 have been obtained with &errno that happened immediately in the
2329 intermediate development versions of OpenSSL; this is no longer the
2330 case, the numeric thread callback will now override the automatic use
2332 [Geoff Thorpe, with help from Bodo Moeller]
2334 *) Initial support for different CRL issuing certificates. This covers a
2335 simple case where the self issued certificates in the chain exist and
2336 the real CRL issuer is higher in the existing chain.
2338 This work was sponsored by Google.
2341 *) Removed effectively defunct crypto/store from the build.
2344 *) Revamp of STACK to provide stronger type-checking. Still to come:
2345 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2346 ASN1_STRING, CONF_VALUE.
2349 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2350 RAM on SSL connections. This option can save about 34k per idle SSL.
2353 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2354 STACK, TXT_DB, bsearch, qsort.
2357 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2358 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2359 support for data, signedData, compressedData, digestedData and
2360 encryptedData, envelopedData types included. Scripts to check against
2361 RFC4134 examples draft and interop and consistency checks of many
2362 content types and variants.
2365 *) Add options to enc utility to support use of zlib compression BIO.
2368 *) Extend mk1mf to support importing of options and assembly language
2369 files from Configure script, currently only included in VC-WIN32.
2370 The assembly language rules can now optionally generate the source
2371 files from the associated perl scripts.
2374 *) Implement remaining functionality needed to support GOST ciphersuites.
2375 Interop testing has been performed using CryptoPro implementations.
2376 [Victor B. Wagner <vitus@cryptocom.ru>]
2378 *) s390x assembler pack.
2381 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2385 *) Implement Opaque PRF Input TLS extension as specified in
2386 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2387 official specification yet and no extension type assignment by
2388 IANA exists, this extension (for now) will have to be explicitly
2389 enabled when building OpenSSL by providing the extension number
2390 to use. For example, specify an option
2392 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2394 to the "config" or "Configure" script to enable the extension,
2395 assuming extension number 0x9527 (which is a completely arbitrary
2396 and unofficial assignment based on the MD5 hash of the Internet
2397 Draft). Note that by doing so, you potentially lose
2398 interoperability with other TLS implementations since these might
2399 be using the same extension number for other purposes.
2401 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2402 opaque PRF input value to use in the handshake. This will create
2403 an interal copy of the length-'len' string at 'src', and will
2404 return non-zero for success.
2406 To get more control and flexibility, provide a callback function
2409 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2410 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2414 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2417 Callback function 'cb' will be called in handshakes, and is
2418 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2419 Argument 'arg' is for application purposes (the value as given to
2420 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2421 be provided to the callback function). The callback function
2422 has to return non-zero to report success: usually 1 to use opaque
2423 PRF input just if possible, or 2 to enforce use of the opaque PRF
2424 input. In the latter case, the library will abort the handshake
2425 if opaque PRF input is not successfully negotiated.
2427 Arguments 'peerinput' and 'len' given to the callback function
2428 will always be NULL and 0 in the case of a client. A server will
2429 see the client's opaque PRF input through these variables if
2430 available (NULL and 0 otherwise). Note that if the server
2431 provides an opaque PRF input, the length must be the same as the
2432 length of the client's opaque PRF input.
2434 Note that the callback function will only be called when creating
2435 a new session (session resumption can resume whatever was
2436 previously negotiated), and will not be called in SSL 2.0
2437 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2438 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2439 for applications that need to enforce opaque PRF input.
2443 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2446 [Victor B. Wagner <vitus@cryptocom.ru>]
2448 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2449 RFC4507bis. The encrypted ticket format is an encrypted encoded
2450 SSL_SESSION structure, that way new session features are automatically
2453 If a client application caches session in an SSL_SESSION structure
2454 support is transparent because tickets are now stored in the encoded
2457 The SSL_CTX structure automatically generates keys for ticket
2458 protection in servers so again support should be possible
2459 with no application modification.
2461 If a client or server wishes to disable RFC4507 support then the option
2462 SSL_OP_NO_TICKET can be set.
2464 Add a TLS extension debugging callback to allow the contents of any client
2465 or server extensions to be examined.
2467 This work was sponsored by Google.
2470 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2471 OpenSSL should now compile cleanly on gcc 4.2
2472 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2474 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2475 support including streaming MAC support: this is required for GOST
2476 ciphersuite support.
2477 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2479 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2480 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2481 to output in BER and PEM format.
2484 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2485 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2486 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2487 ENGINE support for HMAC keys which are unextractable. New -mac and
2488 -macopt options to dgst utility.
2491 *) New option -sigopt to dgst utility. Update dgst to use
2492 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2493 alternative signing paramaters such as X9.31 or PSS in the dgst
2497 *) Change ssl_cipher_apply_rule(), the internal function that does
2498 the work each time a ciphersuite string requests enabling
2499 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2500 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2501 the order of disabled ciphersuites such that those ciphersuites
2502 that most recently went from enabled to disabled not only stay
2503 in order with respect to each other, but also have higher priority
2504 than other disabled ciphersuites the next time ciphersuites are
2507 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2508 the same ciphersuites as with "HIGH" alone, but in a specific
2509 order where the PSK ciphersuites come first (since they are the
2510 most recently disabled ciphersuites when "HIGH" is parsed).
2512 Also, change ssl_create_cipher_list() (using this new
2513 funcionality) such that between otherwise identical
2514 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2518 *) Change ssl_create_cipher_list() so that it automatically
2519 arranges the ciphersuites in reasonable order before starting
2520 to process the rule string. Thus, the definition for "DEFAULT"
2521 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2522 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2523 This makes it much easier to arrive at a reasonable default order
2524 in applications for which anonymous ciphers are OK (meaning
2525 that you can't actually use DEFAULT).
2526 [Bodo Moeller; suggested by Victor Duchovni]
2528 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2529 processing) into multiple integers instead of setting
2530 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2531 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2532 (These masks as well as the individual bit definitions are hidden
2533 away into the non-exported interface ssl/ssl_locl.h, so this
2534 change to the definition of the SSL_CIPHER structure shouldn't
2535 affect applications.) This give us more bits for each of these
2536 categories, so there is no longer a need to coagulate AES128 and
2537 AES256 into a single algorithm bit, and to coagulate Camellia128
2538 and Camellia256 into a single algorithm bit, which has led to all
2541 Thus, among other things, the kludge introduced in 0.9.7m and
2542 0.9.8e for masking out AES256 independently of AES128 or masking
2543 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2545 With the change, we also introduce new ciphersuite aliases that
2546 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2550 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2551 Use the leftmost N bytes of the signature input if the input is
2552 larger than the prime q (with N being the size in bytes of q).
2555 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2556 it yet and it is largely untested.
2559 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2562 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2563 some compilers (gcc 4.2 and later) reject their use. Safestack is
2564 reimplemented. Update ASN1 to avoid use of legacy functions.
2567 *) Win32/64 targets are linked with Winsock2.
2570 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2571 to external functions. This can be used to increase CRL handling
2572 efficiency especially when CRLs are very large by (for example) storing
2573 the CRL revoked certificates in a database.
2576 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2577 new CRLs added to a directory can be used. New command line option
2578 -verify_return_error to s_client and s_server. This causes real errors
2579 to be returned by the verify callback instead of carrying on no matter
2580 what. This reflects the way a "real world" verify callback would behave.
2583 *) GOST engine, supporting several GOST algorithms and public key formats.
2584 Kindly donated by Cryptocom.
2587 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2588 partitioned by DP are handled but no indirect CRL or reason partitioning
2589 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2590 selected via a scoring technique which handles IDP and AKID in CRLs.
2593 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2594 will ultimately be used for all verify operations: this will remove the
2595 X509_STORE dependency on certificate verification and allow alternative
2596 lookup methods. X509_STORE based implementations of these two callbacks.
2599 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2600 Modify get_crl() to find a valid (unexpired) CRL if possible.
2603 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2604 this would be called X509_CRL_cmp() but that name is already used by
2605 a function that just compares CRL issuer names. Cache several CRL
2606 extensions in X509_CRL structure and cache CRLDP in X509.
2609 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2610 this maps equivalent X509_NAME structures into a consistent structure.
2611 Name comparison can then be performed rapidly using memcmp().
2614 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2618 *) Allow digests to supply their own micalg string for S/MIME type using
2619 the ctrl EVP_MD_CTRL_MICALG.
2622 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2623 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2624 ctrl. It can then customise the structure before and/or after signing
2628 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2629 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2630 to free up any added signature OIDs.
2633 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2634 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2635 digest and cipher tables. New options added to openssl utility:
2636 list-message-digest-algorithms and list-cipher-algorithms.
2639 *) Change the array representation of binary polynomials: the list
2640 of degrees of non-zero coefficients is now terminated with -1.
2641 Previously it was terminated with 0, which was also part of the
2642 value; thus, the array representation was not applicable to
2643 polynomials where t^0 has coefficient zero. This change makes
2644 the array representation useful in a more general context.
2647 *) Various modifications and fixes to SSL/TLS cipher string
2648 handling. For ECC, the code now distinguishes between fixed ECDH
2649 with RSA certificates on the one hand and with ECDSA certificates
2650 on the other hand, since these are separate ciphersuites. The
2651 unused code for Fortezza ciphersuites has been removed.
2653 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2654 (not "ECDHE"). For consistency with the code for DH
2655 certificates, use of ECDH certificates is now considered ECDH
2656 authentication, not RSA or ECDSA authentication (the latter is
2657 merely the CA's signing algorithm and not actively used in the
2660 The temporary ciphersuite alias "ECCdraft" is no longer
2661 available, and ECC ciphersuites are no longer excluded from "ALL"
2662 and "DEFAULT". The following aliases now exist for RFC 4492
2663 ciphersuites, most of these by analogy with the DH case:
2665 kECDHr - ECDH cert, signed with RSA
2666 kECDHe - ECDH cert, signed with ECDSA
2667 kECDH - ECDH cert (signed with either RSA or ECDSA)
2668 kEECDH - ephemeral ECDH
2669 ECDH - ECDH cert or ephemeral ECDH
2675 AECDH - anonymous ECDH
2676 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2680 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2681 Use correct micalg parameters depending on digest(s) in signed message.
2684 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2685 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2688 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2689 an engine to register a method. Add ENGINE lookups for methods and
2690 functional reference processing.
2693 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2694 EVP_{Sign,Verify}* which allow an application to customise the signature
2698 *) New -resign option to smime utility. This adds one or more signers
2699 to an existing PKCS#7 signedData structure. Also -md option to use an
2700 alternative message digest algorithm for signing.
2703 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2704 create PKCS7 structures containing multiple signers. Update smime
2705 application to support multiple signers.
2708 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2712 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2713 Reorganize PBE internals to lookup from a static table using NIDs,
2714 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2715 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2716 PRF which will be automatically used with PBES2.
2719 *) Replace the algorithm specific calls to generate keys in "req" with the
2723 *) Update PKCS#7 enveloped data routines to use new API. This is now
2724 supported by any public key method supporting the encrypt operation. A
2725 ctrl is added to allow the public key algorithm to examine or modify
2726 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2730 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2731 a default digest type to use. In most cases this will be SHA1 but some
2732 algorithms (such as GOST) need to specify an alternative digest. The
2733 return value indicates how strong the prefernce is 1 means optional and
2734 2 is mandatory (that is it is the only supported type). Modify
2735 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2736 use the default md. Update openssl utilities to use the default digest
2737 type for signing if it is not explicitly indicated.
2740 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2741 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2742 signing method from the key type. This effectively removes the link
2743 between digests and public key types.
2746 *) Add an OID cross reference table and utility functions. Its purpose is to
2747 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2748 rsaEncryption. This will allow some of the algorithm specific hackery
2749 needed to use the correct OID to be removed.
2752 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2753 structures for PKCS7_sign(). They are now set up by the relevant public
2757 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2760 *) Add support for key derivation (agreement) in the API, DH method and
2764 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2765 public and private key formats. As a side effect these add additional
2766 command line functionality not previously available: DSA signatures can be
2767 generated and verified using pkeyutl and DH key support and generation in
2772 [Oliver Tappe <zooey@hirschkaefer.de>]
2774 *) New make target "install_html_docs" installs HTML renditions of the
2776 [Oliver Tappe <zooey@hirschkaefer.de>]
2778 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
2779 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2780 support key and parameter generation and add initial key generation
2781 functionality for RSA.
2784 *) Add functions for main EVP_PKEY_method operations. The undocumented
2785 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2786 EVP_PKEY_{encrypt,decrypt}_old.
2789 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2790 key API, doesn't do much yet.
2793 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2794 public key algorithms. New option to openssl utility:
2795 "list-public-key-algorithms" to print out info.
2798 *) Implement the Supported Elliptic Curves Extension for
2799 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2802 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2803 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2806 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2807 utilities such as rsa, dsa, dsaparam etc except they process any key
2811 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2812 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2813 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2817 *) Initial support for pluggable public key ASN1.
2818 De-spaghettify the public key ASN1 handling. Move public and private
2819 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2820 algorithm specific handling to a single module within the relevant
2821 algorithm directory. Add functions to allow (near) opaque processing
2822 of public and private key structures.
2825 *) Implement the Supported Point Formats Extension for
2826 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2829 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2830 for the psk identity [hint] and the psk callback functions to the
2831 SSL_SESSION, SSL and SSL_CTX structure.
2834 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2838 SSL_CTX_use_psk_identity_hint
2839 SSL_get_psk_identity_hint
2840 SSL_get_psk_identity
2841 SSL_use_psk_identity_hint
2843 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2845 *) Add RFC 3161 compliant time stamp request creation, response generation
2846 and response verification functionality.
2847 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2849 *) Add initial support for TLS extensions, specifically for the server_name
2850 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2851 have new members for a host name. The SSL data structure has an
2852 additional member SSL_CTX *initial_ctx so that new sessions can be
2853 stored in that context to allow for session resumption, even after the
2854 SSL has been switched to a new SSL_CTX in reaction to a client's
2855 server_name extension.
2857 New functions (subject to change):
2859 SSL_get_servername()
2860 SSL_get_servername_type()
2863 New CTRL codes and macros (subject to change):
2865 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2866 - SSL_CTX_set_tlsext_servername_callback()
2867 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2868 - SSL_CTX_set_tlsext_servername_arg()
2869 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2871 openssl s_client has a new '-servername ...' option.
2873 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2874 '-key2 ...', '-servername_fatal' (subject to change). This allows
2875 testing the HostName extension for a specific single host name ('-cert'
2876 and '-key' remain fallbacks for handshakes without HostName
2877 negotiation). If the unrecogninzed_name alert has to be sent, this by
2878 default is a warning; it becomes fatal with the '-servername_fatal'
2881 [Peter Sylvester, Remy Allais, Christophe Renou]
2883 *) Whirlpool hash implementation is added.
2886 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2887 bn(64,32). Because of instruction set limitations it doesn't have
2888 any negative impact on performance. This was done mostly in order
2889 to make it possible to share assembler modules, such as bn_mul_mont
2890 implementations, between 32- and 64-bit builds without hassle.
2893 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2894 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2898 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2899 dedicated Montgomery multiplication procedure, is introduced.
2900 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2901 "64-bit" performance on certain 32-bit targets.
2904 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2905 in SSL structures. New SSL ctrl to set maximum send fragment size.
2906 Save memory by seeting the I/O buffer sizes dynamically instead of
2907 using the maximum available value.
2910 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2911 in addition to the text details.
2914 *) Very, very preliminary EXPERIMENTAL support for printing of general
2915 ASN1 structures. This currently produces rather ugly output and doesn't
2916 handle several customised structures at all.
2919 *) Integrated support for PVK file format and some related formats such
2920 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2921 these in the 'rsa' and 'dsa' utilities.
2924 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2927 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2928 place for the (very old) "NETSCAPE" format certificates which are now
2929 handled using new ASN1 code equivalents.
2932 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2933 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2934 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2937 *) Modify CRL distribution points extension code to print out previously
2938 unsupported fields. Enhance extension setting code to allow setting of
2942 *) Add print and set support for Issuing Distribution Point CRL extension.
2945 *) Change 'Configure' script to enable Camellia by default.
2948 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2950 *) When rejecting SSL/TLS records due to an incorrect version number, never
2951 update s->server with a new major version number. As of
2952 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2953 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2954 the previous behavior could result in a read attempt at NULL when
2955 receiving specific incorrect SSL/TLS records once record payload
2956 protection is active. (CVE-2010-0740)
2957 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2959 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2960 could be crashed if the relevant tables were not present (e.g. chrooted).
2961 [Tomas Hoger <thoger@redhat.com>]
2963 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2965 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2966 [Martin Olsson, Neel Mehta]
2968 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2969 accommodate for stack sorting, always a write lock!).
2972 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2973 excessive delays in the RAND_poll(): over a minute. As a workaround
2974 include a time check in the inner Heap32Next loop too.
2977 *) The code that handled flushing of data in SSL/TLS originally used the
2978 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2979 the problem outlined in PR#1949. The fix suggested there however can
2980 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2981 of Apache). So instead simplify the code to flush unconditionally.
2982 This should be fine since flushing with no data to flush is a no op.
2985 *) Handle TLS versions 2.0 and later properly and correctly use the
2986 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2987 off ancient servers have a habit of sticking around for a while...
2990 *) Modify compression code so it frees up structures without using the
2991 ex_data callbacks. This works around a problem where some applications
2992 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2993 restarting) then use compression (e.g. SSL with compression) later.
2994 This results in significant per-connection memory leaks and
2995 has caused some security issues including CVE-2008-1678 and
2999 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3000 change when encrypting or decrypting.
3003 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3004 connect and renegotiate with servers which do not support RI.
3005 Until RI is more widely deployed this option is enabled by default.
3008 *) Add "missing" ssl ctrls to clear options and mode.
3011 *) If client attempts to renegotiate and doesn't support RI respond with
3012 a no_renegotiation alert as required by RFC5746. Some renegotiating
3013 TLS clients will continue a connection gracefully when they receive
3014 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3015 waiting for a server hello which it will never receive. Now we treat a
3016 received no_renegotiation alert as a fatal error. This is because
3017 applications requesting a renegotiation might well expect it to succeed
3018 and would have no code in place to handle the server denying it so the
3019 only safe thing to do is to terminate the connection.
3022 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3023 peer supports secure renegotiation and 0 otherwise. Print out peer
3024 renegotiation support in s_client/s_server.
3027 *) Replace the highly broken and deprecated SPKAC certification method with
3028 the updated NID creation version. This should correctly handle UTF8.
3031 *) Implement RFC5746. Re-enable renegotiation but require the extension
3032 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3033 turns out to be a bad idea. It has been replaced by
3034 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3035 SSL_CTX_set_options(). This is really not recommended unless you
3036 know what you are doing.
3037 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3039 *) Fixes to stateless session resumption handling. Use initial_ctx when
3040 issuing and attempting to decrypt tickets in case it has changed during
3041 servername handling. Use a non-zero length session ID when attempting
3042 stateless session resumption: this makes it possible to determine if
3043 a resumption has occurred immediately after receiving server hello
3044 (several places in OpenSSL subtly assume this) instead of later in
3048 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3049 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3050 fixes for a few places where the return code is not checked
3052 [Julia Lawall <julia@diku.dk>]
3054 *) Add --strict-warnings option to Configure script to include devteam
3055 warnings in other configurations.
3058 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3059 makes it possible to install openssl libraries in locations which
3060 have names other than "lib", for example "/usr/lib64" which some
3062 [Steve Henson, based on patch from Jeremy Utley]
3064 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3065 X690 8.9.12 and can produce some misleading textual output of OIDs.
3066 [Steve Henson, reported by Dan Kaminsky]
3068 *) Delete MD2 from algorithm tables. This follows the recommendation in
3069 several standards that it is not used in new applications due to
3070 several cryptographic weaknesses. For binary compatibility reasons
3071 the MD2 API is still compiled in by default.
3074 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3078 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3079 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3081 [Guenter <lists@gknw.net>]
3083 *) Fix the server certificate chain building code to use X509_verify_cert(),
3084 it used to have an ad-hoc builder which was unable to cope with anything
3085 other than a simple chain.
3086 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3088 *) Don't check self signed certificate signatures in X509_verify_cert()
3089 by default (a flag can override this): it just wastes time without
3090 adding any security. As a useful side effect self signed root CAs
3091 with non-FIPS digests are now usable in FIPS mode.
3094 *) In dtls1_process_out_of_seq_message() the check if the current message
3095 is already buffered was missing. For every new message was memory
3096 allocated, allowing an attacker to perform an denial of service attack
3097 with sending out of seq handshake messages until there is no memory
3098 left. Additionally every future messege was buffered, even if the
3099 sequence number made no sense and would be part of another handshake.
3100 So only messages with sequence numbers less than 10 in advance will be
3101 buffered. (CVE-2009-1378)
3102 [Robin Seggelmann, discovered by Daniel Mentz]
3104 *) Records are buffered if they arrive with a future epoch to be
3105 processed after finishing the corresponding handshake. There is
3106 currently no limitation to this buffer allowing an attacker to perform
3107 a DOS attack with sending records with future epochs until there is no
3108 memory left. This patch adds the pqueue_size() function to detemine
3109 the size of a buffer and limits the record buffer to 100 entries.
3111 [Robin Seggelmann, discovered by Daniel Mentz]
3113 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3114 parent structure is freed. (CVE-2009-1379)
3117 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3118 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3121 [Ilya O. <vrghost@gmail.com>]
3123 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3125 *) Disable renegotiation completely - this fixes a severe security
3126 problem (CVE-2009-3555) at the cost of breaking all
3127 renegotiation. Renegotiation can be re-enabled by setting
3128 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3129 run-time. This is really not recommended unless you know what
3133 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3135 *) Don't set val to NULL when freeing up structures, it is freed up by
3136 underlying code. If sizeof(void *) > sizeof(long) this can result in
3137 zeroing past the valid field. (CVE-2009-0789)
3138 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3140 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3141 checked correctly. This would allow some invalid signed attributes to
3142 appear to verify correctly. (CVE-2009-0591)
3143 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3145 *) Reject UniversalString and BMPString types with invalid lengths. This
3146 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3147 a legal length. (CVE-2009-0590)
3150 *) Set S/MIME signing as the default purpose rather than setting it
3151 unconditionally. This allows applications to override it at the store
3155 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3156 to handle some structures.
3159 *) Improve efficiency of mem_gets: don't search whole buffer each time
3161 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3163 *) New -hex option for openssl rand.
3166 *) Print out UTF8String and NumericString when parsing ASN1.
3169 *) Support NumericString type for name components.
3172 *) Allow CC in the environment to override the automatically chosen
3173 compiler. Note that nothing is done to ensure flags work with the
3177 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3179 *) Properly check EVP_VerifyFinal() and similar return values
3181 [Ben Laurie, Bodo Moeller, Google Security Team]
3183 *) Enable TLS extensions by default.
3186 *) Allow the CHIL engine to be loaded, whether the application is
3187 multithreaded or not. (This does not release the developer from the
3188 obligation to set up the dynamic locking callbacks.)
3189 [Sander Temme <sander@temme.net>]
3191 *) Use correct exit code if there is an error in dgst command.
3192 [Steve Henson; problem pointed out by Roland Dirlewanger]
3194 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3195 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3198 *) Add experimental JPAKE support, including demo authentication in
3199 s_client and s_server.
3202 *) Set the comparison function in v3_addr_canonize().
3203 [Rob Austein <sra@hactrn.net>]
3205 *) Add support for XMPP STARTTLS in s_client.
3206 [Philip Paeps <philip@freebsd.org>]
3208 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3209 to ensure that even with this option, only ciphersuites in the
3210 server's preference list will be accepted. (Note that the option
3211 applies only when resuming a session, so the earlier behavior was
3212 just about the algorithm choice for symmetric cryptography.)
3215 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3217 *) Fix NULL pointer dereference if a DTLS server received
3218 ChangeCipherSpec as first record (CVE-2009-1386).
3221 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
3222 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3225 *) The fix in 0.9.8c that supposedly got rid of unsafe
3226 double-checked locking was incomplete for RSA blinding,
3227 addressing just one layer of what turns out to have been
3228 doubly unsafe triple-checked locking.
3230 So now fix this for real by retiring the MONT_HELPER macro
3231 in crypto/rsa/rsa_eay.c.
3233 [Bodo Moeller; problem pointed out by Marius Schilder]
3235 *) Various precautionary measures:
3237 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3239 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3240 (NB: This would require knowledge of the secret session ticket key
3241 to exploit, in which case you'd be SOL either way.)
3243 - Change bn_nist.c so that it will properly handle input BIGNUMs
3244 outside the expected range.
3246 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3249 [Neel Mehta, Bodo Moeller]
3251 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3252 the load fails. Useful for distros.
3253 [Ben Laurie and the FreeBSD team]
3255 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3258 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3261 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3263 This work was sponsored by Logica.
3266 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3267 keystores. Support for SSL/TLS client authentication too.
3268 Not compiled unless enable-capieng specified to Configure.
3270 This work was sponsored by Logica.
3273 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3274 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3275 attribute creation routines such as certifcate requests and PKCS#12
3279 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3281 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3282 handshake which could lead to a cilent crash as found using the
3283 Codenomicon TLS test suite (CVE-2008-1672)
3284 [Steve Henson, Mark Cox]
3286 *) Fix double free in TLS server name extensions which could lead to
3287 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3290 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3292 Clear the error queue to ensure that error entries left from
3293 older function calls do not interfere with the correct operation.
3294 [Lutz Jaenicke, Erik de Castro Lopo]
3296 *) Remove root CA certificates of commercial CAs:
3298 The OpenSSL project does not recommend any specific CA and does not
3299 have any policy with respect to including or excluding any CA.
3300 Therefore it does not make any sense to ship an arbitrary selection
3301 of root CA certificates with the OpenSSL software.
3304 *) RSA OAEP patches to fix two separate invalid memory reads.
3305 The first one involves inputs when 'lzero' is greater than
3306 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3307 before the beginning of from). The second one involves inputs where
3308 the 'db' section contains nothing but zeroes (there is a one-byte
3309 invalid read after the end of 'db').
3310 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3312 *) Partial backport from 0.9.9-dev:
3314 Introduce bn_mul_mont (dedicated Montgomery multiplication
3315 procedure) as a candidate for BIGNUM assembler implementation.
3316 While 0.9.9-dev uses assembler for various architectures, only
3317 x86_64 is available by default here in the 0.9.8 branch, and
3318 32-bit x86 is available through a compile-time setting.
3320 To try the 32-bit x86 assembler implementation, use Configure
3321 option "enable-montasm" (which exists only for this backport).
3323 As "enable-montasm" for 32-bit x86 disclaims code stability
3324 anyway, in this constellation we activate additional code
3325 backported from 0.9.9-dev for further performance improvements,
3326 namely BN_from_montgomery_word. (To enable this otherwise,
3327 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3329 [Andy Polyakov (backport partially by Bodo Moeller)]
3331 *) Add TLS session ticket callback. This allows an application to set
3332 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3333 values. This is useful for key rollover for example where several key
3334 sets may exist with different names.
3337 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3338 This was broken until now in 0.9.8 releases, such that the only way
3339 a registered ENGINE could be used (assuming it initialises
3340 successfully on the host) was to explicitly set it as the default
3341 for the relevant algorithms. This is in contradiction with 0.9.7
3342 behaviour and the documentation. With this fix, when an ENGINE is
3343 registered into a given algorithm's table of implementations, the
3344 'uptodate' flag is reset so that auto-discovery will be used next
3345 time a new context for that algorithm attempts to select an
3347 [Ian Lister (tweaked by Geoff Thorpe)]
3349 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3350 implemention in the following ways:
3352 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3355 Lack of BER streaming support means one pass streaming processing is
3356 only supported if data is detached: setting the streaming flag is
3357 ignored for embedded content.
3359 CMS support is disabled by default and must be explicitly enabled
3360 with the enable-cms configuration option.
3363 *) Update the GMP engine glue to do direct copies between BIGNUM and
3364 mpz_t when openssl and GMP use the same limb size. Otherwise the
3365 existing "conversion via a text string export" trick is still used.
3366 [Paul Sheer <paulsheer@gmail.com>]
3368 *) Zlib compression BIO. This is a filter BIO which compressed and
3369 uncompresses any data passed through it.
3372 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3373 RFC3394 compatible AES key wrapping.
3376 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3377 sets string data without copying. X509_ALGOR_set0() and
3378 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3379 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3380 from an X509_ATTRIBUTE structure optionally checking it occurs only
3381 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3385 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3386 to get the expected BN_FLG_CONSTTIME behavior.
3387 [Bodo Moeller (Google)]
3391 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3392 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3393 - added some more tests to do_tests.pl
3394 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3395 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3396 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3397 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3398 - various changes to netware.pl to enable gcc-cross builds on Win32
3400 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3401 - various changes to fix missing prototype warnings
3402 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3403 - added AES, WHIRLPOOL and CPUID assembler code to build files
3404 - added missing AES assembler make rules to mk1mf.pl
3405 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3406 [Guenter Knauf <eflash@gmx.net>]
3408 *) Implement certificate status request TLS extension defined in RFC3546.
3409 A client can set the appropriate parameters and receive the encoded
3410 OCSP response via a callback. A server can query the supplied parameters
3411 and set the encoded OCSP response in the callback. Add simplified examples
3412 to s_client and s_server.
3415 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3417 *) Fix various bugs:
3418 + Binary incompatibility of ssl_ctx_st structure
3419 + DTLS interoperation with non-compliant servers
3420 + Don't call get_session_cb() without proposed session
3421 + Fix ia64 assembler code
3422 [Andy Polyakov, Steve Henson]
3424 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3426 *) DTLS Handshake overhaul. There were longstanding issues with
3427 OpenSSL DTLS implementation, which were making it impossible for
3428 RFC 4347 compliant client to communicate with OpenSSL server.
3429 Unfortunately just fixing these incompatibilities would "cut off"
3430 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3431 server keeps tolerating non RFC compliant syntax. The opposite is
3432 not true, 0.9.8f client can not communicate with earlier server.
3433 This update even addresses CVE-2007-4995.
3436 *) Changes to avoid need for function casts in OpenSSL: some compilers
3437 (gcc 4.2 and later) reject their use.
3438 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3441 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3442 RFC4507bis. The encrypted ticket format is an encrypted encoded
3443 SSL_SESSION structure, that way new session features are automatically
3446 If a client application caches session in an SSL_SESSION structure
3447 support is transparent because tickets are now stored in the encoded
3450 The SSL_CTX structure automatically generates keys for ticket
3451 protection in servers so again support should be possible
3452 with no application modification.
3454 If a client or server wishes to disable RFC4507 support then the option
3455 SSL_OP_NO_TICKET can be set.
3457 Add a TLS extension debugging callback to allow the contents of any client
3458 or server extensions to be examined.
3460 This work was sponsored by Google.
3463 *) Add initial support for TLS extensions, specifically for the server_name
3464 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3465 have new members for a host name. The SSL data structure has an
3466 additional member SSL_CTX *initial_ctx so that new sessions can be
3467 stored in that context to allow for session resumption, even after the
3468 SSL has been switched to a new SSL_CTX in reaction to a client's
3469 server_name extension.
3471 New functions (subject to change):
3473 SSL_get_servername()
3474 SSL_get_servername_type()
3477 New CTRL codes and macros (subject to change):
3479 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3480 - SSL_CTX_set_tlsext_servername_callback()
3481 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3482 - SSL_CTX_set_tlsext_servername_arg()
3483 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3485 openssl s_client has a new '-servername ...' option.
3487 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3488 '-key2 ...', '-servername_fatal' (subject to change). This allows
3489 testing the HostName extension for a specific single host name ('-cert'
3490 and '-key' remain fallbacks for handshakes without HostName
3491 negotiation). If the unrecogninzed_name alert has to be sent, this by
3492 default is a warning; it becomes fatal with the '-servername_fatal'
3495 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3497 *) Add AES and SSE2 assembly language support to VC++ build.
3500 *) Mitigate attack on final subtraction in Montgomery reduction.
3503 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3504 (which previously caused an internal error).
3507 *) Squeeze another 10% out of IGE mode when in != out.
3510 *) AES IGE mode speedup.
3511 [Dean Gaudet (Google)]
3513 *) Add the Korean symmetric 128-bit cipher SEED (see
3514 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3515 add SEED ciphersuites from RFC 4162:
3517 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3518 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3519 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3520 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3522 To minimize changes between patchlevels in the OpenSSL 0.9.8
3523 series, SEED remains excluded from compilation unless OpenSSL
3524 is configured with 'enable-seed'.
3525 [KISA, Bodo Moeller]
3527 *) Mitigate branch prediction attacks, which can be practical if a
3528 single processor is shared, allowing a spy process to extract
3529 information. For detailed background information, see
3530 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3531 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3532 and Necessary Software Countermeasures"). The core of the change
3533 are new versions BN_div_no_branch() and
3534 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3535 respectively, which are slower, but avoid the security-relevant
3536 conditional branches. These are automatically called by BN_div()
3537 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3538 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3539 remove a conditional branch.
3541 BN_FLG_CONSTTIME is the new name for the previous
3542 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3543 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3544 in the exponent causes BN_mod_exp_mont() to use the alternative
3545 implementation in BN_mod_exp_mont_consttime().) The old name
3546 remains as a deprecated alias.
3548 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3549 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3550 constant-time implementations for more than just exponentiation.
3551 Here too the old name is kept as a deprecated alias.
3553 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3554 the BN_BLINDING structure gets an independent copy of the
3555 modulus. This means that the previous "BIGNUM *m" argument to
3556 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3557 essentially becomes "const BIGNUM *m", although we can't actually
3558 change this in the header file before 0.9.9. It allows
3559 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3560 enable BN_FLG_CONSTTIME.
3562 [Matthew D Wood (Intel Corp)]
3564 *) In the SSL/TLS server implementation, be strict about session ID
3565 context matching (which matters if an application uses a single
3566 external cache for different purposes). Previously,
3567 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3568 set. This did ensure strict client verification, but meant that,
3569 with applications using a single external cache for quite
3570 different requirements, clients could circumvent ciphersuite
3571 restrictions for a given session ID context by starting a session
3572 in a different context.
3575 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3576 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3577 authentication-only ciphersuites.
3580 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3581 not complete and could lead to a possible single byte overflow
3582 (CVE-2007-5135) [Ben Laurie]
3584 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3586 *) Since AES128 and AES256 (and similarly Camellia128 and
3587 Camellia256) share a single mask bit in the logic of
3588 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3589 kludge to work properly if AES128 is available and AES256 isn't
3590 (or if Camellia128 is available and Camellia256 isn't).
3593 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3594 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3595 When a point or a seed is encoded in a BIT STRING, we need to
3596 prevent the removal of trailing zero bits to get the proper DER
3597 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3598 of a NamedBitList, for which trailing 0 bits need to be removed.)
3601 *) Have SSL/TLS server implementation tolerate "mismatched" record
3602 protocol version while receiving ClientHello even if the
3603 ClientHello is fragmented. (The server can't insist on the
3604 particular protocol version it has chosen before the ServerHello
3605 message has informed the client about his choice.)
3608 *) Add RFC 3779 support.
3609 [Rob Austein for ARIN, Ben Laurie]
3611 *) Load error codes if they are not already present instead of using a
3612 static variable. This allows them to be cleanly unloaded and reloaded.
3613 Improve header file function name parsing.
3616 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3617 or CAPABILITY handshake as required by RFCs.
3620 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3622 *) Introduce limits to prevent malicious keys being able to
3623 cause a denial of service. (CVE-2006-2940)
3624 [Steve Henson, Bodo Moeller]
3626 *) Fix ASN.1 parsing of certain invalid structures that can result
3627 in a denial of service. (CVE-2006-2937) [Steve Henson]
3629 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3630 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3632 *) Fix SSL client code which could crash if connecting to a
3633 malicious SSLv2 server. (CVE-2006-4343)
3634 [Tavis Ormandy and Will Drewry, Google Security Team]
3636 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3637 match only those. Before that, "AES256-SHA" would be interpreted
3638 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3639 the same strength classification in 0.9.7h) as we currently only
3640 have a single AES bit in the ciphersuite description bitmap.
3641 That change, however, also applied to ciphersuite strings such as
3642 "RC4-MD5" that intentionally matched multiple ciphersuites --
3643 namely, SSL 2.0 ciphersuites in addition to the more common ones
3644 from SSL 3.0/TLS 1.0.
3646 So we change the selection algorithm again: Naming an explicit
3647 ciphersuite selects this one ciphersuite, and any other similar
3648 ciphersuite (same bitmap) from *other* protocol versions.
3649 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3650 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3652 Since SSL 2.0 does not have any ciphersuites for which the
3653 128/256 bit distinction would be relevant, this works for now.
3654 The proper fix will be to use different bits for AES128 and
3655 AES256, which would have avoided the problems from the beginning;
3656 however, bits are scarce, so we can only do this in a new release
3657 (not just a patchlevel) when we can change the SSL_CIPHER
3658 definition to split the single 'unsigned long mask' bitmap into
3659 multiple values to extend the available space.
3663 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3665 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3666 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3668 *) Add AES IGE and biIGE modes.
3671 *) Change the Unix randomness entropy gathering to use poll() when
3672 possible instead of select(), since the latter has some
3673 undesirable limitations.
3674 [Darryl Miles via Richard Levitte and Bodo Moeller]
3676 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3677 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3678 cannot be implicitly activated as part of, e.g., the "AES" alias.
3679 However, please upgrade to OpenSSL 0.9.9[-dev] for
3680 non-experimental use of the ECC ciphersuites to get TLS extension
3681 support, which is required for curve and point format negotiation
3682 to avoid potential handshake problems.
3685 *) Disable rogue ciphersuites:
3687 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3688 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3689 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3691 The latter two were purportedly from
3692 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3695 Also deactivate the remaining ciphersuites from
3696 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3697 unofficial, and the ID has long expired.
3700 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3701 dual-core machines) and other potential thread-safety issues.
3704 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3705 versions), which is now available for royalty-free use
3706 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3707 Also, add Camellia TLS ciphersuites from RFC 4132.
3709 To minimize changes between patchlevels in the OpenSSL 0.9.8
3710 series, Camellia remains excluded from compilation unless OpenSSL
3711 is configured with 'enable-camellia'.
3714 *) Disable the padding bug check when compression is in use. The padding
3715 bug check assumes the first packet is of even length, this is not
3716 necessarily true if compresssion is enabled and can result in false
3717 positives causing handshake failure. The actual bug test is ancient
3718 code so it is hoped that implementations will either have fixed it by
3719 now or any which still have the bug do not support compression.
3722 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3724 *) When applying a cipher rule check to see if string match is an explicit
3725 cipher suite and only match that one cipher suite if it is.
3728 *) Link in manifests for VC++ if needed.
3729 [Austin Ziegler <halostatue@gmail.com>]
3731 *) Update support for ECC-based TLS ciphersuites according to
3732 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3733 TLS extensions, which are supported starting with the 0.9.9
3734 branch, not in the OpenSSL 0.9.8 branch).
3737 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3738 opaque EVP_CIPHER_CTX handling.
3741 *) Fixes and enhancements to zlib compression code. We now only use
3742 "zlib1.dll" and use the default __cdecl calling convention on Win32
3743 to conform with the standards mentioned here:
3744 http://www.zlib.net/DLL_FAQ.txt
3745 Static zlib linking now works on Windows and the new --with-zlib-include
3746 --with-zlib-lib options to Configure can be used to supply the location
3747 of the headers and library. Gracefully handle case where zlib library
3751 *) Several fixes and enhancements to the OID generation code. The old code
3752 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3753 handle numbers larger than ULONG_MAX, truncated printing and had a
3754 non standard OBJ_obj2txt() behaviour.
3757 *) Add support for building of engines under engine/ as shared libraries
3758 under VC++ build system.
3761 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3762 Hopefully, we will not see any false combination of paths any more.
3765 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3767 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3768 (part of SSL_OP_ALL). This option used to disable the
3769 countermeasure against man-in-the-middle protocol-version
3770 rollback in the SSL 2.0 server implementation, which is a bad
3771 idea. (CVE-2005-2969)
3773 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3774 for Information Security, National Institute of Advanced Industrial
3775 Science and Technology [AIST], Japan)]
3777 *) Add two function to clear and return the verify parameter flags.
3780 *) Keep cipherlists sorted in the source instead of sorting them at
3781 runtime, thus removing the need for a lock.
3784 *) Avoid some small subgroup attacks in Diffie-Hellman.
3785 [Nick Mathewson and Ben Laurie]
3787 *) Add functions for well-known primes.
3790 *) Extended Windows CE support.
3791 [Satoshi Nakamura and Andy Polyakov]
3793 *) Initialize SSL_METHOD structures at compile time instead of during
3794 runtime, thus removing the need for a lock.
3797 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3798 attempting to decrypt each encrypted key in turn. Add support to
3802 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3804 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3807 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3810 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3811 key into the same file any more.
3814 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3817 *) Add -utf8 command line and config file option to 'ca'.
3818 [Stefan <stf@udoma.org]
3820 *) Removed the macro des_crypt(), as it seems to conflict with some
3821 libraries. Use DES_crypt().
3824 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3825 involves renaming the source and generated shared-libs for
3826 both. The engines will accept the corrected or legacy ids
3827 ('ncipher' and '4758_cca' respectively) when binding. NB,
3828 this only applies when building 'shared'.
3829 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3831 *) Add attribute functions to EVP_PKEY structure. Modify
3832 PKCS12_create() to recognize a CSP name attribute and
3833 use it. Make -CSP option work again in pkcs12 utility.
3836 *) Add new functionality to the bn blinding code:
3837 - automatic re-creation of the BN_BLINDING parameters after
3838 a fixed number of uses (currently 32)
3839 - add new function for parameter creation
3840 - introduce flags to control the update behaviour of the
3841 BN_BLINDING parameters
3842 - hide BN_BLINDING structure
3843 Add a second BN_BLINDING slot to the RSA structure to improve
3844 performance when a single RSA object is shared among several
3848 *) Add support for DTLS.
3849 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3851 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3852 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3855 *) Remove buggy and incompletet DH cert support from
3856 ssl/ssl_rsa.c and ssl/s3_both.c
3859 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3860 the apps/openssl applications.
3863 *) Compile clean with "-Wall -Wmissing-prototypes
3864 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3865 DEBUG_SAFESTACK must also be set.
3868 *) Change ./Configure so that certain algorithms can be disabled by default.
3869 The new counterpiece to "no-xxx" is "enable-xxx".
3871 The patented RC5 and MDC2 algorithms will now be disabled unless
3872 "enable-rc5" and "enable-mdc2", respectively, are specified.
3874 (IDEA remains enabled despite being patented. This is because IDEA
3875 is frequently required for interoperability, and there is no license
3876 fee for non-commercial use. As before, "no-idea" can be used to
3877 avoid this algorithm.)
3881 *) Add processing of proxy certificates (see RFC 3820). This work was
3882 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3883 EGEE (Enabling Grids for E-science in Europe).
3886 *) RC4 performance overhaul on modern architectures/implementations, such
3887 as Intel P4, IA-64 and AMD64.
3890 *) New utility extract-section.pl. This can be used specify an alternative
3891 section number in a pod file instead of having to treat each file as
3892 a separate case in Makefile. This can be done by adding two lines to the
3895 =for comment openssl_section:XXX
3897 The blank line is mandatory.
3901 *) New arguments -certform, -keyform and -pass for s_client and s_server
3902 to allow alternative format key and certificate files and passphrase
3906 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3907 update associated structures and add various utility functions.
3909 Add new policy related verify parameters, include policy checking in
3910 standard verify code. Enhance 'smime' application with extra parameters
3911 to support policy checking and print out.
3914 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3915 Nehemiah processors. These extensions support AES encryption in hardware
3916 as well as RNG (though RNG support is currently disabled).
3917 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3919 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3922 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3923 [Andy Polyakov and a number of other people]
3925 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3926 implementation contributed by IBM.
3927 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3929 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3930 exponent rather than 'unsigned long'. There is a corresponding change to
3931 the new 'rsa_keygen' element of the RSA_METHOD structure.
3932 [Jelte Jansen, Geoff Thorpe]
3934 *) Functionality for creating the initial serial number file is now
3935 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3937 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3938 number file to 1, which is bound to cause problems. To avoid
3939 the problems while respecting compatibility between different 0.9.7
3940 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3941 CA.pl for serial number initialization. With the new release 0.9.8,
3942 we can fix the problem directly in the 'ca' utility.)
3945 *) Reduced header interdepencies by declaring more opaque objects in
3946 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3947 give fewer recursive includes, which could break lazy source code - so
3948 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3949 developers should define this symbol when building and using openssl to
3950 ensure they track the recommended behaviour, interfaces, [etc], but
3951 backwards-compatible behaviour prevails when this isn't defined.
3954 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3957 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3958 This will generate a random key of the appropriate length based on the
3959 cipher context. The EVP_CIPHER can provide its own random key generation
3960 routine to support keys of a specific form. This is used in the des and
3961 3des routines to generate a key of the correct parity. Update S/MIME
3962 code to use new functions and hence generate correct parity DES keys.
3963 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3964 valid (weak or incorrect parity).
3967 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3968 as looking them up. This is useful when the verified structure may contain
3969 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3970 present unless the new PKCS7_NO_CRL flag is asserted.
3973 *) Extend ASN1 oid configuration module. It now additionally accepts the
3976 shortName = some long name, 1.2.3.4
3979 *) Reimplemented the BN_CTX implementation. There is now no more static
3980 limitation on the number of variables it can handle nor the depth of the
3981 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3982 information can now expand as required, and rather than having a single
3983 static array of bignums, BN_CTX now uses a linked-list of such arrays
3984 allowing it to expand on demand whilst maintaining the usefulness of
3985 BN_CTX's "bundling".
3988 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3989 to allow all RSA operations to function using a single BN_CTX.
3992 *) Preliminary support for certificate policy evaluation and checking. This
3993 is initially intended to pass the tests outlined in "Conformance Testing
3994 of Relying Party Client Certificate Path Processing Logic" v1.07.
3997 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3998 remained unused and not that useful. A variety of other little bignum
3999 tweaks and fixes have also been made continuing on from the audit (see
4003 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4004 associated ASN1, EVP and SSL functions and old ASN1 macros.
4007 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4008 and this should never fail. So the return value from the use of
4009 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4010 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4013 *) BN_CTX_get() should return zero-valued bignums, providing the same
4014 initialised value as BN_new().
4015 [Geoff Thorpe, suggested by Ulf Möller]
4017 *) Support for inhibitAnyPolicy certificate extension.
4020 *) An audit of the BIGNUM code is underway, for which debugging code is
4021 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4022 is considered valid when processing BIGNUMs, and causes execution to
4023 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4024 further steps are taken to deliberately pollute unused data in BIGNUM
4025 structures to try and expose faulty code further on. For now, openssl will
4026 (in its default mode of operation) continue to tolerate the inconsistent
4027 forms that it has tolerated in the past, but authors and packagers should
4028 consider trying openssl and their own applications when compiled with
4029 these debugging symbols defined. It will help highlight potential bugs in
4030 their own code, and will improve the test coverage for OpenSSL itself. At
4031 some point, these tighter rules will become openssl's default to improve
4032 maintainability, though the assert()s and other overheads will remain only
4033 in debugging configurations. See bn.h for more details.
4034 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4036 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4037 that can only be obtained through BN_CTX_new() (which implicitly
4038 initialises it). The presence of this function only made it possible
4039 to overwrite an existing structure (and cause memory leaks).
4042 *) Because of the callback-based approach for implementing LHASH as a
4043 template type, lh_insert() adds opaque objects to hash-tables and
4044 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4045 to clean up those corresponding objects before destroying the hash table
4046 (and losing the object pointers). So some over-zealous constifications in
4047 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4048 objects as "const" and the lh_doall[_arg] callback wrappers are not
4049 prototyped to have "const" restrictions on the object pointers they are
4050 given (and so aren't required to cast them away any more).
4053 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4054 (speed) prefers to use its own implementation. The two implementations
4055 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4056 its object type properly exposed (MS_TM) instead of casting to/from "char
4057 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4058 aren't necessarily the greatest nomenclatures - but this is what was used
4059 internally to the implementation so I've used that for now.
4062 *) Ensure that deprecated functions do not get compiled when
4063 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4064 the self-tests were still using deprecated key-generation functions so
4065 these have been updated also.
4068 *) Reorganise PKCS#7 code to separate the digest location functionality
4069 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
4070 New function PKCS7_set_digest() to set the digest type for PKCS#7
4071 digestedData type. Add additional code to correctly generate the
4072 digestedData type and add support for this type in PKCS7 initialization
4076 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4077 structure of type "other".
4080 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4081 sure the loop does correctly stop and breaking ("division by zero")
4082 modulus operations are not performed. The (pre-generated) prime
4083 table crypto/bn/bn_prime.h was already correct, but it could not be
4084 re-generated on some platforms because of the "division by zero"
4085 situation in the script.
4086 [Ralf S. Engelschall]
4088 *) Update support for ECC-based TLS ciphersuites according to
4089 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4090 SHA-1 now is only used for "small" curves (where the
4091 representation of a field element takes up to 24 bytes); for
4092 larger curves, the field element resulting from ECDH is directly
4093 used as premaster secret.
4094 [Douglas Stebila (Sun Microsystems Laboratories)]
4096 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4097 curve secp160r1 to the tests.
4098 [Douglas Stebila (Sun Microsystems Laboratories)]
4100 *) Add the possibility to load symbols globally with DSO.
4101 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4103 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4104 control of the error stack.
4107 *) Add support for STORE in ENGINE.
4110 *) Add the STORE type. The intention is to provide a common interface
4111 to certificate and key stores, be they simple file-based stores, or
4112 HSM-type store, or LDAP stores, or...
4113 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4116 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4117 pass a list of arguments to any function as well as provide a way
4118 for a function to pass data back to the caller.
4121 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4122 works like BUF_strdup() but can be used to duplicate a portion of
4123 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4127 *) Add the function sk_find_ex() which works like sk_find(), but will
4128 return an index to an element even if an exact match couldn't be
4129 found. The index is guaranteed to point at the element where the
4130 searched-for key would be inserted to preserve sorting order.
4133 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4134 takes an extra flags argument for optional functionality. Currently,
4135 the following flags are defined:
4137 OBJ_BSEARCH_VALUE_ON_NOMATCH
4138 This one gets OBJ_bsearch_ex() to return a pointer to the first
4139 element where the comparing function returns a negative or zero
4142 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4143 This one gets OBJ_bsearch_ex() to return a pointer to the first
4144 element where the comparing function returns zero. This is useful
4145 if there are more than one element where the comparing function
4149 *) Make it possible to create self-signed certificates with 'openssl ca'
4150 in such a way that the self-signed certificate becomes part of the
4151 CA database and uses the same mechanisms for serial number generation
4152 as all other certificate signing. The new flag '-selfsign' enables
4153 this functionality. Adapt CA.sh and CA.pl.in.
4156 *) Add functionality to check the public key of a certificate request
4157 against a given private. This is useful to check that a certificate
4158 request can be signed by that key (self-signing).
4161 *) Make it possible to have multiple active certificates with the same
4162 subject in the CA index file. This is done only if the keyword
4163 'unique_subject' is set to 'no' in the main CA section (default
4164 if 'CA_default') of the configuration file. The value is saved
4165 with the database itself in a separate index attribute file,
4166 named like the index file with '.attr' appended to the name.
4169 *) Generate muti valued AVAs using '+' notation in config files for
4173 *) Support for nameConstraints certificate extension.
4176 *) Support for policyConstraints certificate extension.
4179 *) Support for policyMappings certificate extension.
4182 *) Make sure the default DSA_METHOD implementation only uses its
4183 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4184 and change its own handlers to be NULL so as to remove unnecessary
4185 indirection. This lets alternative implementations fallback to the
4186 default implementation more easily.
4189 *) Support for directoryName in GeneralName related extensions
4193 *) Make it possible to link applications using Makefile.shared.
4194 Make that possible even when linking against static libraries!
4197 *) Support for single pass processing for S/MIME signing. This now
4198 means that S/MIME signing can be done from a pipe, in addition
4199 cleartext signing (multipart/signed type) is effectively streaming
4200 and the signed data does not need to be all held in memory.
4202 This is done with a new flag PKCS7_STREAM. When this flag is set
4203 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4204 is done after the data is output (and digests calculated) in
4205 SMIME_write_PKCS7().
4208 *) Add full support for -rpath/-R, both in shared libraries and
4209 applications, at least on the platforms where it's known how
4213 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4214 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4215 will now compute a table of multiples of the generator that
4216 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4217 faster (notably in the case of a single point multiplication,
4218 scalar * generator).
4219 [Nils Larsch, Bodo Moeller]
4221 *) IPv6 support for certificate extensions. The various extensions
4222 which use the IP:a.b.c.d can now take IPv6 addresses using the
4223 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4227 *) Added an ENGINE that implements RSA by performing private key
4228 exponentiations with the GMP library. The conversions to and from
4229 GMP's mpz_t format aren't optimised nor are any montgomery forms
4230 cached, and on x86 it appears OpenSSL's own performance has caught up.
4231 However there are likely to be other architectures where GMP could
4232 provide a boost. This ENGINE is not built in by default, but it can be
4233 specified at Configure time and should be accompanied by the necessary
4234 linker additions, eg;
4235 ./config -DOPENSSL_USE_GMP -lgmp
4238 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4239 testing availability of engines with "-t" - the old behaviour is
4240 produced by increasing the feature's verbosity with "-tt".
4243 *) ECDSA routines: under certain error conditions uninitialized BN objects
4244 could be freed. Solution: make sure initialization is performed early
4245 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4249 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4250 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4251 software implementations. For DSA and DH, parameter generation can
4252 also be overriden by providing the appropriate method callbacks.
4255 *) Change the "progress" mechanism used in key-generation and
4256 primality testing to functions that take a new BN_GENCB pointer in
4257 place of callback/argument pairs. The new API functions have "_ex"
4258 postfixes and the older functions are reimplemented as wrappers for
4259 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4260 declarations of the old functions to help (graceful) attempts to
4261 migrate to the new functions. Also, the new key-generation API
4262 functions operate on a caller-supplied key-structure and return
4263 success/failure rather than returning a key or NULL - this is to
4264 help make "keygen" another member function of RSA_METHOD etc.
4266 Example for using the new callback interface:
4268 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4272 BN_GENCB_set(&my_cb, my_callback, my_arg);
4274 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4275 /* For the meaning of a, b in calls to my_callback(), see the
4276 * documentation of the function that calls the callback.
4277 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4278 * my_callback should return 1 if it wants BN_is_prime_ex()
4279 * to continue, or 0 to stop.
4284 *) Change the ZLIB compression method to be stateful, and make it
4285 available to TLS with the number defined in
4286 draft-ietf-tls-compression-04.txt.
4289 *) Add the ASN.1 structures and functions for CertificatePair, which
4290 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4292 CertificatePair ::= SEQUENCE {
4293 forward [0] Certificate OPTIONAL,
4294 reverse [1] Certificate OPTIONAL,
4295 -- at least one of the pair shall be present -- }
4297 Also implement the PEM functions to read and write certificate
4298 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4300 This needed to be defined, mostly for the sake of the LDAP
4301 attribute crossCertificatePair, but may prove useful elsewhere as
4305 *) Make it possible to inhibit symlinking of shared libraries in
4306 Makefile.shared, for Cygwin's sake.
4309 *) Extend the BIGNUM API by creating a function
4310 void BN_set_negative(BIGNUM *a, int neg);
4311 and a macro that behave like
4312 int BN_is_negative(const BIGNUM *a);
4314 to avoid the need to access 'a->neg' directly in applications.
4317 *) Implement fast modular reduction for pseudo-Mersenne primes
4318 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4319 EC_GROUP_new_curve_GFp() will now automatically use this
4321 [Nils Larsch <nla@trustcenter.de>]
4323 *) Add new lock type (CRYPTO_LOCK_BN).
4326 *) Change the ENGINE framework to automatically load engines
4327 dynamically from specific directories unless they could be
4328 found to already be built in or loaded. Move all the
4329 current engines except for the cryptodev one to a new
4331 The engines in engines/ are built as shared libraries if
4332 the "shared" options was given to ./Configure or ./config.
4333 Otherwise, they are inserted in libcrypto.a.
4334 /usr/local/ssl/engines is the default directory for dynamic
4335 engines, but that can be overriden at configure time through
4336 the usual use of --prefix and/or --openssldir, and at run
4337 time with the environment variable OPENSSL_ENGINES.
4338 [Geoff Thorpe and Richard Levitte]
4340 *) Add Makefile.shared, a helper makefile to build shared
4341 libraries. Addapt Makefile.org.
4344 *) Add version info to Win32 DLLs.
4345 [Peter 'Luna' Runestig" <peter@runestig.com>]
4347 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4348 can be added using this API to created arbitrary PKCS#12
4349 files while avoiding the low level API.
4351 New options to PKCS12_create(), key or cert can be NULL and
4352 will then be omitted from the output file. The encryption
4353 algorithm NIDs can be set to -1 for no encryption, the mac
4354 iteration count can be set to 0 to omit the mac.
4356 Enhance pkcs12 utility by making the -nokeys and -nocerts
4357 options work when creating a PKCS#12 file. New option -nomac
4358 to omit the mac, NONE can be set for an encryption algorithm.
4359 New code is modified to use the enhanced PKCS12_create()
4360 instead of the low level API.
4363 *) Extend ASN1 encoder to support indefinite length constructed
4364 encoding. This can output sequences tags and octet strings in
4365 this form. Modify pk7_asn1.c to support indefinite length
4366 encoding. This is experimental and needs additional code to
4367 be useful, such as an ASN1 bio and some enhanced streaming
4370 Extend template encode functionality so that tagging is passed
4371 down to the template encoder.
4374 *) Let 'openssl req' fail if an argument to '-newkey' is not
4375 recognized instead of using RSA as a default.
4378 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4379 As these are not official, they are not included in "ALL";
4380 the "ECCdraft" ciphersuite group alias can be used to select them.
4381 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4383 *) Add ECDH engine support.
4384 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4386 *) Add ECDH in new directory crypto/ecdh/.
4387 [Douglas Stebila (Sun Microsystems Laboratories)]
4389 *) Let BN_rand_range() abort with an error after 100 iterations
4390 without success (which indicates a broken PRNG).
4393 *) Change BN_mod_sqrt() so that it verifies that the input value
4394 is really the square of the return value. (Previously,
4395 BN_mod_sqrt would show GIGO behaviour.)
4398 *) Add named elliptic curves over binary fields from X9.62, SECG,
4399 and WAP/WTLS; add OIDs that were still missing.
4401 [Sheueling Chang Shantz and Douglas Stebila
4402 (Sun Microsystems Laboratories)]
4404 *) Extend the EC library for elliptic curves over binary fields
4405 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4408 EC_GF2m_simple_method
4412 EC_GROUP_new_curve_GF2m
4413 EC_GROUP_set_curve_GF2m
4414 EC_GROUP_get_curve_GF2m
4415 EC_POINT_set_affine_coordinates_GF2m
4416 EC_POINT_get_affine_coordinates_GF2m
4417 EC_POINT_set_compressed_coordinates_GF2m
4419 Point compression for binary fields is disabled by default for
4420 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4423 As binary polynomials are represented as BIGNUMs, various members
4424 of the EC_GROUP and EC_POINT data structures can be shared
4425 between the implementations for prime fields and binary fields;
4426 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4427 are essentially identical to their ..._GFp counterparts.
4428 (For simplicity, the '..._GFp' prefix has been dropped from
4429 various internal method names.)
4431 An internal 'field_div' method (similar to 'field_mul' and
4432 'field_sqr') has been added; this is used only for binary fields.
4434 [Sheueling Chang Shantz and Douglas Stebila
4435 (Sun Microsystems Laboratories)]
4437 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4438 through methods ('mul', 'precompute_mult').
4440 The generic implementations (now internally called 'ec_wNAF_mul'
4441 and 'ec_wNAF_precomputed_mult') remain the default if these
4442 methods are undefined.
4444 [Sheueling Chang Shantz and Douglas Stebila
4445 (Sun Microsystems Laboratories)]
4447 *) New function EC_GROUP_get_degree, which is defined through
4448 EC_METHOD. For curves over prime fields, this returns the bit
4449 length of the modulus.
4451 [Sheueling Chang Shantz and Douglas Stebila
4452 (Sun Microsystems Laboratories)]
4454 *) New functions EC_GROUP_dup, EC_POINT_dup.
4455 (These simply call ..._new and ..._copy).
4457 [Sheueling Chang Shantz and Douglas Stebila
4458 (Sun Microsystems Laboratories)]
4460 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4461 Polynomials are represented as BIGNUMs (where the sign bit is not
4462 used) in the following functions [macros]:
4465 BN_GF2m_sub [= BN_GF2m_add]
4466 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4467 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4468 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4470 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4471 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4472 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4473 BN_GF2m_cmp [= BN_ucmp]
4475 (Note that only the 'mod' functions are actually for fields GF(2^m).
4476 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4478 For some functions, an the irreducible polynomial defining a
4479 field can be given as an 'unsigned int[]' with strictly
4480 decreasing elements giving the indices of those bits that are set;
4481 i.e., p[] represents the polynomial
4482 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4484 p[0] > p[1] > ... > p[k] = 0.
4485 This applies to the following functions:
4490 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4491 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4493 BN_GF2m_mod_sqrt_arr
4494 BN_GF2m_mod_solve_quad_arr
4498 Conversion can be performed by the following functions:
4503 bntest.c has additional tests for binary polynomial arithmetic.
4505 Two implementations for BN_GF2m_mod_div() are available.
4506 The default algorithm simply uses BN_GF2m_mod_inv() and
4507 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4508 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4509 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4511 [Sheueling Chang Shantz and Douglas Stebila
4512 (Sun Microsystems Laboratories)]
4514 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4515 functionality is disabled at compile-time.
4516 [Douglas Stebila <douglas.stebila@sun.com>]
4518 *) Change default behaviour of 'openssl asn1parse' so that more
4519 information is visible when viewing, e.g., a certificate:
4521 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4522 mode the content of non-printable OCTET STRINGs is output in a
4523 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4524 avoid the appearance of a printable string.
4525 [Nils Larsch <nla@trustcenter.de>]
4527 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4529 EC_GROUP_set_asn1_flag()
4530 EC_GROUP_get_asn1_flag()
4531 EC_GROUP_set_point_conversion_form()
4532 EC_GROUP_get_point_conversion_form()
4533 These control ASN1 encoding details:
4534 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4535 has been set to OPENSSL_EC_NAMED_CURVE.
4536 - Points are encoded in uncompressed form by default; options for
4537 asn1_for are as for point2oct, namely
4538 POINT_CONVERSION_COMPRESSED
4539 POINT_CONVERSION_UNCOMPRESSED
4540 POINT_CONVERSION_HYBRID
4542 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4545 EC_GROUP_get0_seed()
4546 EC_GROUP_get_seed_len()
4547 This is used only for ASN1 purposes (so far).
4548 [Nils Larsch <nla@trustcenter.de>]
4550 *) Add 'field_type' member to EC_METHOD, which holds the NID
4551 of the appropriate field type OID. The new function
4552 EC_METHOD_get_field_type() returns this value.
4553 [Nils Larsch <nla@trustcenter.de>]
4558 EC_POINT_point2hex()
4559 EC_POINT_hex2point()
4560 providing useful interfaces to EC_POINT_point2oct() and
4561 EC_POINT_oct2point().
4562 [Nils Larsch <nla@trustcenter.de>]
4564 *) Change internals of the EC library so that the functions
4565 EC_GROUP_set_generator()
4566 EC_GROUP_get_generator()
4567 EC_GROUP_get_order()
4568 EC_GROUP_get_cofactor()
4569 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4570 to methods, which would lead to unnecessary code duplication when
4571 adding different types of curves.
4572 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4574 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4575 arithmetic, and such that modified wNAFs are generated
4576 (which avoid length expansion in many cases).
4579 *) Add a function EC_GROUP_check_discriminant() (defined via
4580 EC_METHOD) that verifies that the curve discriminant is non-zero.
4582 Add a function EC_GROUP_check() that makes some sanity tests
4583 on a EC_GROUP, its generator and order. This includes
4584 EC_GROUP_check_discriminant().
4585 [Nils Larsch <nla@trustcenter.de>]
4587 *) Add ECDSA in new directory crypto/ecdsa/.
4589 Add applications 'openssl ecparam' and 'openssl ecdsa'
4590 (these are based on 'openssl dsaparam' and 'openssl dsa').
4592 ECDSA support is also included in various other files across the
4593 library. Most notably,
4594 - 'openssl req' now has a '-newkey ecdsa:file' option;
4595 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4596 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4597 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4598 them suitable for ECDSA where domain parameters must be
4599 extracted before the specific public key;
4600 - ECDSA engine support has been added.
4601 [Nils Larsch <nla@trustcenter.de>]
4603 *) Include some named elliptic curves, and add OIDs from X9.62,
4604 SECG, and WAP/WTLS. Each curve can be obtained from the new
4606 EC_GROUP_new_by_curve_name(),
4607 and the list of available named curves can be obtained with
4608 EC_get_builtin_curves().
4609 Also add a 'curve_name' member to EC_GROUP objects, which can be
4611 EC_GROUP_set_curve_name()
4612 EC_GROUP_get_curve_name()
4613 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4615 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4616 was actually never needed) and in BN_mul(). The removal in BN_mul()
4617 required a small change in bn_mul_part_recursive() and the addition
4618 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4619 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4620 bn_sub_words() and bn_add_words() except they take arrays with
4624 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4626 *) Cleanse PEM buffers before freeing them since they may contain
4628 [Benjamin Bennett <ben@psc.edu>]
4630 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4631 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4632 authentication-only ciphersuites.
4635 *) Since AES128 and AES256 share a single mask bit in the logic of
4636 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4637 kludge to work properly if AES128 is available and AES256 isn't.
4640 *) Expand security boundary to match 1.1.1 module.
4643 *) Remove redundant features: hash file source, editing of test vectors
4644 modify fipsld to use external fips_premain.c signature.
4647 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4648 run algorithm test programs.
4651 *) Make algorithm test programs more tolerant of whitespace.
4654 *) Have SSL/TLS server implementation tolerate "mismatched" record
4655 protocol version while receiving ClientHello even if the
4656 ClientHello is fragmented. (The server can't insist on the
4657 particular protocol version it has chosen before the ServerHello
4658 message has informed the client about his choice.)
4661 *) Load error codes if they are not already present instead of using a
4662 static variable. This allows them to be cleanly unloaded and reloaded.
4665 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4667 *) Introduce limits to prevent malicious keys being able to
4668 cause a denial of service. (CVE-2006-2940)
4669 [Steve Henson, Bodo Moeller]
4671 *) Fix ASN.1 parsing of certain invalid structures that can result
4672 in a denial of service. (CVE-2006-2937) [Steve Henson]
4674 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4675 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4677 *) Fix SSL client code which could crash if connecting to a
4678 malicious SSLv2 server. (CVE-2006-4343)
4679 [Tavis Ormandy and Will Drewry, Google Security Team]
4681 *) Change ciphersuite string processing so that an explicit
4682 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4683 will no longer include "AES128-SHA"), and any other similar
4684 ciphersuite (same bitmap) from *other* protocol versions (so that
4685 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4686 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4687 changes from 0.9.8b and 0.9.8d.
4690 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4692 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4693 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4695 *) Change the Unix randomness entropy gathering to use poll() when
4696 possible instead of select(), since the latter has some
4697 undesirable limitations.
4698 [Darryl Miles via Richard Levitte and Bodo Moeller]
4700 *) Disable rogue ciphersuites:
4702 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4703 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4704 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4706 The latter two were purportedly from
4707 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4710 Also deactive the remaining ciphersuites from
4711 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4712 unofficial, and the ID has long expired.
4715 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4716 dual-core machines) and other potential thread-safety issues.
4719 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4721 *) Adapt fipsld and the build system to link against the validated FIPS
4722 module in FIPS mode.
4725 *) Fixes for VC++ 2005 build under Windows.
4728 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4729 from a Windows bash shell such as MSYS. It is autodetected from the
4730 "config" script when run from a VC++ environment. Modify standard VC++
4731 build to use fipscanister.o from the GNU make build.
4734 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4736 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4737 The value now differs depending on if you build for FIPS or not.
4738 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4739 safely run with a non-FIPSed libcrypto, as it may crash because of
4740 the difference induced by this change.
4743 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4745 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4746 (part of SSL_OP_ALL). This option used to disable the
4747 countermeasure against man-in-the-middle protocol-version
4748 rollback in the SSL 2.0 server implementation, which is a bad
4749 idea. (CVE-2005-2969)
4751 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4752 for Information Security, National Institute of Advanced Industrial
4753 Science and Technology [AIST], Japan)]
4755 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4756 mainly for FIPS compliance and not fully integrated at this stage.
4759 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4760 the exponentiation using a fixed-length exponent. (Otherwise,
4761 the information leaked through timing could expose the secret key
4762 after many signatures; cf. Bleichenbacher's attack on DSA with
4766 *) Make a new fixed-window mod_exp implementation the default for
4767 RSA, DSA, and DH private-key operations so that the sequence of
4768 squares and multiplies and the memory access pattern are
4769 independent of the particular secret key. This will mitigate
4770 cache-timing and potential related attacks.
4772 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4773 and this is automatically used by BN_mod_exp_mont() if the new flag
4774 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4775 will use this BN flag for private exponents unless the flag
4776 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4777 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4779 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4781 *) Change the client implementation for SSLv23_method() and
4782 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4783 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4784 (Previously, the SSL 2.0 backwards compatible Client Hello
4785 message format would be used even with SSL_OP_NO_SSLv2.)
4788 *) Add support for smime-type MIME parameter in S/MIME messages which some
4792 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4793 a threadsafe manner. Modify rsa code to use new function and add calls
4794 to dsa and dh code (which had race conditions before).
4797 *) Include the fixed error library code in the C error file definitions
4798 instead of fixing them up at runtime. This keeps the error code
4799 structures constant.
4802 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4804 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4807 *) Fixes for newer kerberos headers. NB: the casts are needed because
4808 the 'length' field is signed on one version and unsigned on another
4809 with no (?) obvious way to tell the difference, without these VC++
4810 complains. Also the "definition" of FAR (blank) is no longer included
4811 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4812 some needed definitions.
4815 *) Undo Cygwin change.
4818 *) Added support for proxy certificates according to RFC 3820.
4819 Because they may be a security thread to unaware applications,
4820 they must be explicitely allowed in run-time. See
4821 docs/HOWTO/proxy_certificates.txt for further information.
4824 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4826 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4827 server and client random values. Previously
4828 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4829 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4831 This change has negligible security impact because:
4833 1. Server and client random values still have 24 bytes of pseudo random
4836 2. Server and client random values are sent in the clear in the initial
4839 3. The master secret is derived using the premaster secret (48 bytes in
4840 size for static RSA ciphersuites) as well as client server and random
4843 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4846 [Stephen Henson, reported by UK NISCC]
4848 *) Use Windows randomness collection on Cygwin.
4851 *) Fix hang in EGD/PRNGD query when communication socket is closed
4852 prematurely by EGD/PRNGD.
4853 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4855 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4858 *) Back-port of selected performance improvements from development
4859 branch, as well as improved support for PowerPC platforms.
4862 *) Add lots of checks for memory allocation failure, error codes to indicate
4863 failure and freeing up memory if a failure occurs.
4864 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4866 *) Add new -passin argument to dgst.
4869 *) Perform some character comparisons of different types in X509_NAME_cmp:
4870 this is needed for some certificates that reencode DNs into UTF8Strings
4871 (in violation of RFC3280) and can't or wont issue name rollover
4875 *) Make an explicit check during certificate validation to see that
4876 the CA setting in each certificate on the chain is correct. As a
4877 side effect always do the following basic checks on extensions,
4878 not just when there's an associated purpose to the check:
4880 - if there is an unhandled critical extension (unless the user
4881 has chosen to ignore this fault)
4882 - if the path length has been exceeded (if one is set at all)
4883 - that certain extensions fit the associated purpose (if one has
4887 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4889 *) Avoid a race condition when CRLs are checked in a multi threaded
4890 environment. This would happen due to the reordering of the revoked
4891 entries during signature checking and serial number lookup. Now the
4892 encoding is cached and the serial number sort performed under a lock.
4893 Add new STACK function sk_is_sorted().
4896 *) Add Delta CRL to the extension code.
4899 *) Various fixes to s3_pkt.c so alerts are sent properly.
4900 [David Holmes <d.holmes@f5.com>]
4902 *) Reduce the chances of duplicate issuer name and serial numbers (in
4903 violation of RFC3280) using the OpenSSL certificate creation utilities.
4904 This is done by creating a random 64 bit value for the initial serial
4905 number when a serial number file is created or when a self signed
4906 certificate is created using 'openssl req -x509'. The initial serial
4907 number file is created using 'openssl x509 -next_serial' in CA.pl
4908 rather than being initialized to 1.
4911 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4913 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4914 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4915 [Joe Orton, Steve Henson]
4917 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4919 [Joe Orton, Steve Henson]
4921 *) Make it possible to have multiple active certificates with the same
4922 subject in the CA index file. This is done only if the keyword
4923 'unique_subject' is set to 'no' in the main CA section (default
4924 if 'CA_default') of the configuration file. The value is saved
4925 with the database itself in a separate index attribute file,
4926 named like the index file with '.attr' appended to the name.
4929 *) X509 verify fixes. Disable broken certificate workarounds when
4930 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4931 keyUsage extension present. Don't accept CRLs with unhandled critical
4932 extensions: since verify currently doesn't process CRL extensions this
4933 rejects a CRL with *any* critical extensions. Add new verify error codes
4937 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4938 A clarification of RFC2560 will require the use of OCTET STRINGs and
4939 some implementations cannot handle the current raw format. Since OpenSSL
4940 copies and compares OCSP nonces as opaque blobs without any attempt at
4941 parsing them this should not create any compatibility issues.
4944 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4945 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4946 this HMAC (and other) operations are several times slower than OpenSSL
4950 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4951 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4953 *) Use the correct content when signing type "other".
4956 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4958 *) Fix various bugs revealed by running the NISCC test suite:
4960 Stop out of bounds reads in the ASN1 code when presented with
4961 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4963 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4965 If verify callback ignores invalid public key errors don't try to check
4966 certificate signature with the NULL public key.
4970 *) New -ignore_err option in ocsp application to stop the server
4971 exiting on the first error in a request.
4974 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4975 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4979 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4980 extra data after the compression methods not only for TLS 1.0
4981 but also for SSL 3.0 (as required by the specification).
4982 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4984 *) Change X509_certificate_type() to mark the key as exported/exportable
4985 when it's 512 *bits* long, not 512 bytes.
4988 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4989 blocks during encryption.
4992 *) Various fixes to base64 BIO and non blocking I/O. On write
4993 flushes were not handled properly if the BIO retried. On read
4994 data was not being buffered properly and had various logic bugs.
4995 This also affects blocking I/O when the data being decoded is a
4999 *) Various S/MIME bugfixes and compatibility changes:
5000 output correct application/pkcs7 MIME type if
5001 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5002 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5003 of files as .eml work). Correctly handle very long lines in MIME
5007 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5009 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5010 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5011 a protocol version number mismatch like a decryption error
5012 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5015 *) Turn on RSA blinding by default in the default implementation
5016 to avoid a timing attack. Applications that don't want it can call
5017 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5018 They would be ill-advised to do so in most cases.
5019 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5021 *) Change RSA blinding code so that it works when the PRNG is not
5022 seeded (in this case, the secret RSA exponent is abused as
5023 an unpredictable seed -- if it is not unpredictable, there
5024 is no point in blinding anyway). Make RSA blinding thread-safe
5025 by remembering the creator's thread ID in rsa->blinding and
5026 having all other threads use local one-time blinding factors
5027 (this requires more computation than sharing rsa->blinding, but
5028 avoids excessive locking; and if an RSA object is not shared
5029 between threads, blinding will still be very fast).
5032 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5033 ENGINE as defaults for all supported algorithms irrespective of
5034 the 'flags' parameter. 'flags' is now honoured, so applications
5035 should make sure they are passing it correctly.
5038 *) Target "mingw" now allows native Windows code to be generated in
5039 the Cygwin environment as well as with the MinGW compiler.
5042 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5044 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5045 via timing by performing a MAC computation even if incorrrect
5046 block cipher padding has been found. This is a countermeasure
5047 against active attacks where the attacker has to distinguish
5048 between bad padding and a MAC verification error. (CVE-2003-0078)
5050 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5051 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5052 Martin Vuagnoux (EPFL, Ilion)]
5054 *) Make the no-err option work as intended. The intention with no-err
5055 is not to have the whole error stack handling routines removed from
5056 libcrypto, it's only intended to remove all the function name and
5057 reason texts, thereby removing some of the footprint that may not
5058 be interesting if those errors aren't displayed anyway.
5060 NOTE: it's still possible for any application or module to have it's
5061 own set of error texts inserted. The routines are there, just not
5062 used by default when no-err is given.
5065 *) Add support for FreeBSD on IA64.
5066 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5068 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5069 Kerberos function mit_des_cbc_cksum(). Before this change,
5070 the value returned by DES_cbc_cksum() was like the one from
5071 mit_des_cbc_cksum(), except the bytes were swapped.
5072 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5074 *) Allow an application to disable the automatic SSL chain building.
5075 Before this a rather primitive chain build was always performed in
5076 ssl3_output_cert_chain(): an application had no way to send the
5077 correct chain if the automatic operation produced an incorrect result.
5079 Now the chain builder is disabled if either:
5081 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5083 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5085 The reasoning behind this is that an application would not want the
5086 auto chain building to take place if extra chain certificates are
5087 present and it might also want a means of sending no additional
5088 certificates (for example the chain has two certificates and the
5092 *) Add the possibility to build without the ENGINE framework.
5093 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5095 *) Under Win32 gmtime() can return NULL: check return value in
5096 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5099 *) DSA routines: under certain error conditions uninitialized BN objects
5100 could be freed. Solution: make sure initialization is performed early
5101 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5102 Nils Larsch <nla@trustcenter.de> via PR#459)
5105 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5106 checked on reconnect on the client side, therefore session resumption
5107 could still fail with a "ssl session id is different" error. This
5108 behaviour is masked when SSL_OP_ALL is used due to
5109 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5110 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5111 followup to PR #377.
5114 *) IA-32 assembler support enhancements: unified ELF targets, support
5115 for SCO/Caldera platforms, fix for Cygwin shared build.
5118 *) Add support for FreeBSD on sparc64. As a consequence, support for
5119 FreeBSD on non-x86 processors is separate from x86 processors on
5120 the config script, much like the NetBSD support.
5121 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5123 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5125 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5128 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5129 code (06) was taken as the first octet of the session ID and the last
5130 octet was ignored consequently. As a result SSLv2 client side session
5131 caching could not have worked due to the session ID mismatch between
5133 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5137 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5138 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5142 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5143 seems that in spite of existing for more than a year, many application
5144 author have done nothing to provide the necessary callbacks, which
5145 means that this particular engine will not work properly anywhere.
5146 This is a very unfortunate situation which forces us, in the name
5147 of usability, to give the hw_ncipher.c a static lock, which is part
5149 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5150 appear in 0.9.8 or later. We EXPECT application authors to have
5151 dealt properly with this when 0.9.8 is released (unless we actually
5152 make such changes in the libcrypto locking code that changes will
5153 have to be made anyway).
5156 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5157 octets have been read, EOF or an error occurs. Without this change
5158 some truncated ASN1 structures will not produce an error.
5161 *) Disable Heimdal support, since it hasn't been fully implemented.
5162 Still give the possibility to force the use of Heimdal, but with
5163 warnings and a request that patches get sent to openssl-dev.
5166 *) Add the VC-CE target, introduce the WINCE sysname, and add
5167 INSTALL.WCE and appropriate conditionals to make it build.
5168 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5170 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5171 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5172 edit numbers of the version.
5173 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5175 *) Introduce safe string copy and catenation functions
5176 (BUF_strlcpy() and BUF_strlcat()).
5177 [Ben Laurie (CHATS) and Richard Levitte]
5179 *) Avoid using fixed-size buffers for one-line DNs.
5180 [Ben Laurie (CHATS)]
5182 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5183 resizing buffers containing secrets, and use where appropriate.
5184 [Ben Laurie (CHATS)]
5186 *) Avoid using fixed size buffers for configuration file location.
5187 [Ben Laurie (CHATS)]
5189 *) Avoid filename truncation for various CA files.
5190 [Ben Laurie (CHATS)]
5192 *) Use sizeof in preference to magic numbers.
5193 [Ben Laurie (CHATS)]
5195 *) Avoid filename truncation in cert requests.
5196 [Ben Laurie (CHATS)]
5198 *) Add assertions to check for (supposedly impossible) buffer
5200 [Ben Laurie (CHATS)]
5202 *) Don't cache truncated DNS entries in the local cache (this could
5203 potentially lead to a spoofing attack).
5204 [Ben Laurie (CHATS)]
5206 *) Fix various buffers to be large enough for hex/decimal
5207 representations in a platform independent manner.
5208 [Ben Laurie (CHATS)]
5210 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5211 resizing buffers containing secrets, and use where appropriate.
5212 [Ben Laurie (CHATS)]
5214 *) Add BIO_indent() to avoid much slightly worrying code to do
5216 [Ben Laurie (CHATS)]
5218 *) Convert sprintf()/BIO_puts() to BIO_printf().
5219 [Ben Laurie (CHATS)]
5221 *) buffer_gets() could terminate with the buffer only half
5223 [Ben Laurie (CHATS)]
5225 *) Add assertions to prevent user-supplied crypto functions from
5226 overflowing internal buffers by having large block sizes, etc.
5227 [Ben Laurie (CHATS)]
5229 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5231 [Ben Laurie (CHATS)]
5233 *) Eliminate unused copy of key in RC4.
5234 [Ben Laurie (CHATS)]
5236 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5237 [Ben Laurie (CHATS)]
5239 *) Fix off-by-one error in EGD path.
5240 [Ben Laurie (CHATS)]
5242 *) If RANDFILE path is too long, ignore instead of truncating.
5243 [Ben Laurie (CHATS)]
5245 *) Eliminate unused and incorrectly sized X.509 structure
5247 [Ben Laurie (CHATS)]
5249 *) Eliminate unused and dangerous function knumber().
5250 [Ben Laurie (CHATS)]
5252 *) Eliminate unused and dangerous structure, KSSL_ERR.
5253 [Ben Laurie (CHATS)]
5255 *) Protect against overlong session ID context length in an encoded
5256 session object. Since these are local, this does not appear to be
5258 [Ben Laurie (CHATS)]
5260 *) Change from security patch (see 0.9.6e below) that did not affect
5261 the 0.9.6 release series:
5263 Remote buffer overflow in SSL3 protocol - an attacker could
5264 supply an oversized master key in Kerberos-enabled versions.
5266 [Ben Laurie (CHATS)]
5268 *) Change the SSL kerb5 codes to match RFC 2712.
5271 *) Make -nameopt work fully for req and add -reqopt switch.
5272 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5274 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5275 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5277 *) Make sure tests can be performed even if the corresponding algorithms
5278 have been removed entirely. This was also the last step to make
5279 OpenSSL compilable with DJGPP under all reasonable conditions.
5280 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5282 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5283 to allow version independent disabling of normally unselected ciphers,
5284 which may be activated as a side-effect of selecting a single cipher.
5286 (E.g., cipher list string "RSA" enables ciphersuites that are left
5287 out of "ALL" because they do not provide symmetric encryption.
5288 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5289 [Lutz Jaenicke, Bodo Moeller]
5291 *) Add appropriate support for separate platform-dependent build
5292 directories. The recommended way to make a platform-dependent
5293 build directory is the following (tested on Linux), maybe with
5296 # Place yourself outside of the OpenSSL source tree. In
5297 # this example, the environment variable OPENSSL_SOURCE
5298 # is assumed to contain the absolute OpenSSL source directory.
5299 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5300 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5301 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5302 mkdir -p `dirname $F`
5303 ln -s $OPENSSL_SOURCE/$F $F
5306 To be absolutely sure not to disturb the source tree, a "make clean"
5307 is a good thing. If it isn't successfull, don't worry about it,
5308 it probably means the source directory is very clean.
5311 *) Make sure any ENGINE control commands make local copies of string
5312 pointers passed to them whenever necessary. Otherwise it is possible
5313 the caller may have overwritten (or deallocated) the original string
5314 data when a later ENGINE operation tries to use the stored values.
5315 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5317 *) Improve diagnostics in file reading and command-line digests.
5318 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5320 *) Add AES modes CFB and OFB to the object database. Correct an
5321 error in AES-CFB decryption.
5324 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5325 allows existing EVP_CIPHER_CTX structures to be reused after
5326 calling EVP_*Final(). This behaviour is used by encryption
5327 BIOs and some applications. This has the side effect that
5328 applications must explicitly clean up cipher contexts with
5329 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5332 *) Check the values of dna and dnb in bn_mul_recursive before calling
5333 bn_mul_comba (a non zero value means the a or b arrays do not contain
5334 n2 elements) and fallback to bn_mul_normal if either is not zero.
5337 *) Fix escaping of non-ASCII characters when using the -subj option
5338 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5341 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5342 form for "surname", serialNumber has no short form.
5343 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5344 therefore remove "mail" short name for "internet 7".
5345 The OID for unique identifiers in X509 certificates is
5346 x500UniqueIdentifier, not uniqueIdentifier.
5347 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5350 *) Add an "init" command to the ENGINE config module and auto initialize
5351 ENGINEs. Without any "init" command the ENGINE will be initialized
5352 after all ctrl commands have been executed on it. If init=1 the
5353 ENGINE is initailized at that point (ctrls before that point are run
5354 on the uninitialized ENGINE and after on the initialized one). If
5355 init=0 then the ENGINE will not be iniatialized at all.
5358 *) Fix the 'app_verify_callback' interface so that the user-defined
5359 argument is actually passed to the callback: In the
5360 SSL_CTX_set_cert_verify_callback() prototype, the callback
5361 declaration has been changed from
5364 int (*cb)(X509_STORE_CTX *,void *);
5365 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5366 i=s->ctx->app_verify_callback(&ctx)
5367 has been changed into
5368 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5370 To update applications using SSL_CTX_set_cert_verify_callback(),
5371 a dummy argument can be added to their callback functions.
5372 [D. K. Smetters <smetters@parc.xerox.com>]
5374 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5375 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5377 *) Add and OPENSSL_LOAD_CONF define which will cause
5378 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5379 This allows older applications to transparently support certain
5380 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5381 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5382 load the config file and OPENSSL_add_all_algorithms_conf() which will
5383 always load it have also been added.
5386 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5387 Adjust NIDs and EVP layer.
5388 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5390 *) Config modules support in openssl utility.
5392 Most commands now load modules from the config file,
5393 though in a few (such as version) this isn't done
5394 because it couldn't be used for anything.
5396 In the case of ca and req the config file used is
5397 the same as the utility itself: that is the -config
5398 command line option can be used to specify an
5402 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5403 use "openssl_conf" if filename is NULL use default openssl config file.
5406 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5407 config section name. Add a new flag to tolerate a missing config file
5408 and move code to CONF_modules_load_file().
5411 *) Support for crypto accelerator cards from Accelerated Encryption
5412 Processing, www.aep.ie. (Use engine 'aep')
5413 The support was copied from 0.9.6c [engine] and adapted/corrected
5414 to work with the new engine framework.
5415 [AEP Inc. and Richard Levitte]
5417 *) Support for SureWare crypto accelerator cards from Baltimore
5418 Technologies. (Use engine 'sureware')
5419 The support was copied from 0.9.6c [engine] and adapted
5420 to work with the new engine framework.
5423 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5424 make the newer ENGINE framework commands for the CHIL engine work.
5425 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5427 *) Make it possible to produce shared libraries on ReliantUNIX.
5428 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5430 *) Add the configuration target debug-linux-ppro.
5431 Make 'openssl rsa' use the general key loading routines
5432 implemented in apps.c, and make those routines able to
5433 handle the key format FORMAT_NETSCAPE and the variant
5435 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5437 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5438 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5440 *) Add -keyform to rsautl, and document -engine.
5441 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5443 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5444 BIO_R_NO_SUCH_FILE error code rather than the generic
5445 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5448 *) Add new functions
5450 ERR_peek_last_error_line
5451 ERR_peek_last_error_line_data.
5452 These are similar to
5455 ERR_peek_error_line_data,
5456 but report on the latest error recorded rather than the first one
5457 still in the error queue.
5458 [Ben Laurie, Bodo Moeller]
5460 *) default_algorithms option in ENGINE config module. This allows things
5462 default_algorithms = ALL
5463 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5466 *) Prelminary ENGINE config module.
5469 *) New experimental application configuration code.
5472 *) Change the AES code to follow the same name structure as all other
5473 symmetric ciphers, and behave the same way. Move everything to
5474 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5475 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5477 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5478 [Ben Laurie and Theo de Raadt]
5480 *) Add option to output public keys in req command.
5481 [Massimiliano Pala madwolf@openca.org]
5483 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5484 (up to about 10% better than before for P-192 and P-224).
5487 *) New functions/macros
5489 SSL_CTX_set_msg_callback(ctx, cb)
5490 SSL_CTX_set_msg_callback_arg(ctx, arg)
5491 SSL_set_msg_callback(ssl, cb)
5492 SSL_set_msg_callback_arg(ssl, arg)
5494 to request calling a callback function
5496 void cb(int write_p, int version, int content_type,
5497 const void *buf, size_t len, SSL *ssl, void *arg)
5499 whenever a protocol message has been completely received
5500 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5501 protocol version according to which the SSL library interprets
5502 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5503 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5504 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5505 specification (change_cipher_spec(20), alert(21), handshake(22)).
5506 'buf' and 'len' point to the actual message, 'ssl' to the
5507 SSL object, and 'arg' is the application-defined value set by
5508 SSL[_CTX]_set_msg_callback_arg().
5510 'openssl s_client' and 'openssl s_server' have new '-msg' options
5511 to enable a callback that displays all protocol messages.
5514 *) Change the shared library support so shared libraries are built as
5515 soon as the corresponding static library is finished, and thereby get
5516 openssl and the test programs linked against the shared library.
5517 This still only happens when the keyword "shard" has been given to
5518 the configuration scripts.
5520 NOTE: shared library support is still an experimental thing, and
5521 backward binary compatibility is still not guaranteed.
5522 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5524 *) Add support for Subject Information Access extension.
5525 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5527 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5528 additional bytes when new memory had to be allocated, not just
5529 when reusing an existing buffer.
5532 *) New command line and configuration option 'utf8' for the req command.
5533 This allows field values to be specified as UTF8 strings.
5536 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5537 runs for the former and machine-readable output for the latter.
5540 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5541 of the e-mail address in the DN (i.e., it will go into a certificate
5542 extension only). The new configuration file option 'email_in_dn = no'
5543 has the same effect.
5544 [Massimiliano Pala madwolf@openca.org]
5546 *) Change all functions with names starting with des_ to be starting
5547 with DES_ instead. Add wrappers that are compatible with libdes,
5548 but are named _ossl_old_des_*. Finally, add macros that map the
5549 des_* symbols to the corresponding _ossl_old_des_* if libdes
5550 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5551 desired, the des_* symbols will be mapped to DES_*, with one
5554 Since we provide two compatibility mappings, the user needs to
5555 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5556 compatibility is desired. The default (i.e., when that macro
5557 isn't defined) is OpenSSL 0.9.6c compatibility.
5559 There are also macros that enable and disable the support of old
5560 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5561 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5562 are defined, the default will apply: to support the old des routines.
5564 In either case, one must include openssl/des.h to get the correct
5565 definitions. Do not try to just include openssl/des_old.h, that
5568 NOTE: This is a major break of an old API into a new one. Software
5569 authors are encouraged to switch to the DES_ style functions. Some
5570 time in the future, des_old.h and the libdes compatibility functions
5571 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5572 default), and then completely removed.
5575 *) Test for certificates which contain unsupported critical extensions.
5576 If such a certificate is found during a verify operation it is
5577 rejected by default: this behaviour can be overridden by either
5578 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5579 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5580 X509_supported_extension() has also been added which returns 1 if a
5581 particular extension is supported.
5584 *) Modify the behaviour of EVP cipher functions in similar way to digests
5585 to retain compatibility with existing code.
5588 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5589 compatibility with existing code. In particular the 'ctx' parameter does
5590 not have to be to be initialized before the call to EVP_DigestInit() and
5591 it is tidied up after a call to EVP_DigestFinal(). New function
5592 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5593 EVP_MD_CTX_copy() changed to not require the destination to be
5594 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5595 requires the destination to be valid.
5597 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5598 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5601 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5602 so that complete 'Handshake' protocol structures are kept in memory
5603 instead of overwriting 'msg_type' and 'length' with 'body' data.
5606 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5607 [Massimo Santin via Richard Levitte]
5609 *) Major restructuring to the underlying ENGINE code. This includes
5610 reduction of linker bloat, separation of pure "ENGINE" manipulation
5611 (initialisation, etc) from functionality dealing with implementations
5612 of specific crypto iterfaces. This change also introduces integrated
5613 support for symmetric ciphers and digest implementations - so ENGINEs
5614 can now accelerate these by providing EVP_CIPHER and EVP_MD
5615 implementations of their own. This is detailed in crypto/engine/README
5616 as it couldn't be adequately described here. However, there are a few
5617 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5618 were changed in the original introduction of ENGINE code have now
5619 reverted back - the hooking from this code to ENGINE is now a good
5620 deal more passive and at run-time, operations deal directly with
5621 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5622 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5623 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5624 they were not being used by the framework as there is no concept of a
5625 BIGNUM_METHOD and they could not be generalised to the new
5626 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5627 ENGINE_cpy() has been removed as it cannot be consistently defined in
5631 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5634 *) Change mkdef.pl to sort symbols that get the same entry number,
5635 and make sure the automatically generated functions ERR_load_*
5636 become part of libeay.num as well.
5639 *) New function SSL_renegotiate_pending(). This returns true once
5640 renegotiation has been requested (either SSL_renegotiate() call
5641 or HelloRequest/ClientHello receveived from the peer) and becomes
5642 false once a handshake has been completed.
5643 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5644 sends a HelloRequest, but does not ensure that a handshake takes
5645 place. SSL_renegotiate_pending() is useful for checking if the
5646 client has followed the request.)
5649 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5650 By default, clients may request session resumption even during
5651 renegotiation (if session ID contexts permit); with this option,
5652 session resumption is possible only in the first handshake.
5654 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5655 more bits available for options that should not be part of
5656 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5659 *) Add some demos for certificate and certificate request creation.
5662 *) Make maximum certificate chain size accepted from the peer application
5663 settable (SSL*_get/set_max_cert_list()), as proposed by
5664 "Douglas E. Engert" <deengert@anl.gov>.
5667 *) Add support for shared libraries for Unixware-7
5668 (Boyd Lynn Gerber <gerberb@zenez.com>).
5671 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5672 be done prior to destruction. Use this to unload error strings from
5673 ENGINEs that load their own error strings. NB: This adds two new API
5674 functions to "get" and "set" this destroy handler in an ENGINE.
5677 *) Alter all existing ENGINE implementations (except "openssl" and
5678 "openbsd") to dynamically instantiate their own error strings. This
5679 makes them more flexible to be built both as statically-linked ENGINEs
5680 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5681 Also, add stub code to each that makes building them as self-contained
5682 shared-libraries easier (see README.ENGINE).
5685 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5686 implementations into applications that are completely implemented in
5687 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5688 commands that can be used to configure what shared-library to load and
5689 to control aspects of the way it is handled. Also, made an update to
5690 the README.ENGINE file that brings its information up-to-date and
5691 provides some information and instructions on the "dynamic" ENGINE
5692 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5695 *) Make it possible to unload ranges of ERR strings with a new
5696 "ERR_unload_strings" function.
5699 *) Add a copy() function to EVP_MD.
5702 *) Make EVP_MD routines take a context pointer instead of just the
5703 md_data void pointer.
5706 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5707 that the digest can only process a single chunk of data
5708 (typically because it is provided by a piece of
5709 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5710 is only going to provide a single chunk of data, and hence the
5711 framework needn't accumulate the data for oneshot drivers.
5714 *) As with "ERR", make it possible to replace the underlying "ex_data"
5715 functions. This change also alters the storage and management of global
5716 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5717 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5718 index counters. The API functions that use this state have been changed
5719 to take a "class_index" rather than pointers to the class's local STACK
5720 and counter, and there is now an API function to dynamically create new
5721 classes. This centralisation allows us to (a) plug a lot of the
5722 thread-safety problems that existed, and (b) makes it possible to clean
5723 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5724 such data would previously have always leaked in application code and
5725 workarounds were in place to make the memory debugging turn a blind eye
5726 to it. Application code that doesn't use this new function will still
5727 leak as before, but their memory debugging output will announce it now
5728 rather than letting it slide.
5730 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5731 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5732 has a return value to indicate success or failure.
5735 *) Make it possible to replace the underlying "ERR" functions such that the
5736 global state (2 LHASH tables and 2 locks) is only used by the "default"
5737 implementation. This change also adds two functions to "get" and "set"
5738 the implementation prior to it being automatically set the first time
5739 any other ERR function takes place. Ie. an application can call "get",
5740 pass the return value to a module it has just loaded, and that module
5741 can call its own "set" function using that value. This means the
5742 module's "ERR" operations will use (and modify) the error state in the
5743 application and not in its own statically linked copy of OpenSSL code.
5746 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5747 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5748 the operation, and provides a more encapsulated way for external code
5749 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5750 to use these functions rather than manually incrementing the counts.
5752 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5755 *) Add EVP test program.
5758 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5761 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5762 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5763 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5764 These allow a CRL to be built without having to access X509_CRL fields
5765 directly. Modify 'ca' application to use new functions.
5768 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5769 bug workarounds. Rollback attack detection is a security feature.
5770 The problem will only arise on OpenSSL servers when TLSv1 is not
5771 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5772 Software authors not wanting to support TLSv1 will have special reasons
5773 for their choice and can explicitly enable this option.
5774 [Bodo Moeller, Lutz Jaenicke]
5776 *) Rationalise EVP so it can be extended: don't include a union of
5777 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5778 (similar to those existing for EVP_CIPHER_CTX).
5783 EVP_MD_CTX_init(&md); /* new function call */
5784 EVP_DigestInit(&md, EVP_sha1());
5785 EVP_DigestUpdate(&md, in, len);
5786 EVP_DigestFinal(&md, out, NULL);
5787 EVP_MD_CTX_cleanup(&md); /* new function call */
5791 *) Make DES key schedule conform to the usual scheme, as well as
5792 correcting its structure. This means that calls to DES functions
5793 now have to pass a pointer to a des_key_schedule instead of a
5794 plain des_key_schedule (which was actually always a pointer
5797 des_key_schedule ks;
5799 des_set_key_checked(..., &ks);
5800 des_ncbc_encrypt(..., &ks, ...);
5802 (Note that a later change renames 'des_...' into 'DES_...'.)
5805 *) Initial reduction of linker bloat: the use of some functions, such as
5806 PEM causes large amounts of unused functions to be linked in due to
5807 poor organisation. For example pem_all.c contains every PEM function
5808 which has a knock on effect of linking in large amounts of (unused)
5809 ASN1 code. Grouping together similar functions and splitting unrelated
5810 functions prevents this.
5813 *) Cleanup of EVP macros.
5816 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5817 correct _ecb suffix.
5820 *) Add initial OCSP responder support to ocsp application. The
5821 revocation information is handled using the text based index
5822 use by the ca application. The responder can either handle
5823 requests generated internally, supplied in files (for example
5824 via a CGI script) or using an internal minimal server.
5827 *) Add configuration choices to get zlib compression for TLS.
5830 *) Changes to Kerberos SSL for RFC 2712 compliance:
5831 1. Implemented real KerberosWrapper, instead of just using
5832 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5833 2. Implemented optional authenticator field of KerberosWrapper.
5835 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5836 and authenticator structs; see crypto/krb5/.
5838 Generalized Kerberos calls to support multiple Kerberos libraries.
5839 [Vern Staats <staatsvr@asc.hpc.mil>,
5840 Jeffrey Altman <jaltman@columbia.edu>
5841 via Richard Levitte]
5843 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5844 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5845 values for each of the key sizes rather than having just
5846 parameters (and 'speed' generating keys each time).
5849 *) Speed up EVP routines.
5852 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5853 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5854 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5855 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5857 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5858 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5859 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5862 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5864 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5867 *) Added the OS2-EMX target.
5868 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5870 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5871 to support NCONF routines in extension code. New function CONF_set_nconf()
5872 to allow functions which take an NCONF to also handle the old LHASH
5873 structure: this means that the old CONF compatible routines can be
5874 retained (in particular wrt extensions) without having to duplicate the
5875 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5878 *) Enhance the general user interface with mechanisms for inner control
5879 and with possibilities to have yes/no kind of prompts.
5882 *) Change all calls to low level digest routines in the library and
5883 applications to use EVP. Add missing calls to HMAC_cleanup() and
5884 don't assume HMAC_CTX can be copied using memcpy().
5885 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5887 *) Add the possibility to control engines through control names but with
5888 arbitrary arguments instead of just a string.
5889 Change the key loaders to take a UI_METHOD instead of a callback
5890 function pointer. NOTE: this breaks binary compatibility with earlier
5891 versions of OpenSSL [engine].
5892 Adapt the nCipher code for these new conditions and add a card insertion
5896 *) Enhance the general user interface with mechanisms to better support
5897 dialog box interfaces, application-defined prompts, the possibility
5898 to use defaults (for example default passwords from somewhere else)
5899 and interrupts/cancellations.
5902 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5903 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5906 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5907 tidy up some unnecessarily weird code in 'sk_new()').
5908 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5910 *) Change the key loading routines for ENGINEs to use the same kind
5911 callback (pem_password_cb) as all other routines that need this
5915 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5916 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5917 than this minimum value is recommended.
5920 *) New random seeder for OpenVMS, using the system process statistics
5921 that are easily reachable.
5924 *) Windows apparently can't transparently handle global
5925 variables defined in DLLs. Initialisations such as:
5927 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5929 wont compile. This is used by the any applications that need to
5930 declare their own ASN1 modules. This was fixed by adding the option
5931 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5932 needed for static libraries under Win32.
5935 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5936 setting of purpose and trust fields. New X509_STORE trust and
5937 purpose functions and tidy up setting in other SSL functions.
5940 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5941 structure. These are inherited by X509_STORE_CTX when it is
5942 initialised. This allows various defaults to be set in the
5943 X509_STORE structure (such as flags for CRL checking and custom
5944 purpose or trust settings) for functions which only use X509_STORE_CTX
5945 internally such as S/MIME.
5947 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5948 trust settings if they are not set in X509_STORE. This allows X509_STORE
5949 purposes and trust (in S/MIME for example) to override any set by default.
5951 Add command line options for CRL checking to smime, s_client and s_server
5955 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5956 are set then the CRL is looked up in the X509_STORE structure and
5957 its validity and signature checked, then if the certificate is found
5958 in the CRL the verify fails with a revoked error.
5960 Various new CRL related callbacks added to X509_STORE_CTX structure.
5962 Command line options added to 'verify' application to support this.
5964 This needs some additional work, such as being able to handle multiple
5965 CRLs with different times, extension based lookup (rather than just
5966 by subject name) and ultimately more complete V2 CRL extension
5970 *) Add a general user interface API (crypto/ui/). This is designed
5971 to replace things like des_read_password and friends (backward
5972 compatibility functions using this new API are provided).
5973 The purpose is to remove prompting functions from the DES code
5974 section as well as provide for prompting through dialog boxes in
5975 a window system and the like.
5978 *) Add "ex_data" support to ENGINE so implementations can add state at a
5979 per-structure level rather than having to store it globally.
5982 *) Make it possible for ENGINE structures to be copied when retrieved by
5983 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5984 This causes the "original" ENGINE structure to act like a template,
5985 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5986 operational state can be localised to each ENGINE structure, despite the
5987 fact they all share the same "methods". New ENGINE structures returned in
5988 this case have no functional references and the return value is the single
5989 structural reference. This matches the single structural reference returned
5990 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5994 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5995 needs to match any other type at all we need to manually clear the
5999 *) Changes to the "openssl engine" utility to include;
6000 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6001 about an ENGINE's available control commands.
6002 - executing control commands from command line arguments using the
6003 '-pre' and '-post' switches. '-post' is only used if '-t' is
6004 specified and the ENGINE is successfully initialised. The syntax for
6005 the individual commands are colon-separated, for example;
6006 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6009 *) New dynamic control command support for ENGINEs. ENGINEs can now
6010 declare their own commands (numbers), names (strings), descriptions,
6011 and input types for run-time discovery by calling applications. A
6012 subset of these commands are implicitly classed as "executable"
6013 depending on their input type, and only these can be invoked through
6014 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6015 can be based on user input, config files, etc). The distinction is
6016 that "executable" commands cannot return anything other than a boolean
6017 result and can only support numeric or string input, whereas some
6018 discoverable commands may only be for direct use through
6019 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6020 pointers, or other custom uses. The "executable" commands are to
6021 support parameterisations of ENGINE behaviour that can be
6022 unambiguously defined by ENGINEs and used consistently across any
6023 OpenSSL-based application. Commands have been added to all the
6024 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6025 control over shared-library paths without source code alterations.
6028 *) Changed all ENGINE implementations to dynamically allocate their
6029 ENGINEs rather than declaring them statically. Apart from this being
6030 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6031 this also allows the implementations to compile without using the
6032 internal engine_int.h header.
6035 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6036 'const' value. Any code that should be able to modify a RAND_METHOD
6037 should already have non-const pointers to it (ie. they should only
6038 modify their own ones).
6041 *) Made a variety of little tweaks to the ENGINE code.
6042 - "atalla" and "ubsec" string definitions were moved from header files
6043 to C code. "nuron" string definitions were placed in variables
6044 rather than hard-coded - allowing parameterisation of these values
6045 later on via ctrl() commands.
6046 - Removed unused "#if 0"'d code.
6047 - Fixed engine list iteration code so it uses ENGINE_free() to release
6048 structural references.
6049 - Constified the RAND_METHOD element of ENGINE structures.
6050 - Constified various get/set functions as appropriate and added
6051 missing functions (including a catch-all ENGINE_cpy that duplicates
6052 all ENGINE values onto a new ENGINE except reference counts/state).
6053 - Removed NULL parameter checks in get/set functions. Setting a method
6054 or function to NULL is a way of cancelling out a previously set
6055 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6056 and doesn't justify the extra error symbols and code.
6057 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6058 flags from engine_int.h to engine.h.
6059 - Changed prototypes for ENGINE handler functions (init(), finish(),
6060 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6063 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6064 to the algorithm using long division. The binary algorithm can be
6065 used only if the modulus is odd. On 32-bit systems, it is faster
6066 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6067 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6068 up to 450 bits. In 64-bit environments, the binary algorithm
6069 appears to be advantageous for much longer moduli; here we use it
6070 for moduli up to 2048 bits.
6073 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6074 could not support the combine flag in choice fields.
6077 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6078 extensions from a certificate request to the certificate.
6081 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6082 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6083 file: this allows the display of the certificate about to be
6084 signed to be customised, to allow certain fields to be included
6085 or excluded and extension details. The old system didn't display
6086 multicharacter strings properly, omitted fields not in the policy
6087 and couldn't display additional details such as extensions.
6090 *) Function EC_POINTs_mul for multiple scalar multiplication
6091 of an arbitrary number of elliptic curve points
6092 \sum scalars[i]*points[i],
6093 optionally including the generator defined for the EC_GROUP:
6094 scalar*generator + \sum scalars[i]*points[i].
6096 EC_POINT_mul is a simple wrapper function for the typical case
6097 that the point list has just one item (besides the optional
6101 *) First EC_METHODs for curves over GF(p):
6103 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6104 operations and provides various method functions that can also
6105 operate with faster implementations of modular arithmetic.
6107 EC_GFp_mont_method() reuses most functions that are part of
6108 EC_GFp_simple_method, but uses Montgomery arithmetic.
6110 [Bodo Moeller; point addition and point doubling
6111 implementation directly derived from source code provided by
6112 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6114 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6115 crypto/ec/ec_lib.c):
6117 Curves are EC_GROUP objects (with an optional group generator)
6118 based on EC_METHODs that are built into the library.
6120 Points are EC_POINT objects based on EC_GROUP objects.
6122 Most of the framework would be able to handle curves over arbitrary
6123 finite fields, but as there are no obvious types for fields other
6124 than GF(p), some functions are limited to that for now.
6127 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6128 that the file contains a complete HTTP response.
6131 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6132 change the def and num file printf format specifier from "%-40sXXX"
6133 to "%-39s XXX". The latter will always guarantee a space after the
6134 field while the former will cause them to run together if the field
6135 is 40 of more characters long.
6138 *) Constify the cipher and digest 'method' functions and structures
6139 and modify related functions to take constant EVP_MD and EVP_CIPHER
6143 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6144 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6147 *) Modify EVP_Digest*() routines so they now return values. Although the
6148 internal software routines can never fail additional hardware versions
6152 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6154 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6155 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6158 ERR_R_NESTED_ASN1_ERROR
6160 ERR_R_MISSING_ASN1_EOS
6161 were 4 .. 9, conflicting with
6162 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6164 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6165 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6167 Add new error code 'ERR_R_INTERNAL_ERROR'.
6170 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6174 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6175 sets the subject name for a new request or supersedes the
6176 subject name in a given request. Formats that can be parsed are
6177 'CN=Some Name, OU=myOU, C=IT'
6179 'CN=Some Name/OU=myOU/C=IT'.
6181 Add options '-batch' and '-verbose' to 'openssl req'.
6182 [Massimiliano Pala <madwolf@hackmasters.net>]
6184 *) Introduce the possibility to access global variables through
6185 functions on platform were that's the best way to handle exporting
6186 global variables in shared libraries. To enable this functionality,
6187 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6188 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6189 is normally done by Configure or something similar).
6191 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6192 in the source file (foo.c) like this:
6194 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6195 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6197 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6198 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6200 OPENSSL_DECLARE_GLOBAL(int,foo);
6201 #define foo OPENSSL_GLOBAL_REF(foo)
6202 OPENSSL_DECLARE_GLOBAL(double,bar);
6203 #define bar OPENSSL_GLOBAL_REF(bar)
6205 The #defines are very important, and therefore so is including the
6206 header file everywhere where the defined globals are used.
6208 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6209 of ASN.1 items, but that structure is a bit different.
6211 The largest change is in util/mkdef.pl which has been enhanced with
6212 better and easier to understand logic to choose which symbols should
6213 go into the Windows .def files as well as a number of fixes and code
6214 cleanup (among others, algorithm keywords are now sorted
6215 lexicographically to avoid constant rewrites).
6218 *) In BN_div() keep a copy of the sign of 'num' before writing the
6219 result to 'rm' because if rm==num the value will be overwritten
6220 and produce the wrong result if 'num' is negative: this caused
6221 problems with BN_mod() and BN_nnmod().
6224 *) Function OCSP_request_verify(). This checks the signature on an
6225 OCSP request and verifies the signer certificate. The signer
6226 certificate is just checked for a generic purpose and OCSP request
6230 *) Add OCSP_check_validity() function to check the validity of OCSP
6231 responses. OCSP responses are prepared in real time and may only
6232 be a few seconds old. Simply checking that the current time lies
6233 between thisUpdate and nextUpdate max reject otherwise valid responses
6234 caused by either OCSP responder or client clock inaccuracy. Instead
6235 we allow thisUpdate and nextUpdate to fall within a certain period of
6236 the current time. The age of the response can also optionally be
6237 checked. Two new options -validity_period and -status_age added to
6241 *) If signature or public key algorithm is unrecognized print out its
6242 OID rather that just UNKNOWN.
6245 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6246 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6247 ID to be generated from the issuer certificate alone which can then be
6248 passed to OCSP_id_issuer_cmp().
6251 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6252 ASN1 modules to export functions returning ASN1_ITEM pointers
6253 instead of the ASN1_ITEM structures themselves. This adds several
6254 new macros which allow the underlying ASN1 function/structure to
6255 be accessed transparently. As a result code should not use ASN1_ITEM
6256 references directly (such as &X509_it) but instead use the relevant
6257 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6258 use of the new ASN1 code on platforms where exporting structures
6259 is problematical (for example in shared libraries) but exporting
6260 functions returning pointers to structures is not.
6263 *) Add support for overriding the generation of SSL/TLS session IDs.
6264 These callbacks can be registered either in an SSL_CTX or per SSL.
6265 The purpose of this is to allow applications to control, if they wish,
6266 the arbitrary values chosen for use as session IDs, particularly as it
6267 can be useful for session caching in multiple-server environments. A
6268 command-line switch for testing this (and any client code that wishes
6269 to use such a feature) has been added to "s_server".
6270 [Geoff Thorpe, Lutz Jaenicke]
6272 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6273 of the form '#if defined(...) || defined(...) || ...' and
6274 '#if !defined(...) && !defined(...) && ...'. This also avoids
6275 the growing number of special cases it was previously handling.
6278 *) Make all configuration macros available for application by making
6279 sure they are available in opensslconf.h, by giving them names starting
6280 with "OPENSSL_" to avoid conflicts with other packages and by making
6281 sure e_os2.h will cover all platform-specific cases together with
6283 Additionally, it is now possible to define configuration/platform-
6284 specific names (called "system identities"). In the C code, these
6285 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6286 macro with the name beginning with "OPENSSL_SYS_", which is determined
6287 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6291 *) New option -set_serial to 'req' and 'x509' this allows the serial
6292 number to use to be specified on the command line. Previously self
6293 signed certificates were hard coded with serial number 0 and the
6294 CA options of 'x509' had to use a serial number in a file which was
6298 *) New options to 'ca' utility to support V2 CRL entry extensions.
6299 Currently CRL reason, invalidity date and hold instruction are
6300 supported. Add new CRL extensions to V3 code and some new objects.
6303 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6304 disable standard block padding (aka PKCS#5 padding) in the EVP
6305 API, which was previously mandatory. This means that the data is
6306 not padded in any way and so the total length much be a multiple
6307 of the block size, otherwise an error occurs.
6310 *) Initial (incomplete) OCSP SSL support.
6313 *) New function OCSP_parse_url(). This splits up a URL into its host,
6314 port and path components: primarily to parse OCSP URLs. New -url
6315 option to ocsp utility.
6318 *) New nonce behavior. The return value of OCSP_check_nonce() now
6319 reflects the various checks performed. Applications can decide
6320 whether to tolerate certain situations such as an absent nonce
6321 in a response when one was present in a request: the ocsp application
6322 just prints out a warning. New function OCSP_add1_basic_nonce()
6323 this is to allow responders to include a nonce in a response even if
6324 the request is nonce-less.
6327 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6328 skipped when using openssl x509 multiple times on a single input file,
6329 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6332 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6333 set string type: to handle setting ASN1_TIME structures. Fix ca
6334 utility to correctly initialize revocation date of CRLs.
6337 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6338 the clients preferred ciphersuites and rather use its own preferences.
6339 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6340 Internet Explorer by ensuring unchanged hash method during stepup.
6341 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6344 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6345 to aes and add a new 'exist' option to print out symbols that don't
6349 *) Additional options to ocsp utility to allow flags to be set and
6350 additional certificates supplied.
6353 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6354 OCSP client a number of certificate to only verify the response
6358 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6359 handle the new API. Currently only ECB, CBC modes supported. Add new
6362 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6363 Encryption Standard (AES) Ciphersuites for Transport Layer
6364 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6365 not enabled by default and were not part of the "ALL" ciphersuite
6366 alias because they were not yet official; they could be
6367 explicitly requested by specifying the "AESdraft" ciphersuite
6368 group alias. In the final release of OpenSSL 0.9.7, the group
6369 alias is called "AES" and is part of "ALL".)
6370 [Ben Laurie, Steve Henson, Bodo Moeller]
6372 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6373 request to response.
6376 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6377 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6378 extract information from a certificate request. OCSP_response_create()
6379 creates a response and optionally adds a basic response structure.
6380 OCSP_basic_add1_status() adds a complete single response to a basic
6381 response and returns the OCSP_SINGLERESP structure just added (to allow
6382 extensions to be included for example). OCSP_basic_add1_cert() adds a
6383 certificate to a basic response and OCSP_basic_sign() signs a basic
6384 response with various flags. New helper functions ASN1_TIME_check()
6385 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6386 (converts ASN1_TIME to GeneralizedTime).
6389 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6390 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6391 structure from a certificate. X509_pubkey_digest() digests the public_key
6392 contents: this is used in various key identifiers.
6395 *) Make sk_sort() tolerate a NULL argument.
6396 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6398 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6399 passed by the function are trusted implicitly. If any of them signed the
6400 response then it is assumed to be valid and is not verified.
6403 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6404 to data. This was previously part of the PKCS7 ASN1 code. This
6405 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6406 [Steve Henson, reported by Kenneth R. Robinette
6407 <support@securenetterm.com>]
6409 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6410 routines: without these tracing memory leaks is very painful.
6411 Fix leaks in PKCS12 and PKCS7 routines.
6414 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6415 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6416 effectively meant GeneralizedTime would never be used. Now it
6417 is initialised to -1 but X509_time_adj() now has to check the value
6418 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6419 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6420 [Steve Henson, reported by Kenneth R. Robinette
6421 <support@securenetterm.com>]
6423 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6424 result in a zero length in the ASN1_INTEGER structure which was
6425 not consistent with the structure when d2i_ASN1_INTEGER() was used
6426 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6427 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6428 where it did not print out a minus for negative ASN1_INTEGER.
6431 *) Add summary printout to ocsp utility. The various functions which
6432 convert status values to strings have been renamed to:
6433 OCSP_response_status_str(), OCSP_cert_status_str() and
6434 OCSP_crl_reason_str() and are no longer static. New options
6435 to verify nonce values and to disable verification. OCSP response
6436 printout format cleaned up.
6439 *) Add additional OCSP certificate checks. These are those specified
6440 in RFC2560. This consists of two separate checks: the CA of the
6441 certificate being checked must either be the OCSP signer certificate
6442 or the issuer of the OCSP signer certificate. In the latter case the
6443 OCSP signer certificate must contain the OCSP signing extended key
6444 usage. This check is performed by attempting to match the OCSP
6445 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6446 in the OCSP_CERTID structures of the response.
6449 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6450 and related routines. This uses the standard OpenSSL certificate
6451 verify routines to perform initial checks (just CA validity) and
6452 to obtain the certificate chain. Then additional checks will be
6453 performed on the chain. Currently the root CA is checked to see
6454 if it is explicitly trusted for OCSP signing. This is used to set
6455 a root CA as a global signing root: that is any certificate that
6456 chains to that CA is an acceptable OCSP signing certificate.
6459 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6460 extensions from a separate configuration file.
6461 As when reading extensions from the main configuration file,
6462 the '-extensions ...' option may be used for specifying the
6464 [Massimiliano Pala <madwolf@comune.modena.it>]
6466 *) New OCSP utility. Allows OCSP requests to be generated or
6467 read. The request can be sent to a responder and the output
6468 parsed, outputed or printed in text form. Not complete yet:
6469 still needs to check the OCSP response validity.
6472 *) New subcommands for 'openssl ca':
6473 'openssl ca -status <serial>' prints the status of the cert with
6474 the given serial number (according to the index file).
6475 'openssl ca -updatedb' updates the expiry status of certificates
6477 [Massimiliano Pala <madwolf@comune.modena.it>]
6479 *) New '-newreq-nodes' command option to CA.pl. This is like
6480 '-newreq', but calls 'openssl req' with the '-nodes' option
6481 so that the resulting key is not encrypted.
6482 [Damien Miller <djm@mindrot.org>]
6484 *) New configuration for the GNU Hurd.
6485 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6487 *) Initial code to implement OCSP basic response verify. This
6488 is currently incomplete. Currently just finds the signer's
6489 certificate and verifies the signature on the response.
6492 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6493 value of OPENSSLDIR. This is available via the new '-d' option
6494 to 'openssl version', and is also included in 'openssl version -a'.
6497 *) Allowing defining memory allocation callbacks that will be given
6498 file name and line number information in additional arguments
6499 (a const char* and an int). The basic functionality remains, as
6500 well as the original possibility to just replace malloc(),
6501 realloc() and free() by functions that do not know about these
6502 additional arguments. To register and find out the current
6503 settings for extended allocation functions, the following
6504 functions are provided:
6506 CRYPTO_set_mem_ex_functions
6507 CRYPTO_set_locked_mem_ex_functions
6508 CRYPTO_get_mem_ex_functions
6509 CRYPTO_get_locked_mem_ex_functions
6511 These work the same way as CRYPTO_set_mem_functions and friends.
6512 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6513 extended allocation function is enabled.
6514 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6515 a conventional allocation function is enabled.
6516 [Richard Levitte, Bodo Moeller]
6518 *) Finish off removing the remaining LHASH function pointer casts.
6519 There should no longer be any prototype-casting required when using
6520 the LHASH abstraction, and any casts that remain are "bugs". See
6521 the callback types and macros at the head of lhash.h for details
6522 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6525 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6526 If /dev/[u]random devices are not available or do not return enough
6527 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6529 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6530 /etc/entropy will be queried once each in this sequence, quering stops
6531 when enough entropy was collected without querying more sockets.
6534 *) Change the Unix RAND_poll() variant to be able to poll several
6535 random devices, as specified by DEVRANDOM, until a sufficient amount
6536 of data has been collected. We spend at most 10 ms on each file
6537 (select timeout) and read in non-blocking mode. DEVRANDOM now
6538 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6539 (previously it was just the string "/dev/urandom"), so on typical
6540 platforms the 10 ms delay will never occur.
6541 Also separate out the Unix variant to its own file, rand_unix.c.
6542 For VMS, there's a currently-empty rand_vms.c.
6545 *) Move OCSP client related routines to ocsp_cl.c. These
6546 provide utility functions which an application needing
6547 to issue a request to an OCSP responder and analyse the
6548 response will typically need: as opposed to those which an
6549 OCSP responder itself would need which will be added later.
6551 OCSP_request_sign() signs an OCSP request with an API similar
6552 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6553 response. OCSP_response_get1_basic() extracts basic response
6554 from response. OCSP_resp_find_status(): finds and extracts status
6555 information from an OCSP_CERTID structure (which will be created
6556 when the request structure is built). These are built from lower
6557 level functions which work on OCSP_SINGLERESP structures but
6558 wont normally be used unless the application wishes to examine
6559 extensions in the OCSP response for example.
6561 Replace nonce routines with a pair of functions.
6562 OCSP_request_add1_nonce() adds a nonce value and optionally
6563 generates a random value. OCSP_check_nonce() checks the
6564 validity of the nonce in an OCSP response.
6567 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6568 This doesn't copy the supplied OCSP_CERTID and avoids the
6569 need to free up the newly created id. Change return type
6570 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6571 This can then be used to add extensions to the request.
6572 Deleted OCSP_request_new(), since most of its functionality
6573 is now in OCSP_REQUEST_new() (and the case insensitive name
6574 clash) apart from the ability to set the request name which
6575 will be added elsewhere.
6578 *) Update OCSP API. Remove obsolete extensions argument from
6579 various functions. Extensions are now handled using the new
6580 OCSP extension code. New simple OCSP HTTP function which
6581 can be used to send requests and parse the response.
6584 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6585 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6586 uses the special reorder version of SET OF to sort the attributes
6587 and reorder them to match the encoded order. This resolves a long
6588 standing problem: a verify on a PKCS7 structure just after signing
6589 it used to fail because the attribute order did not match the
6590 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6591 it uses the received order. This is necessary to tolerate some broken
6592 software that does not order SET OF. This is handled by encoding
6593 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6594 to produce the required SET OF.
6597 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6598 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6599 files to get correct declarations of the ASN.1 item variables.
6602 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6603 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6604 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6605 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6606 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6607 ASN1_ITEM and no wrapper functions.
6610 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6611 replace the old function pointer based I/O routines. Change most of
6612 the *_d2i_bio() and *_d2i_fp() functions to use these.
6615 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6616 lines, recognice more "algorithms" that can be deselected, and make
6617 it complain about algorithm deselection that isn't recognised.
6620 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6621 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6622 to use new functions. Add NO_ASN1_OLD which can be set to remove
6623 some old style ASN1 functions: this can be used to determine if old
6624 code will still work when these eventually go away.
6627 *) New extension functions for OCSP structures, these follow the
6628 same conventions as certificates and CRLs.
6631 *) New function X509V3_add1_i2d(). This automatically encodes and
6632 adds an extension. Its behaviour can be customised with various
6633 flags to append, replace or delete. Various wrappers added for
6634 certifcates and CRLs.
6637 *) Fix to avoid calling the underlying ASN1 print routine when
6638 an extension cannot be parsed. Correct a typo in the
6639 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6642 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
6643 entries for variables.
6646 *) Add functionality to apps/openssl.c for detecting locking
6647 problems: As the program is single-threaded, all we have
6648 to do is register a locking callback using an array for
6649 storing which locks are currently held by the program.
6652 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6653 SSL_get_ex_data_X509_STORE_idx(), which is used in
6654 ssl_verify_cert_chain() and thus can be called at any time
6655 during TLS/SSL handshakes so that thread-safety is essential.
6656 Unfortunately, the ex_data design is not at all suited
6657 for multi-threaded use, so it probably should be abolished.
6660 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6661 [Broadcom, tweaked and integrated by Geoff Thorpe]
6663 *) Move common extension printing code to new function
6664 X509V3_print_extensions(). Reorganise OCSP print routines and
6665 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6668 *) New function X509_signature_print() to remove duplication in some
6672 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6673 set (this was treated exactly the same as SET OF previously). This
6674 is used to reorder the STACK representing the structure to match the
6675 encoding. This will be used to get round a problem where a PKCS7
6676 structure which was signed could not be verified because the STACK
6677 order did not reflect the encoded order.
6680 *) Reimplement the OCSP ASN1 module using the new code.
6683 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6684 for its ASN1 operations. The old style function pointers still exist
6685 for now but they will eventually go away.
6688 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6689 completely replaces the old ASN1 functionality with a table driven
6690 encoder and decoder which interprets an ASN1_ITEM structure describing
6691 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6692 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6693 has also been converted to the new form.
6696 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6697 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6698 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6699 for negative moduli.
6702 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6703 of not touching the result's sign bit.
6706 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6710 *) Changed the LHASH code to use prototypes for callbacks, and created
6711 macros to declare and implement thin (optionally static) functions
6712 that provide type-safety and avoid function pointer casting for the
6713 type-specific callbacks.
6716 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6718 [Veers Staats <staatsvr@asc.hpc.mil>,
6719 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6721 *) Reformat the FAQ so the different questions and answers can be divided
6722 in sections depending on the subject.
6725 *) Have the zlib compression code load ZLIB.DLL dynamically under
6729 *) New function BN_mod_sqrt for computing square roots modulo a prime
6730 (using the probabilistic Tonelli-Shanks algorithm unless
6731 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6732 be handled deterministically).
6733 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6735 *) Make BN_mod_inverse faster by explicitly handling small quotients
6736 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6737 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6740 *) New function BN_kronecker.
6743 *) Fix BN_gcd so that it works on negative inputs; the result is
6744 positive unless both parameters are zero.
6745 Previously something reasonably close to an infinite loop was
6746 possible because numbers could be growing instead of shrinking
6747 in the implementation of Euclid's algorithm.
6750 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6751 sign of the number in question.
6753 Fix BN_is_word(a,w) to work correctly for w == 0.
6755 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6756 because its test if the absolute value of 'a' equals 'w'.
6757 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6758 it exists mostly for use in the implementations of BN_is_zero(),
6759 BN_is_one(), and BN_is_word().
6762 *) New function BN_swap.
6765 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6766 the exponentiation functions are more likely to produce reasonable
6767 results on negative inputs.
6770 *) Change BN_mod_mul so that the result is always non-negative.
6771 Previously, it could be negative if one of the factors was negative;
6772 I don't think anyone really wanted that behaviour.
6775 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6776 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6777 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6778 and add new functions:
6787 BN_mod_lshift1_quick
6791 These functions always generate non-negative results.
6793 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6794 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6796 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6797 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6798 be reduced modulo m.
6799 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6802 The following entry accidentily appeared in the CHANGES file
6803 distributed with OpenSSL 0.9.7. The modifications described in
6804 it do *not* apply to OpenSSL 0.9.7.
6806 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6807 was actually never needed) and in BN_mul(). The removal in BN_mul()
6808 required a small change in bn_mul_part_recursive() and the addition
6809 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6810 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6811 bn_sub_words() and bn_add_words() except they take arrays with
6816 *) In 'openssl passwd', verify passwords read from the terminal
6817 unless the '-salt' option is used (which usually means that
6818 verification would just waste user's time since the resulting
6819 hash is going to be compared with some given password hash)
6820 or the new '-noverify' option is used.
6822 This is an incompatible change, but it does not affect
6823 non-interactive use of 'openssl passwd' (passwords on the command
6824 line, '-stdin' option, '-in ...' option) and thus should not
6828 *) Remove all references to RSAref, since there's no more need for it.
6831 *) Make DSO load along a path given through an environment variable
6832 (SHLIB_PATH) with shl_load().
6835 *) Constify the ENGINE code as a result of BIGNUM constification.
6836 Also constify the RSA code and most things related to it. In a
6837 few places, most notable in the depth of the ASN.1 code, ugly
6838 casts back to non-const were required (to be solved at a later
6842 *) Make it so the openssl application has all engines loaded by default.
6845 *) Constify the BIGNUM routines a little more.
6848 *) Add the following functions:
6850 ENGINE_load_cswift()
6852 ENGINE_load_atalla()
6854 ENGINE_load_builtin_engines()
6856 That way, an application can itself choose if external engines that
6857 are built-in in OpenSSL shall ever be used or not. The benefit is
6858 that applications won't have to be linked with libdl or other dso
6859 libraries unless it's really needed.
6861 Changed 'openssl engine' to load all engines on demand.
6862 Changed the engine header files to avoid the duplication of some
6863 declarations (they differed!).
6866 *) 'openssl engine' can now list capabilities.
6869 *) Better error reporting in 'openssl engine'.
6872 *) Never call load_dh_param(NULL) in s_server.
6875 *) Add engine application. It can currently list engines by name and
6876 identity, and test if they are actually available.
6879 *) Improve RPM specification file by forcing symbolic linking and making
6880 sure the installed documentation is also owned by root.root.
6881 [Damien Miller <djm@mindrot.org>]
6883 *) Give the OpenSSL applications more possibilities to make use of
6884 keys (public as well as private) handled by engines.
6887 *) Add OCSP code that comes from CertCo.
6890 *) Add VMS support for the Rijndael code.
6893 *) Added untested support for Nuron crypto accelerator.
6896 *) Add support for external cryptographic devices. This code was
6897 previously distributed separately as the "engine" branch.
6898 [Geoff Thorpe, Richard Levitte]
6900 *) Rework the filename-translation in the DSO code. It is now possible to
6901 have far greater control over how a "name" is turned into a filename
6902 depending on the operating environment and any oddities about the
6903 different shared library filenames on each system.
6906 *) Support threads on FreeBSD-elf in Configure.
6909 *) Fix for SHA1 assembly problem with MASM: it produces
6910 warnings about corrupt line number information when assembling
6911 with debugging information. This is caused by the overlapping
6913 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6916 NCONF_get_number() has no error checking at all. As a replacement,
6917 NCONF_get_number_e() is defined (_e for "error checking") and is
6918 promoted strongly. The old NCONF_get_number is kept around for
6919 binary backward compatibility.
6920 Make it possible for methods to load from something other than a BIO,
6921 by providing a function pointer that is given a name instead of a BIO.
6922 For example, this could be used to load configuration data from an
6926 *) Fix for non blocking accept BIOs. Added new I/O special reason
6927 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6928 with non blocking I/O was not possible because no retry code was
6929 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6933 *) Added the beginnings of Rijndael support.
6936 *) Fix for bug in DirectoryString mask setting. Add support for
6937 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6938 to allow certificate printing to more controllable, additional
6939 'certopt' option to 'x509' to allow new printing options to be
6943 *) Clean old EAY MD5 hack from e_os.h.
6946 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6948 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6949 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6950 [Joe Orton, Steve Henson]
6952 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6954 *) Fix additional bug revealed by the NISCC test suite:
6956 Stop bug triggering large recursion when presented with
6957 certain ASN.1 tags (CVE-2003-0851)
6960 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6962 *) Fix various bugs revealed by running the NISCC test suite:
6964 Stop out of bounds reads in the ASN1 code when presented with
6965 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6967 If verify callback ignores invalid public key errors don't try to check
6968 certificate signature with the NULL public key.
6972 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6973 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6977 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6978 extra data after the compression methods not only for TLS 1.0
6979 but also for SSL 3.0 (as required by the specification).
6980 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6982 *) Change X509_certificate_type() to mark the key as exported/exportable
6983 when it's 512 *bits* long, not 512 bytes.
6986 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6988 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6989 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6990 a protocol version number mismatch like a decryption error
6991 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6994 *) Turn on RSA blinding by default in the default implementation
6995 to avoid a timing attack. Applications that don't want it can call
6996 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6997 They would be ill-advised to do so in most cases.
6998 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7000 *) Change RSA blinding code so that it works when the PRNG is not
7001 seeded (in this case, the secret RSA exponent is abused as
7002 an unpredictable seed -- if it is not unpredictable, there
7003 is no point in blinding anyway). Make RSA blinding thread-safe
7004 by remembering the creator's thread ID in rsa->blinding and
7005 having all other threads use local one-time blinding factors
7006 (this requires more computation than sharing rsa->blinding, but
7007 avoids excessive locking; and if an RSA object is not shared
7008 between threads, blinding will still be very fast).
7011 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7013 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7014 via timing by performing a MAC computation even if incorrrect
7015 block cipher padding has been found. This is a countermeasure
7016 against active attacks where the attacker has to distinguish
7017 between bad padding and a MAC verification error. (CVE-2003-0078)
7019 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7020 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7021 Martin Vuagnoux (EPFL, Ilion)]
7023 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7025 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7026 memory from it's contents. This is done with a counter that will
7027 place alternating values in each byte. This can be used to solve
7028 two issues: 1) the removal of calls to memset() by highly optimizing
7029 compilers, and 2) cleansing with other values than 0, since those can
7030 be read through on certain media, for example a swap space on disk.
7033 *) Bugfix: client side session caching did not work with external caching,
7034 because the session->cipher setting was not restored when reloading
7035 from the external cache. This problem was masked, when
7036 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7037 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7040 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7041 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7042 [Zeev Lieber <zeev-l@yahoo.com>]
7044 *) Undo an undocumented change introduced in 0.9.6e which caused
7045 repeated calls to OpenSSL_add_all_ciphers() and
7046 OpenSSL_add_all_digests() to be ignored, even after calling
7050 *) Change the default configuration reader to deal with last line not
7051 being properly terminated.
7054 *) Change X509_NAME_cmp() so it applies the special rules on handling
7055 DN values that are of type PrintableString, as well as RDNs of type
7056 emailAddress where the value has the type ia5String.
7057 [stefank@valicert.com via Richard Levitte]
7059 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7060 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7061 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7062 the bitwise-OR of the two for use by the majority of applications
7063 wanting this behaviour, and update the docs. The documented
7064 behaviour and actual behaviour were inconsistent and had been
7065 changing anyway, so this is more a bug-fix than a behavioural
7067 [Geoff Thorpe, diagnosed by Nadav Har'El]
7069 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7070 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7073 *) Fix initialization code race conditions in
7074 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7075 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7076 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7077 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7078 ssl2_get_cipher_by_char(),
7079 ssl3_get_cipher_by_char().
7080 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7082 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7083 the cached sessions are flushed, as the remove_cb() might use ex_data
7084 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7085 (see [openssl.org #212]).
7086 [Geoff Thorpe, Lutz Jaenicke]
7088 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7089 length, instead of the encoding length to d2i_ASN1_OBJECT.
7092 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7094 *) [In 0.9.6g-engine release:]
7095 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7096 [Lynn Gazis <lgazis@rainbow.com>]
7098 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7100 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7101 and get fix the header length calculation.
7102 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7103 Alon Kantor <alonk@checkpoint.com> (and others),
7106 *) Use proper error handling instead of 'assertions' in buffer
7107 overflow checks added in 0.9.6e. This prevents DoS (the
7108 assertions could call abort()).
7109 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7111 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7113 *) Add various sanity checks to asn1_get_length() to reject
7114 the ASN1 length bytes if they exceed sizeof(long), will appear
7115 negative or the content length exceeds the length of the
7117 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7119 *) Fix cipher selection routines: ciphers without encryption had no flags
7120 for the cipher strength set and where therefore not handled correctly
7121 by the selection routines (PR #130).
7124 *) Fix EVP_dsa_sha macro.
7128 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7129 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7130 that was added in OpenSSL 0.9.6d.
7132 As the countermeasure turned out to be incompatible with some
7133 broken SSL implementations, the new option is part of SSL_OP_ALL.
7134 SSL_OP_ALL is usually employed when compatibility with weird SSL
7135 implementations is desired (e.g. '-bugs' option to 's_client' and
7136 's_server'), so the new option is automatically set in many
7140 *) Changes in security patch:
7142 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7143 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7144 Air Force Materiel Command, USAF, under agreement number
7147 *) Add various sanity checks to asn1_get_length() to reject
7148 the ASN1 length bytes if they exceed sizeof(long), will appear
7149 negative or the content length exceeds the length of the
7150 supplied buffer. (CVE-2002-0659)
7151 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7153 *) Assertions for various potential buffer overflows, not known to
7155 [Ben Laurie (CHATS)]
7157 *) Various temporary buffers to hold ASCII versions of integers were
7158 too small for 64 bit platforms. (CVE-2002-0655)
7159 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7161 *) Remote buffer overflow in SSL3 protocol - an attacker could
7162 supply an oversized session ID to a client. (CVE-2002-0656)
7163 [Ben Laurie (CHATS)]
7165 *) Remote buffer overflow in SSL2 protocol - an attacker could
7166 supply an oversized client master key. (CVE-2002-0656)
7167 [Ben Laurie (CHATS)]
7169 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7171 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7172 encoded as NULL) with id-dsa-with-sha1.
7173 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7175 *) Check various X509_...() return values in apps/req.c.
7176 [Nils Larsch <nla@trustcenter.de>]
7178 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7179 an end-of-file condition would erronously be flagged, when the CRLF
7180 was just at the end of a processed block. The bug was discovered when
7181 processing data through a buffering memory BIO handing the data to a
7182 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7183 <ptsekov@syntrex.com> and Nedelcho Stanev.
7186 *) Implement a countermeasure against a vulnerability recently found
7187 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7188 before application data chunks to avoid the use of known IVs
7189 with data potentially chosen by the attacker.
7192 *) Fix length checks in ssl3_get_client_hello().
7195 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7196 to prevent ssl3_read_internal() from incorrectly assuming that
7197 ssl3_read_bytes() found application data while handshake
7198 processing was enabled when in fact s->s3->in_read_app_data was
7199 merely automatically cleared during the initial handshake.
7200 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7202 *) Fix object definitions for Private and Enterprise: they were not
7203 recognized in their shortname (=lowercase) representation. Extend
7204 obj_dat.pl to issue an error when using undefined keywords instead
7205 of silently ignoring the problem (Svenning Sorensen
7206 <sss@sss.dnsalias.net>).
7209 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7210 generators, i.e. generators other than 2 and 5. (Previously, the
7211 code did not properly initialise the 'add' and 'rem' values to
7212 BN_generate_prime().)
7214 In the new general case, we do not insist that 'generator' is
7215 actually a primitive root: This requirement is rather pointless;
7216 a generator of the order-q subgroup is just as good, if not
7220 *) Map new X509 verification errors to alerts. Discovered and submitted by
7221 Tom Wu <tom@arcot.com>.
7224 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7225 returning non-zero before the data has been completely received
7226 when using non-blocking I/O.
7227 [Bodo Moeller; problem pointed out by John Hughes]
7229 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7230 [Ben Laurie, Lutz Jaenicke]
7232 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7233 Yoram Zahavi <YoramZ@gilian.com>).
7236 *) Add information about CygWin 1.3 and on, and preserve proper
7237 configuration for the versions before that.
7238 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7240 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7241 check whether we deal with a copy of a session and do not delete from
7242 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7243 <izhar@checkpoint.com>.
7246 *) Do not store session data into the internal session cache, if it
7247 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7248 flag is set). Proposed by Aslam <aslam@funk.com>.
7251 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7255 *) [In 0.9.6d-engine release:]
7256 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7257 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7259 *) Add the configuration target linux-s390x.
7260 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7262 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7263 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7264 variable as an indication that a ClientHello message has been
7265 received. As the flag value will be lost between multiple
7266 invocations of ssl3_accept when using non-blocking I/O, the
7267 function may not be aware that a handshake has actually taken
7268 place, thus preventing a new session from being added to the
7271 To avoid this problem, we now set s->new_session to 2 instead of
7272 using a local variable.
7273 [Lutz Jaenicke, Bodo Moeller]
7275 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7276 if the SSL_R_LENGTH_MISMATCH error is detected.
7277 [Geoff Thorpe, Bodo Moeller]
7279 *) New 'shared_ldflag' column in Configure platform table.
7282 *) Fix EVP_CIPHER_mode macro.
7283 ["Dan S. Camper" <dan@bti.net>]
7285 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7286 type, we must throw them away by setting rr->length to 0.
7287 [D P Chang <dpc@qualys.com>]
7289 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7291 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7292 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7293 worked incorrectly for those cases where range = 10..._2 and
7294 3*range is two bits longer than range.)
7297 *) Only add signing time to PKCS7 structures if it is not already
7301 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7302 OBJ_ld_ce should be OBJ_id_ce.
7303 Also some ip-pda OIDs in crypto/objects/objects.txt were
7304 incorrect (cf. RFC 3039).
7305 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7307 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7308 returns early because it has nothing to do.
7309 [Andy Schneider <andy.schneider@bjss.co.uk>]
7311 *) [In 0.9.6c-engine release:]
7312 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7313 [Andy Schneider <andy.schneider@bjss.co.uk>]
7315 *) [In 0.9.6c-engine release:]
7316 Add support for Cryptographic Appliance's keyserver technology.
7317 (Use engine 'keyclient')
7318 [Cryptographic Appliances and Geoff Thorpe]
7320 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7321 is called via tools/c89.sh because arguments have to be
7322 rearranged (all '-L' options must appear before the first object
7324 [Richard Shapiro <rshapiro@abinitio.com>]
7326 *) [In 0.9.6c-engine release:]
7327 Add support for Broadcom crypto accelerator cards, backported
7329 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7331 *) [In 0.9.6c-engine release:]
7332 Add support for SureWare crypto accelerator cards from
7333 Baltimore Technologies. (Use engine 'sureware')
7334 [Baltimore Technologies and Mark Cox]
7336 *) [In 0.9.6c-engine release:]
7337 Add support for crypto accelerator cards from Accelerated
7338 Encryption Processing, www.aep.ie. (Use engine 'aep')
7339 [AEP Inc. and Mark Cox]
7341 *) Add a configuration entry for gcc on UnixWare.
7342 [Gary Benson <gbenson@redhat.com>]
7344 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7345 messages are stored in a single piece (fixed-length part and
7346 variable-length part combined) and fix various bugs found on the way.
7349 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7350 instead. BIO_gethostbyname() does not know what timeouts are
7351 appropriate, so entries would stay in cache even when they have
7353 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7355 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7356 faced with a pathologically small ClientHello fragment that does
7357 not contain client_version: Instead of aborting with an error,
7358 simply choose the highest available protocol version (i.e.,
7359 TLS 1.0 unless it is disabled). In practice, ClientHello
7360 messages are never sent like this, but this change gives us
7361 strictly correct behaviour at least for TLS.
7364 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7365 never resets s->method to s->ctx->method when called from within
7366 one of the SSL handshake functions.
7367 [Bodo Moeller; problem pointed out by Niko Baric]
7369 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7370 (sent using the client's version number) if client_version is
7371 smaller than the protocol version in use. Also change
7372 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7373 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7374 the client will at least see that alert.
7377 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7381 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7382 client receives HelloRequest while in a handshake.
7383 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7385 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7386 should end in 'break', not 'goto end' which circuments various
7387 cleanups done in state SSL_ST_OK. But session related stuff
7388 must be disabled for SSL_ST_OK in the case that we just sent a
7391 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7392 before just sending a HelloRequest.
7393 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7395 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7396 reveal whether illegal block cipher padding was found or a MAC
7397 verification error occured. (Neither SSLerr() codes nor alerts
7398 are directly visible to potential attackers, but the information
7399 may leak via logfiles.)
7401 Similar changes are not required for the SSL 2.0 implementation
7402 because the number of padding bytes is sent in clear for SSL 2.0,
7403 and the extra bytes are just ignored. However ssl/s2_pkt.c
7404 failed to verify that the purported number of padding bytes is in
7408 *) Add OpenUNIX-8 support including shared libraries
7409 (Boyd Lynn Gerber <gerberb@zenez.com>).
7412 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7413 'wristwatch attack' using huge encoding parameters (cf.
7414 James H. Manger's CRYPTO 2001 paper). Note that the
7415 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7416 encoding parameters and hence was not vulnerable.
7419 *) BN_sqr() bug fix.
7420 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7422 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7423 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7424 followed by modular reduction.
7425 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7427 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7428 equivalent based on BN_pseudo_rand() instead of BN_rand().
7431 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7432 This function was broken, as the check for a new client hello message
7433 to handle SGC did not allow these large messages.
7434 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7437 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7440 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7441 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7444 *) Rework the configuration and shared library support for Tru64 Unix.
7445 The configuration part makes use of modern compiler features and
7446 still retains old compiler behavior for those that run older versions
7447 of the OS. The shared library support part includes a variant that
7448 uses the RPATH feature, and is available through the special
7449 configuration target "alpha-cc-rpath", which will never be selected
7451 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7453 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7454 with the same message size as in ssl3_get_certificate_request().
7455 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7456 messages might inadvertently be reject as too long.
7457 [Petr Lampa <lampa@fee.vutbr.cz>]
7459 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7462 *) Modified SSL library such that the verify_callback that has been set
7463 specificly for an SSL object with SSL_set_verify() is actually being
7464 used. Before the change, a verify_callback set with this function was
7465 ignored and the verify_callback() set in the SSL_CTX at the time of
7466 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7467 to allow the necessary settings.
7470 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7471 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7472 done automatically (in contradiction to the requirements of the C
7473 standard). This made problems when used from OpenSSH.
7476 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7477 dh->length and always used
7479 BN_rand_range(priv_key, dh->p).
7481 BN_rand_range() is not necessary for Diffie-Hellman, and this
7482 specific range makes Diffie-Hellman unnecessarily inefficient if
7483 dh->length (recommended exponent length) is much smaller than the
7484 length of dh->p. We could use BN_rand_range() if the order of
7485 the subgroup was stored in the DH structure, but we only have
7490 BN_rand(priv_key, l, ...)
7492 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7498 RSA_eay_public_encrypt
7499 RSA_eay_private_decrypt
7500 RSA_eay_private_encrypt (signing)
7501 RSA_eay_public_decrypt (signature verification)
7503 (default implementations for RSA_public_encrypt,
7504 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7505 always reject numbers >= n.
7508 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7509 to synchronize access to 'locking_thread'. This is necessary on
7510 systems where access to 'locking_thread' (an 'unsigned long'
7511 variable) is not atomic.
7514 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7515 *before* setting the 'crypto_lock_rand' flag. The previous code had
7516 a race condition if 0 is a valid thread ID.
7517 [Travis Vitek <vitek@roguewave.com>]
7519 *) Add support for shared libraries under Irix.
7520 [Albert Chin-A-Young <china@thewrittenword.com>]
7522 *) Add configuration option to build on Linux on both big-endian and
7524 [Ralf Baechle <ralf@uni-koblenz.de>]
7526 *) Add the possibility to create shared libraries on HP-UX.
7529 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7531 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7532 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7533 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7534 PRNG state recovery was possible based on the output of
7535 one PRNG request appropriately sized to gain knowledge on
7536 'md' followed by enough consecutive 1-byte PRNG requests
7537 to traverse all of 'state'.
7539 1. When updating 'md_local' (the current thread's copy of 'md')
7540 during PRNG output generation, hash all of the previous
7541 'md_local' value, not just the half used for PRNG output.
7543 2. Make the number of bytes from 'state' included into the hash
7544 independent from the number of PRNG bytes requested.
7546 The first measure alone would be sufficient to avoid
7547 Markku-Juhani's attack. (Actually it had never occurred
7548 to me that the half of 'md_local' used for chaining was the
7549 half from which PRNG output bytes were taken -- I had always
7550 assumed that the secret half would be used.) The second
7551 measure makes sure that additional data from 'state' is never
7552 mixed into 'md_local' in small portions; this heuristically
7553 further strengthens the PRNG.
7556 *) Fix crypto/bn/asm/mips3.s.
7559 *) When only the key is given to "enc", the IV is undefined. Print out
7560 an error message in this case.
7563 *) Handle special case when X509_NAME is empty in X509 printing routines.
7566 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7567 positive and less than q.
7570 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7571 used: it isn't thread safe and the add_lock_callback should handle
7573 [Paul Rose <Paul.Rose@bridge.com>]
7575 *) Verify that incoming data obeys the block size in
7576 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7580 [Ulf Möller, Bodo Möller]
7582 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7583 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7584 when fixing the server behaviour for backwards-compatible 'client
7585 hello' messages. (Note that the attack is impractical against
7586 SSL 3.0 and TLS 1.0 anyway because length and version checking
7587 means that the probability of guessing a valid ciphertext is
7588 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7591 Before 0.9.5, the countermeasure (hide the error by generating a
7592 random 'decryption result') did not work properly because
7593 ERR_clear_error() was missing, meaning that SSL_get_error() would
7594 detect the supposedly ignored error.
7596 Both problems are now fixed.
7599 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7600 (previously it was 1024).
7603 *) Fix for compatibility mode trust settings: ignore trust settings
7604 unless some valid trust or reject settings are present.
7607 *) Fix for blowfish EVP: its a variable length cipher.
7610 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7611 parameters in DSA public key structures and return an error in the
7612 DSA routines if parameters are absent.
7615 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7616 in the current directory if neither $RANDFILE nor $HOME was set.
7617 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7618 caused some confusion to Windows users who haven't defined $HOME.
7619 Thus RAND_file_name() is changed again: e_os.h can define a
7620 DEFAULT_HOME, which will be used if $HOME is not set.
7621 For Windows, we use "C:"; on other platforms, we still require
7622 environment variables.
7624 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7625 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7626 having multiple threads call RAND_poll() concurrently.
7629 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7630 combination of a flag and a thread ID variable.
7631 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7632 flag), *other* threads can enter ssleay_add_bytes without obeying
7633 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7634 that they do not hold after the first thread unsets add_do_not_lock).
7637 *) Change bctest again: '-x' expressions are not available in all
7641 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7643 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7644 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7646 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7647 the default extension for executables, if any. Also, make the perl
7648 scripts that use symlink() to test if it really exists and use "cp"
7649 if it doesn't. All this made OpenSSL compilable and installable in
7653 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7654 If SEQUENCE is length is indefinite just set c->slen to the total
7655 amount of data available.
7656 [Steve Henson, reported by shige@FreeBSD.org]
7657 [This change does not apply to 0.9.7.]
7659 *) Change bctest to avoid here-documents inside command substitution
7660 (workaround for FreeBSD /bin/sh bug).
7661 For compatibility with Ultrix, avoid shell functions (introduced
7662 in the bctest version that searches along $PATH).
7665 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7666 with des_encrypt() defined on some operating systems, like Solaris
7670 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7671 On the Importance of Eliminating Errors in Cryptographic
7672 Computations, J. Cryptology 14 (2001) 2, 101-119,
7673 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7676 *) MIPS assembler BIGNUM division bug fix.
7679 *) Disabled incorrect Alpha assembler code.
7682 *) Fix PKCS#7 decode routines so they correctly update the length
7683 after reading an EOC for the EXPLICIT tag.
7685 [This change does not apply to 0.9.7.]
7687 *) Fix bug in PKCS#12 key generation routines. This was triggered
7688 if a 3DES key was generated with a 0 initial byte. Include
7689 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7690 (but broken) behaviour.
7693 *) Enhance bctest to search for a working bc along $PATH and print
7695 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7697 *) Fix memory leaks in err.c: free err_data string if necessary;
7698 don't write to the wrong index in ERR_set_error_data.
7701 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7705 *) Replace rdtsc with _emit statements for VC++ version 5.
7706 [Jeremy Cooper <jeremy@baymoo.org>]
7708 *) Make it possible to reuse SSLv2 sessions.
7711 *) In copy_email() check for >= 0 as a return value for
7712 X509_NAME_get_index_by_NID() since 0 is a valid index.
7713 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7715 *) Avoid coredump with unsupported or invalid public keys by checking if
7716 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7717 PKCS7_verify() fails with non detached data.
7720 *) Don't use getenv in library functions when run as setuid/setgid.
7721 New function OPENSSL_issetugid().
7724 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7725 due to incorrect handling of multi-threading:
7727 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7729 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7731 3. Count how many times MemCheck_off() has been called so that
7732 nested use can be treated correctly. This also avoids
7733 inband-signalling in the previous code (which relied on the
7734 assumption that thread ID 0 is impossible).
7737 *) Add "-rand" option also to s_client and s_server.
7740 *) Fix CPU detection on Irix 6.x.
7741 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7742 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7744 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7747 [This change does not apply to 0.9.7.]
7749 *) Use the cached encoding of an X509_NAME structure rather than
7750 copying it. This is apparently the reason for the libsafe "errors"
7751 but the code is actually correct.
7754 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7755 Bleichenbacher's DSA attack.
7756 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7757 to be set and top=0 forces the highest bit to be set; top=-1 is new
7758 and leaves the highest bit random.
7759 [Ulf Moeller, Bodo Moeller]
7761 *) In the NCONF_...-based implementations for CONF_... queries
7762 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7763 a temporary CONF structure with the data component set to NULL
7764 (which gives segmentation faults in lh_retrieve).
7765 Instead, use NULL for the CONF pointer in CONF_get_string and
7766 CONF_get_number (which may use environment variables) and directly
7767 return NULL from CONF_get_section.
7770 *) Fix potential buffer overrun for EBCDIC.
7773 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7774 keyUsage if basicConstraints absent for a CA.
7777 *) Make SMIME_write_PKCS7() write mail header values with a format that
7778 is more generally accepted (no spaces before the semicolon), since
7779 some programs can't parse those values properly otherwise. Also make
7780 sure BIO's that break lines after each write do not create invalid
7784 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7785 macros previously used would not encode an empty SEQUENCE OF
7786 and break the signature.
7788 [This change does not apply to 0.9.7.]
7790 *) Zero the premaster secret after deriving the master secret in
7794 *) Add some EVP_add_digest_alias registrations (as found in
7795 OpenSSL_add_all_digests()) to SSL_library_init()
7796 aka OpenSSL_add_ssl_algorithms(). This provides improved
7797 compatibility with peers using X.509 certificates
7798 with unconventional AlgorithmIdentifier OIDs.
7801 *) Fix for Irix with NO_ASM.
7802 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7804 *) ./config script fixes.
7805 [Ulf Moeller, Richard Levitte]
7807 *) Fix 'openssl passwd -1'.
7810 *) Change PKCS12_key_gen_asc() so it can cope with non null
7811 terminated strings whose length is passed in the passlen
7812 parameter, for example from PEM callbacks. This was done
7813 by adding an extra length parameter to asc2uni().
7814 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7816 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7817 call failed, free the DSA structure.
7820 *) Fix to uni2asc() to cope with zero length Unicode strings.
7821 These are present in some PKCS#12 files.
7824 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7825 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7826 when writing a 32767 byte record.
7827 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7829 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7830 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7832 (RSA objects have a reference count access to which is protected
7833 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7834 so they are meant to be shared between threads.)
7835 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7836 "Reddie, Steven" <Steven.Reddie@ca.com>]
7838 *) Fix a deadlock in CRYPTO_mem_leaks().
7841 *) Use better test patterns in bntest.
7844 *) rand_win.c fix for Borland C.
7847 *) BN_rshift bugfix for n == 0.
7850 *) Add a 'bctest' script that checks for some known 'bc' bugs
7851 so that 'make test' does not abort just because 'bc' is broken.
7854 *) Store verify_result within SSL_SESSION also for client side to
7855 avoid potential security hole. (Re-used sessions on the client side
7856 always resulted in verify_result==X509_V_OK, not using the original
7857 result of the server certificate verification.)
7860 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7861 SSL3_RT_APPLICATION_DATA, return 0.
7862 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7866 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7867 releases, have been re-implemented by renaming the previous
7868 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7869 and ssl3_read_internal, respectively, and adding 'peek' parameters
7870 to them. The new ssl[23]_{read,peek} functions are calls to
7871 ssl[23]_read_internal with the 'peek' flag set appropriately.
7872 A 'peek' parameter has also been added to ssl3_read_bytes, which
7873 does the actual work for ssl3_read_internal.
7876 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7877 the method-specific "init()" handler. Also clean up ex_data after
7878 calling the method-specific "finish()" handler. Previously, this was
7879 happening the other way round.
7882 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7883 The previous value, 12, was not always sufficient for BN_mod_exp().
7886 *) Make sure that shared libraries get the internal name engine with
7887 the full version number and not just 0. This should mark the
7888 shared libraries as not backward compatible. Of course, this should
7889 be changed again when we can guarantee backward binary compatibility.
7892 *) Fix typo in get_cert_by_subject() in by_dir.c
7893 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7895 *) Rework the system to generate shared libraries:
7897 - Make note of the expected extension for the shared libraries and
7898 if there is a need for symbolic links from for example libcrypto.so.0
7899 to libcrypto.so.0.9.7. There is extended info in Configure for
7902 - Make as few rebuilds of the shared libraries as possible.
7904 - Still avoid linking the OpenSSL programs with the shared libraries.
7906 - When installing, install the shared libraries separately from the
7910 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7912 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7913 and not in SSL_clear because the latter is also used by the
7914 accept/connect functions; previously, the settings made by
7915 SSL_set_read_ahead would be lost during the handshake.
7916 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7918 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7919 Previously, it would create entries for disableed algorithms no
7923 *) Added several new manual pages for SSL_* function.
7926 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7928 *) In ssl23_get_client_hello, generate an error message when faced
7929 with an initial SSL 3.0/TLS record that is too small to contain the
7930 first two bytes of the ClientHello message, i.e. client_version.
7931 (Note that this is a pathologic case that probably has never happened
7932 in real life.) The previous approach was to use the version number
7933 from the record header as a substitute; but our protocol choice
7934 should not depend on that one because it is not authenticated
7935 by the Finished messages.
7938 *) More robust randomness gathering functions for Windows.
7939 [Jeffrey Altman <jaltman@columbia.edu>]
7941 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7942 not set then we don't setup the error code for issuer check errors
7943 to avoid possibly overwriting other errors which the callback does
7944 handle. If an application does set the flag then we assume it knows
7945 what it is doing and can handle the new informational codes
7949 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7950 a general "ANY" type, as such it should be able to decode anything
7951 including tagged types. However it didn't check the class so it would
7952 wrongly interpret tagged types in the same way as their universal
7953 counterpart and unknown types were just rejected. Changed so that the
7954 tagged and unknown types are handled in the same way as a SEQUENCE:
7955 that is the encoding is stored intact. There is also a new type
7956 "V_ASN1_OTHER" which is used when the class is not universal, in this
7957 case we have no idea what the actual type is so we just lump them all
7961 *) On VMS, stdout may very well lead to a file that is written to
7962 in a record-oriented fashion. That means that every write() will
7963 write a separate record, which will be read separately by the
7964 programs trying to read from it. This can be very confusing.
7966 The solution is to put a BIO filter in the way that will buffer
7967 text until a linefeed is reached, and then write everything a
7968 line at a time, so every record written will be an actual line,
7969 not chunks of lines and not (usually doesn't happen, but I've
7970 seen it once) several lines in one record. BIO_f_linebuffer() is
7973 Currently, it's a VMS-only method, because that's where it has
7974 been tested well enough.
7977 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7978 it can return incorrect results.
7979 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7980 but it was in 0.9.6-beta[12].)
7983 *) Disable the check for content being present when verifying detached
7984 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7985 include zero length content when signing messages.
7988 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7989 BIO_ctrl (for BIO pairs).
7992 *) Add DSO method for VMS.
7995 *) Bug fix: Montgomery multiplication could produce results with the
7999 *) Add RPM specification openssl.spec and modify it to build three
8000 packages. The default package contains applications, application
8001 documentation and run-time libraries. The devel package contains
8002 include files, static libraries and function documentation. The
8003 doc package contains the contents of the doc directory. The original
8004 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8007 *) Add a large number of documentation files for many SSL routines.
8008 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8010 *) Add a configuration entry for Sony News 4.
8011 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8013 *) Don't set the two most significant bits to one when generating a
8014 random number < q in the DSA library.
8017 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8018 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8019 the underlying transport is blocking) if a handshake took place.
8020 (The default behaviour is needed by applications such as s_client
8021 and s_server that use select() to determine when to use SSL_read;
8022 but for applications that know in advance when to expect data, it
8023 just makes things more complicated.)
8026 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8030 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8031 work better on such systems.
8032 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8034 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8035 Update PKCS12_parse() so it copies the friendlyName and the
8036 keyid to the certificates aux info.
8039 *) Fix bug in PKCS7_verify() which caused an infinite loop
8040 if there was more than one signature.
8041 [Sven Uszpelkat <su@celocom.de>]
8043 *) Major change in util/mkdef.pl to include extra information
8044 about each symbol, as well as presentig variables as well
8045 as functions. This change means that there's n more need
8046 to rebuild the .num files when some algorithms are excluded.
8049 *) Allow the verify time to be set by an application,
8050 rather than always using the current time.
8053 *) Phase 2 verify code reorganisation. The certificate
8054 verify code now looks up an issuer certificate by a
8055 number of criteria: subject name, authority key id
8056 and key usage. It also verifies self signed certificates
8057 by the same criteria. The main comparison function is
8058 X509_check_issued() which performs these checks.
8060 Lot of changes were necessary in order to support this
8061 without completely rewriting the lookup code.
8063 Authority and subject key identifier are now cached.
8065 The LHASH 'certs' is X509_STORE has now been replaced
8066 by a STACK_OF(X509_OBJECT). This is mainly because an
8067 LHASH can't store or retrieve multiple objects with
8068 the same hash value.
8070 As a result various functions (which were all internal
8071 use only) have changed to handle the new X509_STORE
8072 structure. This will break anything that messed round
8073 with X509_STORE internally.
8075 The functions X509_STORE_add_cert() now checks for an
8076 exact match, rather than just subject name.
8078 The X509_STORE API doesn't directly support the retrieval
8079 of multiple certificates matching a given criteria, however
8080 this can be worked round by performing a lookup first
8081 (which will fill the cache with candidate certificates)
8082 and then examining the cache for matches. This is probably
8083 the best we can do without throwing out X509_LOOKUP
8084 entirely (maybe later...).
8086 The X509_VERIFY_CTX structure has been enhanced considerably.
8088 All certificate lookup operations now go via a get_issuer()
8089 callback. Although this currently uses an X509_STORE it
8090 can be replaced by custom lookups. This is a simple way
8091 to bypass the X509_STORE hackery necessary to make this
8092 work and makes it possible to use more efficient techniques
8093 in future. A very simple version which uses a simple
8094 STACK for its trusted certificate store is also provided
8095 using X509_STORE_CTX_trusted_stack().
8097 The verify_cb() and verify() callbacks now have equivalents
8098 in the X509_STORE_CTX structure.
8100 X509_STORE_CTX also has a 'flags' field which can be used
8101 to customise the verify behaviour.
8104 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8105 excludes S/MIME capabilities.
8108 *) When a certificate request is read in keep a copy of the
8109 original encoding of the signed data and use it when outputing
8110 again. Signatures then use the original encoding rather than
8111 a decoded, encoded version which may cause problems if the
8112 request is improperly encoded.
8115 *) For consistency with other BIO_puts implementations, call
8116 buffer_write(b, ...) directly in buffer_puts instead of calling
8119 In BIO_puts, increment b->num_write as in BIO_write.
8120 [Peter.Sylvester@EdelWeb.fr]
8122 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8123 BN_zero, we may not return a BIGNUM with an array consisting of
8127 *) Avoid calling abort() from within the library when problems are
8128 detected, except if preprocessor symbols have been defined
8129 (such as REF_CHECK, BN_DEBUG etc.).
8132 *) New openssl application 'rsautl'. This utility can be
8133 used for low level RSA operations. DER public key
8134 BIO/fp routines also added.
8137 *) New Configure entry and patches for compiling on QNX 4.
8138 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8140 *) A demo state-machine implementation was sponsored by
8141 Nuron (http://www.nuron.com/) and is now available in
8142 demos/state_machine.
8145 *) New options added to the 'dgst' utility for signature
8146 generation and verification.
8149 *) Unrecognized PKCS#7 content types are now handled via a
8150 catch all ASN1_TYPE structure. This allows unsupported
8151 types to be stored as a "blob" and an application can
8152 encode and decode it manually.
8155 *) Fix various signed/unsigned issues to make a_strex.c
8157 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8159 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8160 length if passed a buffer. ASN1_INTEGER_to_BN failed
8161 if passed a NULL BN and its argument was negative.
8162 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8164 *) Modification to PKCS#7 encoding routines to output definite
8165 length encoding. Since currently the whole structures are in
8166 memory there's not real point in using indefinite length
8167 constructed encoding. However if OpenSSL is compiled with
8168 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8171 *) Added BIO_vprintf() and BIO_vsnprintf().
8174 *) Added more prefixes to parse for in the the strings written
8175 through a logging bio, to cover all the levels that are available
8176 through syslog. The prefixes are now:
8178 PANIC, EMERG, EMR => LOG_EMERG
8179 ALERT, ALR => LOG_ALERT
8180 CRIT, CRI => LOG_CRIT
8181 ERROR, ERR => LOG_ERR
8182 WARNING, WARN, WAR => LOG_WARNING
8183 NOTICE, NOTE, NOT => LOG_NOTICE
8184 INFO, INF => LOG_INFO
8185 DEBUG, DBG => LOG_DEBUG
8187 and as before, if none of those prefixes are present at the
8188 beginning of the string, LOG_ERR is chosen.
8190 On Win32, the LOG_* levels are mapped according to this:
8192 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8193 LOG_WARNING => EVENTLOG_WARNING_TYPE
8194 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8198 *) Made it possible to reconfigure with just the configuration
8199 argument "reconf" or "reconfigure". The command line arguments
8200 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8201 and are retrieved from there when reconfiguring.
8205 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8207 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8210 *) The obj_dat.pl script was messing up the sorting of object
8211 names. The reason was that it compared the quoted version
8212 of strings as a result "OCSP" > "OCSP Signing" because
8213 " > SPACE. Changed script to store unquoted versions of
8214 names and add quotes on output. It was also omitting some
8215 names from the lookup table if they were given a default
8216 value (that is if SN is missing it is given the same
8217 value as LN and vice versa), these are now added on the
8218 grounds that if an object has a name we should be able to
8219 look it up. Finally added warning output when duplicate
8220 short or long names are found.
8223 *) Changes needed for Tandem NSK.
8224 [Scott Uroff <scott@xypro.com>]
8226 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8227 RSA_padding_check_SSLv23(), special padding was never detected
8228 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8229 version rollback attacks was not effective.
8231 In s23_clnt.c, don't use special rollback-attack detection padding
8232 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8233 client; similarly, in s23_srvr.c, don't do the rollback check if
8234 SSL 2.0 is the only protocol enabled in the server.
8237 *) Make it possible to get hexdumps of unprintable data with 'openssl
8238 asn1parse'. By implication, the functions ASN1_parse_dump() and
8239 BIO_dump_indent() are added.
8242 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8243 these print out strings and name structures based on various
8244 flags including RFC2253 support and proper handling of
8245 multibyte characters. Added options to the 'x509' utility
8246 to allow the various flags to be set.
8249 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8250 Also change the functions X509_cmp_current_time() and
8251 X509_gmtime_adj() work with an ASN1_TIME structure,
8252 this will enable certificates using GeneralizedTime in validity
8253 dates to be checked.
8256 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8257 negative public key encodings) on by default,
8258 NO_NEG_PUBKEY_BUG can be set to disable it.
8261 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8262 content octets. An i2c_ASN1_OBJECT is unnecessary because
8263 the encoding can be trivially obtained from the structure.
8266 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8267 not read locks (CRYPTO_r_[un]lock).
8270 *) A first attempt at creating official support for shared
8271 libraries through configuration. I've kept it so the
8272 default is static libraries only, and the OpenSSL programs
8273 are always statically linked for now, but there are
8274 preparations for dynamic linking in place.
8275 This has been tested on Linux and Tru64.
8278 *) Randomness polling function for Win9x, as described in:
8279 Peter Gutmann, Software Generation of Practically Strong
8283 *) Fix so PRNG is seeded in req if using an already existing
8287 *) New options to smime application. -inform and -outform
8288 allow alternative formats for the S/MIME message including
8289 PEM and DER. The -content option allows the content to be
8290 specified separately. This should allow things like Netscape
8291 form signing output easier to verify.
8294 *) Fix the ASN1 encoding of tags using the 'long form'.
8297 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8298 STRING types. These convert content octets to and from the
8299 underlying type. The actual tag and length octets are
8300 already assumed to have been read in and checked. These
8301 are needed because all other string types have virtually
8302 identical handling apart from the tag. By having versions
8303 of the ASN1 functions that just operate on content octets
8304 IMPLICIT tagging can be handled properly. It also allows
8305 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8306 and ASN1_INTEGER are identical apart from the tag.
8309 *) Change the handling of OID objects as follows:
8311 - New object identifiers are inserted in objects.txt, following
8312 the syntax given in objects.README.
8313 - objects.pl is used to process obj_mac.num and create a new
8315 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8318 This is currently kind of a hack, and the perl code in objects.pl
8319 isn't very elegant, but it works as I intended. The simplest way
8320 to check that it worked correctly is to look in obj_dat.h and
8321 check the array nid_objs and make sure the objects haven't moved
8322 around (this is important!). Additions are OK, as well as
8323 consistent name changes.
8326 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8329 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8330 The given file adds to whatever has already been seeded into the
8331 random pool through the RANDFILE configuration file option or
8332 environment variable, or the default random state file.
8335 *) mkstack.pl now sorts each macro group into lexical order.
8336 Previously the output order depended on the order the files
8337 appeared in the directory, resulting in needless rewriting
8341 *) Patches to make OpenSSL compile under Win32 again. Mostly
8342 work arounds for the VC++ problem that it treats func() as
8343 func(void). Also stripped out the parts of mkdef.pl that
8344 added extra typesafe functions: these no longer exist.
8347 *) Reorganisation of the stack code. The macros are now all
8348 collected in safestack.h . Each macro is defined in terms of
8349 a "stack macro" of the form SKM_<name>(type, a, b). The
8350 DEBUG_SAFESTACK is now handled in terms of function casts,
8351 this has the advantage of retaining type safety without the
8352 use of additional functions. If DEBUG_SAFESTACK is not defined
8353 then the non typesafe macros are used instead. Also modified the
8354 mkstack.pl script to handle the new form. Needs testing to see
8355 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8356 the default if no major problems. Similar behaviour for ASN1_SET_OF
8357 and PKCS12_STACK_OF.
8360 *) When some versions of IIS use the 'NET' form of private key the
8361 key derivation algorithm is different. Normally MD5(password) is
8362 used as a 128 bit RC4 key. In the modified case
8363 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
8364 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8365 as the old Netscape_RSA functions except they have an additional
8366 'sgckey' parameter which uses the modified algorithm. Also added
8367 an -sgckey command line option to the rsa utility. Thanks to
8368 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8369 algorithm to openssl-dev.
8372 *) The evp_local.h macros were using 'c.##kname' which resulted in
8373 invalid expansion on some systems (SCO 5.0.5 for example).
8374 Corrected to 'c.kname'.
8375 [Phillip Porch <root@theporch.com>]
8377 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8378 a STACK of email addresses from a certificate or request, these look
8379 in the subject name and the subject alternative name extensions and
8380 omit any duplicate addresses.
8383 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8384 This makes DSA verification about 2 % faster.
8387 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8388 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8389 plus overhead for 1024 bit moduli).
8390 This makes exponentiations about 0.5 % faster for 1024 bit
8391 exponents (as measured by "openssl speed rsa2048").
8394 *) Rename memory handling macros to avoid conflicts with other
8396 Malloc => OPENSSL_malloc
8397 Malloc_locked => OPENSSL_malloc_locked
8398 Realloc => OPENSSL_realloc
8399 Free => OPENSSL_free
8402 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8403 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8406 *) CygWin32 support.
8407 [John Jarvie <jjarvie@newsguy.com>]
8409 *) The type-safe stack code has been rejigged. It is now only compiled
8410 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8411 by default all type-specific stack functions are "#define"d back to
8412 standard stack functions. This results in more streamlined output
8413 but retains the type-safety checking possibilities of the original
8417 *) The STACK code has been cleaned up, and certain type declarations
8418 that didn't make a lot of sense have been brought in line. This has
8419 also involved a cleanup of sorts in safestack.h to more correctly
8420 map type-safe stack functions onto their plain stack counterparts.
8421 This work has also resulted in a variety of "const"ifications of
8422 lots of the code, especially "_cmp" operations which should normally
8423 be prototyped with "const" parameters anyway.
8426 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8427 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8428 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8429 where all of 'md' is used each time the PRNG is used, but 'state'
8430 is used only indexed by a cyclic counter. As entropy may not be
8431 well distributed from the beginning, 'md' is important as a
8432 chaining variable. However, the output function chains only half
8433 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8434 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8435 in all of 'state' being rewritten, with the new values depending
8436 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8439 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8440 the handshake is continued after ssl_verify_cert_chain();
8441 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8442 can lead to 'unexplainable' connection aborts later.
8443 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8445 *) Major EVP API cipher revision.
8446 Add hooks for extra EVP features. This allows various cipher
8447 parameters to be set in the EVP interface. Support added for variable
8448 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8449 setting of RC2 and RC5 parameters.
8451 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8454 Remove lots of duplicated code from the EVP library. For example *every*
8455 cipher init() function handles the 'iv' in the same way according to the
8456 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8457 for CFB and OFB modes they zero ctx->num.
8459 New functionality allows removal of S/MIME code RC2 hack.
8461 Most of the routines have the same form and so can be declared in terms
8464 By shifting this to the top level EVP_CipherInit() it can be removed from
8465 all individual ciphers. If the cipher wants to handle IVs or keys
8466 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8469 Change lots of functions like EVP_EncryptUpdate() to now return a
8470 value: although software versions of the algorithms cannot fail
8471 any installed hardware versions can.
8474 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8475 this option is set, tolerate broken clients that send the negotiated
8476 protocol version number instead of the requested protocol version
8480 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8481 i.e. non-zero for export ciphersuites, zero otherwise.
8482 Previous versions had this flag inverted, inconsistent with
8483 rsa_tmp_cb (..._TMP_RSA_CB).
8484 [Bodo Moeller; problem reported by Amit Chopra]
8486 *) Add missing DSA library text string. Work around for some IIS
8487 key files with invalid SEQUENCE encoding.
8490 *) Add a document (doc/standards.txt) that list all kinds of standards
8491 and so on that are implemented in OpenSSL.
8494 *) Enhance c_rehash script. Old version would mishandle certificates
8495 with the same subject name hash and wouldn't handle CRLs at all.
8496 Added -fingerprint option to crl utility, to support new c_rehash
8500 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8503 *) Fix for SSL server purpose checking. Server checking was
8504 rejecting certificates which had extended key usage present
8505 but no ssl client purpose.
8506 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8508 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8509 is a little unclear about how a blank password is handled.
8510 Since the password in encoded as a BMPString with terminating
8511 double NULL a zero length password would end up as just the
8512 double NULL. However no password at all is different and is
8513 handled differently in the PKCS#12 key generation code. NS
8514 treats a blank password as zero length. MSIE treats it as no
8515 password on export: but it will try both on import. We now do
8516 the same: PKCS12_parse() tries zero length and no password if
8517 the password is set to "" or NULL (NULL is now a valid password:
8518 it wasn't before) as does the pkcs12 application.
8521 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8522 perror when PEM_read_bio_X509_REQ fails, the error message must
8523 be obtained from the error queue.
8526 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8527 it in ERR_remove_state if appropriate, and change ERR_get_state
8528 accordingly to avoid race conditions (this is necessary because
8529 thread_hash is no longer constant once set).
8532 *) Bugfix for linux-elf makefile.one.
8535 *) RSA_get_default_method() will now cause a default
8536 RSA_METHOD to be chosen if one doesn't exist already.
8537 Previously this was only set during a call to RSA_new()
8538 or RSA_new_method(NULL) meaning it was possible for
8539 RSA_get_default_method() to return NULL.
8542 *) Added native name translation to the existing DSO code
8543 that will convert (if the flag to do so is set) filenames
8544 that are sufficiently small and have no path information
8545 into a canonical native form. Eg. "blah" converted to
8546 "libblah.so" or "blah.dll" etc.
8549 *) New function ERR_error_string_n(e, buf, len) which is like
8550 ERR_error_string(e, buf), but writes at most 'len' bytes
8551 including the 0 terminator. For ERR_error_string_n, 'buf'
8553 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8555 *) CONF library reworked to become more general. A new CONF
8556 configuration file reader "class" is implemented as well as a
8557 new functions (NCONF_*, for "New CONF") to handle it. The now
8558 old CONF_* functions are still there, but are reimplemented to
8559 work in terms of the new functions. Also, a set of functions
8560 to handle the internal storage of the configuration data is
8561 provided to make it easier to write new configuration file
8562 reader "classes" (I can definitely see something reading a
8563 configuration file in XML format, for example), called _CONF_*,
8564 or "the configuration storage API"...
8566 The new configuration file reading functions are:
8568 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8569 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8571 NCONF_default, NCONF_WIN32
8573 NCONF_dump_fp, NCONF_dump_bio
8575 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8576 NCONF_new creates a new CONF object. This works in the same way
8577 as other interfaces in OpenSSL, like the BIO interface.
8578 NCONF_dump_* dump the internal storage of the configuration file,
8579 which is useful for debugging. All other functions take the same
8580 arguments as the old CONF_* functions wth the exception of the
8581 first that must be a `CONF *' instead of a `LHASH *'.
8583 To make it easer to use the new classes with the old CONF_* functions,
8584 the function CONF_set_default_method is provided.
8587 *) Add '-tls1' option to 'openssl ciphers', which was already
8588 mentioned in the documentation but had not been implemented.
8589 (This option is not yet really useful because even the additional
8590 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8593 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8594 OpenSSL-based applications) load shared libraries and bind to
8595 them in a portable way.
8596 [Geoff Thorpe, with contributions from Richard Levitte]
8598 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8600 *) Make sure _lrotl and _lrotr are only used with MSVC.
8602 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8603 (the default implementation of RAND_status).
8605 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8606 to '-clrext' (= clear extensions), as intended and documented.
8607 [Bodo Moeller; inconsistency pointed out by Michael Attili
8610 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8611 was larger than the MD block size.
8612 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8614 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8615 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8616 using the passed key: if the passed key was a private key the result
8617 of X509_print(), for example, would be to print out all the private key
8621 *) des_quad_cksum() byte order bug fix.
8622 [Ulf Möller, using the problem description in krb4-0.9.7, where
8623 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8625 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8627 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8629 *) For easily testing in shell scripts whether some command
8630 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8631 returns with exit code 0 iff no command of the given name is available.
8632 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8633 the output goes to stdout and nothing is printed to stderr.
8634 Additional arguments are always ignored.
8636 Since for each cipher there is a command of the same name,
8637 the 'no-cipher' compilation switches can be tested this way.
8639 ('openssl no-XXX' is not able to detect pseudo-commands such
8640 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8643 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8646 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8647 is set; it will be thrown away anyway because each handshake creates
8649 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8650 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8651 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
8652 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8655 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8656 'Q' and 'R' lose their special meanings (quit/renegotiate).
8657 This is part of what -quiet does; unlike -quiet, -ign_eof
8658 does not suppress any output.
8661 *) Add compatibility options to the purpose and trust code. The
8662 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8663 accepts a certificate or CA, this was the previous behaviour,
8664 with all the associated security issues.
8666 X509_TRUST_COMPAT is the old trust behaviour: only and
8667 automatically trust self signed roots in certificate store. A
8668 new trust setting X509_TRUST_DEFAULT is used to specify that
8669 a purpose has no associated trust setting and it should instead
8670 use the value in the default purpose.
8673 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8674 and fix a memory leak.
8677 *) In util/mkerr.pl (which implements 'make errors'), preserve
8678 reason strings from the previous version of the .c file, as
8679 the default to have only downcase letters (and digits) in
8680 automatically generated reasons codes is not always appropriate.
8683 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8684 using strerror. Previously, ERR_reason_error_string() returned
8685 library names as reason strings for SYSerr; but SYSerr is a special
8686 case where small numbers are errno values, not library numbers.
8689 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8690 converts DSA parameters into DH parameters. (When creating parameters,
8691 DSA_generate_parameters is used.)
8694 *) Include 'length' (recommended exponent length) in C code generated
8695 by 'openssl dhparam -C'.
8698 *) The second argument to set_label in perlasm was already being used
8699 so couldn't be used as a "file scope" flag. Moved to third argument
8703 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8704 instead of RAND_bytes for encryption IVs and salts.
8707 *) Include RAND_status() into RAND_METHOD instead of implementing
8708 it only for md_rand.c Otherwise replacing the PRNG by calling
8709 RAND_set_rand_method would be impossible.
8712 *) Don't let DSA_generate_key() enter an infinite loop if the random
8713 number generation fails.
8716 *) New 'rand' application for creating pseudo-random output.
8719 *) Added configuration support for Linux/IA64
8720 [Rolf Haberrecker <rolf@suse.de>]
8722 *) Assembler module support for Mingw32.
8725 *) Shared library support for HPUX (in shlib/).
8726 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8728 *) Shared library support for Solaris gcc.
8729 [Lutz Behnke <behnke@trustcenter.de>]
8731 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8733 *) PKCS7_encrypt() was adding text MIME headers twice because they
8734 were added manually and by SMIME_crlf_copy().
8737 *) In bntest.c don't call BN_rand with zero bits argument.
8738 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8740 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8741 case was implemented. This caused BN_div_recp() to fail occasionally.
8744 *) Add an optional second argument to the set_label() in the perl
8745 assembly language builder. If this argument exists and is set
8746 to 1 it signals that the assembler should use a symbol whose
8747 scope is the entire file, not just the current function. This
8748 is needed with MASM which uses the format label:: for this scope.
8749 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8751 *) Change the ASN1 types so they are typedefs by default. Before
8752 almost all types were #define'd to ASN1_STRING which was causing
8753 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8757 *) Change names of new functions to the new get1/get0 naming
8758 convention: After 'get1', the caller owns a reference count
8759 and has to call ..._free; 'get0' returns a pointer to some
8760 data structure without incrementing reference counters.
8761 (Some of the existing 'get' functions increment a reference
8762 counter, some don't.)
8763 Similarly, 'set1' and 'add1' functions increase reference
8764 counters or duplicate objects.
8767 *) Allow for the possibility of temp RSA key generation failure:
8768 the code used to assume it always worked and crashed on failure.
8771 *) Fix potential buffer overrun problem in BIO_printf().
8772 [Ulf Möller, using public domain code by Patrick Powell; problem
8773 pointed out by David Sacerdote <das33@cornell.edu>]
8775 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8776 RAND_egd() and RAND_status(). In the command line application,
8777 the EGD socket can be specified like a seed file using RANDFILE
8781 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8782 Some CAs (e.g. Verisign) distribute certificates in this form.
8785 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8786 list to exclude them. This means that no special compilation option
8787 is needed to use anonymous DH: it just needs to be included in the
8791 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8792 EVP_MD_type. The old functionality is available in a new macro called
8793 EVP_MD_md(). Change code that uses it and update docs.
8796 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8797 where the 'void *' argument is replaced by a function pointer argument.
8798 Previously 'void *' was abused to point to functions, which works on
8799 many platforms, but is not correct. As these functions are usually
8800 called by macros defined in OpenSSL header files, most source code
8801 should work without changes.
8804 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8805 sections with information on -D... compiler switches used for
8806 compiling the library so that applications can see them. To enable
8807 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8808 must be defined. E.g.,
8809 #define OPENSSL_ALGORITHM_DEFINES
8810 #include <openssl/opensslconf.h>
8811 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8812 [Richard Levitte, Ulf and Bodo Möller]
8814 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8818 *) Change the 'other' type in certificate aux info to a STACK_OF
8819 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8820 the required ASN1 format: arbitrary types determined by an OID.
8823 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8824 argument to 'req'. This is not because the function is newer or
8825 better than others it just uses the work 'NEW' in the certificate
8826 request header lines. Some software needs this.
8829 *) Reorganise password command line arguments: now passwords can be
8830 obtained from various sources. Delete the PEM_cb function and make
8831 it the default behaviour: i.e. if the callback is NULL and the
8832 usrdata argument is not NULL interpret it as a null terminated pass
8833 phrase. If usrdata and the callback are NULL then the pass phrase
8834 is prompted for as usual.
8837 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8838 the support is automatically enabled. The resulting binaries will
8839 autodetect the card and use it if present.
8840 [Ben Laurie and Compaq Inc.]
8842 *) Work around for Netscape hang bug. This sends certificate request
8843 and server done in one record. Since this is perfectly legal in the
8844 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8845 the bugs/SSLv3 entry for more info.
8848 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8851 *) Add -rand argument to smime and pkcs12 applications and read/write
8855 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8858 *) Add command line password options to the remaining applications.
8861 *) Bug fix for BN_div_recp() for numerators with an even number of
8865 *) More tests in bntest.c, and changed test_bn output.
8868 *) ./config recognizes MacOS X now.
8871 *) Bug fix for BN_div() when the first words of num and divsor are
8872 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8875 *) Add support for various broken PKCS#8 formats, and command line
8876 options to produce them.
8879 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8880 get temporary BIGNUMs from a BN_CTX.
8883 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8887 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8888 include a #define from the old name to the new. The original intent
8889 was that statically linked binaries could for example just call
8890 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8891 link with digests. This never worked becayse SSLeay_add_all_digests()
8892 and SSLeay_add_all_ciphers() were in the same source file so calling
8893 one would link with the other. They are now in separate source files.
8896 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8899 *) Use a less unusual form of the Miller-Rabin primality test (it used
8900 a binary algorithm for exponentiation integrated into the Miller-Rabin
8901 loop, our standard modexp algorithms are faster).
8904 *) Support for the EBCDIC character set completed.
8905 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8907 *) Source code cleanups: use const where appropriate, eliminate casts,
8908 use void * instead of char * in lhash.
8911 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8912 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8913 this the server could overwrite ephemeral keys that the client
8917 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8918 using 50 iterations of the Rabin-Miller test.
8920 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8921 iterations of the Rabin-Miller test as required by the appendix
8922 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8923 As BN_is_prime_fasttest includes trial division, DSA parameter
8924 generation becomes much faster.
8926 This implies a change for the callback functions in DSA_is_prime
8927 and DSA_generate_parameters: The callback function is called once
8928 for each positive witness in the Rabin-Miller test, not just
8929 occasionally in the inner loop; and the parameters to the
8930 callback function now provide an iteration count for the outer
8931 loop rather than for the current invocation of the inner loop.
8932 DSA_generate_parameters additionally can call the callback
8933 function with an 'iteration count' of -1, meaning that a
8934 candidate has passed the trial division test (when q is generated
8935 from an application-provided seed, trial division is skipped).
8938 *) New function BN_is_prime_fasttest that optionally does trial
8939 division before starting the Rabin-Miller test and has
8940 an additional BN_CTX * argument (whereas BN_is_prime always
8941 has to allocate at least one BN_CTX).
8942 'callback(1, -1, cb_arg)' is called when a number has passed the
8943 trial division stage.
8946 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8950 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8953 *) New function BN_pseudo_rand().
8956 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8957 bignum version of BN_from_montgomery() with the working code from
8958 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8962 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8963 made it impossible to use the same SSL_SESSION data structure in
8964 SSL2 clients in multiple threads.
8967 *) The return value of RAND_load_file() no longer counts bytes obtained
8968 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8969 to seed the PRNG (previously an explicit byte count was required).
8970 [Ulf Möller, Bodo Möller]
8972 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8973 used (char *) instead of (void *) and had casts all over the place.
8976 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8979 *) Retain source code compatibility for BN_prime_checks macro:
8980 BN_is_prime(..., BN_prime_checks, ...) now uses
8981 BN_prime_checks_for_size to determine the appropriate number of
8982 Rabin-Miller iterations.
8985 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8986 DH_CHECK_P_NOT_SAFE_PRIME.
8987 (Check if this is true? OpenPGP calls them "strong".)
8990 *) Merge the functionality of "dh" and "gendh" programs into a new program
8991 "dhparam". The old programs are retained for now but will handle DH keys
8992 (instead of parameters) in future.
8995 *) Make the ciphers, s_server and s_client programs check the return values
8996 when a new cipher list is set.
8999 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9000 ciphers. Before when the 56bit ciphers were enabled the sorting was
9003 The syntax for the cipher sorting has been extended to support sorting by
9004 cipher-strength (using the strength_bits hard coded in the tables).
9005 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9007 Fix a bug in the cipher-command parser: when supplying a cipher command
9008 string with an "undefined" symbol (neither command nor alphanumeric
9009 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9010 an error is flagged.
9012 Due to the strength-sorting extension, the code of the
9013 ssl_create_cipher_list() function was completely rearranged. I hope that
9014 the readability was also increased :-)
9015 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9017 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9018 for the first serial number and places 2 in the serial number file. This
9019 avoids problems when the root CA is created with serial number zero and
9020 the first user certificate has the same issuer name and serial number
9024 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9025 the new code. Add documentation for this stuff.
9028 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9029 X509_*() to X509at_*() on the grounds that they don't handle X509
9030 structures and behave in an analagous way to the X509v3 functions:
9031 they shouldn't be called directly but wrapper functions should be used
9034 So we also now have some wrapper functions that call the X509at functions
9035 when passed certificate requests. (TO DO: similar things can be done with
9036 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9037 things. Some of these need some d2i or i2d and print functionality
9038 because they handle more complex structures.)
9041 *) Add missing #ifndefs that caused missing symbols when building libssl
9042 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9043 NO_RSA in ssl/s2*.c.
9044 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9046 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9047 has a return value which indicates the quality of the random data
9048 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9049 error queue. New function RAND_pseudo_bytes() generates output that is
9050 guaranteed to be unique but not unpredictable. RAND_add is like
9051 RAND_seed, but takes an extra argument for an entropy estimate
9052 (RAND_seed always assumes full entropy).
9055 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9056 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9057 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9058 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9059 false-positive rate of at most 2^-80 for random input.
9062 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9065 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9066 in the 0.9.5 release), this returns the chain
9067 from an X509_CTX structure with a dup of the stack and all
9068 the X509 reference counts upped: so the stack will exist
9069 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9072 Also make SSL_SESSION_print() print out the verify return
9076 *) Add manpage for the pkcs12 command. Also change the default
9077 behaviour so MAC iteration counts are used unless the new
9078 -nomaciter option is used. This improves file security and
9079 only older versions of MSIE (4.0 for example) need it.
9082 *) Honor the no-xxx Configure options when creating .DEF files.
9085 *) Add PKCS#10 attributes to field table: challengePassword,
9086 unstructuredName and unstructuredAddress. These are taken from
9087 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9088 international characters are used.
9090 More changes to X509_ATTRIBUTE code: allow the setting of types
9091 based on strings. Remove the 'loc' parameter when adding
9092 attributes because these will be a SET OF encoding which is sorted
9096 *) Initial changes to the 'req' utility to allow request generation
9097 automation. This will allow an application to just generate a template
9098 file containing all the field values and have req construct the
9101 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9102 used all over the place including certificate requests and PKCS#7
9103 structures. They are currently handled manually where necessary with
9104 some primitive wrappers for PKCS#7. The new functions behave in a
9105 manner analogous to the X509 extension functions: they allow
9106 attributes to be looked up by NID and added.
9108 Later something similar to the X509V3 code would be desirable to
9109 automatically handle the encoding, decoding and printing of the
9110 more complex types. The string types like challengePassword can
9111 be handled by the string table functions.
9113 Also modified the multi byte string table handling. Now there is
9114 a 'global mask' which masks out certain types. The table itself
9115 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9116 is useful when for example there is only one permissible type
9117 (as in countryName) and using the mask might result in no valid
9121 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9122 SSL_get_peer_finished to allow applications to obtain the latest
9123 Finished messages sent to the peer or expected from the peer,
9124 respectively. (SSL_get_peer_finished is usually the Finished message
9125 actually received from the peer, otherwise the protocol will be aborted.)
9127 As the Finished message are message digests of the complete handshake
9128 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9129 be used for external authentication procedures when the authentication
9130 provided by SSL/TLS is not desired or is not enough.
9133 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9134 the host supports BWX extension and if Compaq C is present on the
9135 $PATH. Just exploiting of the BWX extension results in 20-30%
9136 performance kick for some algorithms, e.g. DES and RC4 to mention
9137 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9141 *) Add support for MS "fast SGC". This is arguably a violation of the
9142 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9143 weak crypto and after checking the certificate is SGC a second one
9144 with strong crypto. MS SGC stops the first handshake after receiving
9145 the server certificate message and sends a second client hello. Since
9146 a server will typically do all the time consuming operations before
9147 expecting any further messages from the client (server key exchange
9148 is the most expensive) there is little difference between the two.
9150 To get OpenSSL to support MS SGC we have to permit a second client
9151 hello message after we have sent server done. In addition we have to
9152 reset the MAC if we do get this second client hello.
9155 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9156 if a DER encoded private key is RSA or DSA traditional format. Changed
9157 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9158 format DER encoded private key. Newer code should use PKCS#8 format which
9159 has the key type encoded in the ASN1 structure. Added DER private key
9160 support to pkcs8 application.
9163 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9164 ciphersuites has been selected (as required by the SSL 3/TLS 1
9165 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9166 is set, we interpret this as a request to violate the specification
9167 (the worst that can happen is a handshake failure, and 'correct'
9168 behaviour would result in a handshake failure anyway).
9171 *) In SSL_CTX_add_session, take into account that there might be multiple
9172 SSL_SESSION structures with the same session ID (e.g. when two threads
9173 concurrently obtain them from an external cache).
9174 The internal cache can handle only one SSL_SESSION with a given ID,
9175 so if there's a conflict, we now throw out the old one to achieve
9179 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9180 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9181 some routines that use cipher OIDs: some ciphers do not have OIDs
9182 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9186 *) Simplify the trust setting structure and code. Now we just have
9187 two sequences of OIDs for trusted and rejected settings. These will
9188 typically have values the same as the extended key usage extension
9189 and any application specific purposes.
9191 The trust checking code now has a default behaviour: it will just
9192 check for an object with the same NID as the passed id. Functions can
9193 be provided to override either the default behaviour or the behaviour
9194 for a given id. SSL client, server and email already have functions
9195 in place for compatibility: they check the NID and also return "trusted"
9196 if the certificate is self signed.
9199 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9200 traditional format into an EVP_PKEY structure.
9203 *) Add a password callback function PEM_cb() which either prompts for
9204 a password if usr_data is NULL or otherwise assumes it is a null
9205 terminated password. Allow passwords to be passed on command line
9206 environment or config files in a few more utilities.
9209 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9210 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9211 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9212 Update documentation.
9215 *) Support for ASN1 "NULL" type. This could be handled before by using
9216 ASN1_TYPE but there wasn't any function that would try to read a NULL
9217 and produce an error if it couldn't. For compatibility we also have
9218 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9219 don't allocate anything because they don't need to.
9222 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9224 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9226 *) Rebuild of the memory allocation routines used by OpenSSL code and
9227 possibly others as well. The purpose is to make an interface that
9228 provide hooks so anyone can build a separate set of allocation and
9229 deallocation routines to be used by OpenSSL, for example memory
9230 pool implementations, or something else, which was previously hard
9231 since Malloc(), Realloc() and Free() were defined as macros having
9232 the values malloc, realloc and free, respectively (except for Win32
9233 compilations). The same is provided for memory debugging code.
9234 OpenSSL already comes with functionality to find memory leaks, but
9235 this gives people a chance to debug other memory problems.
9237 With these changes, a new set of functions and macros have appeared:
9239 CRYPTO_set_mem_debug_functions() [F]
9240 CRYPTO_get_mem_debug_functions() [F]
9241 CRYPTO_dbg_set_options() [F]
9242 CRYPTO_dbg_get_options() [F]
9243 CRYPTO_malloc_debug_init() [M]
9245 The memory debug functions are NULL by default, unless the library
9246 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9247 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9248 gives the standard debugging functions that come with OpenSSL) or
9249 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9250 provided by the library user) must be used. When the standard
9251 debugging functions are used, CRYPTO_dbg_set_options can be used to
9252 request additional information:
9253 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9254 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9256 Also, things like CRYPTO_set_mem_functions will always give the
9257 expected result (the new set of functions is used for allocation
9258 and deallocation) at all times, regardless of platform and compiler
9261 To finish it up, some functions that were never use in any other
9262 way than through macros have a new API and new semantic:
9265 CRYPTO_dbg_realloc()
9268 All macros of value have retained their old syntax.
9269 [Richard Levitte and Bodo Moeller]
9271 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9272 ordering of SMIMECapabilities wasn't in "strength order" and there
9273 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9277 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9278 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9279 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9281 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9282 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9283 functionality to handle multipart/signed properly) and a utility
9284 called 'smime' to call all this stuff. This is based on code I
9285 originally wrote for Celo who have kindly allowed it to be
9286 included in OpenSSL.
9289 *) Add variants des_set_key_checked and des_set_key_unchecked of
9290 des_set_key (aka des_key_sched). Global variable des_check_key
9291 decides which of these is called by des_set_key; this way
9292 des_check_key behaves as it always did, but applications and
9293 the library itself, which was buggy for des_check_key == 1,
9294 have a cleaner way to pick the version they need.
9297 *) New function PKCS12_newpass() which changes the password of a
9301 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9302 dynamic mix. In both cases the ids can be used as an index into the
9303 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9304 functions so they accept a list of the field values and the
9305 application doesn't need to directly manipulate the X509_TRUST
9309 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9313 *) Modify the way the V3 extension code looks up extensions. This now
9314 works in a similar way to the object code: we have some "standard"
9315 extensions in a static table which is searched with OBJ_bsearch()
9316 and the application can add dynamic ones if needed. The file
9317 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9318 updated whenever a new extension is added to the core code and kept
9319 in ext_nid order. There is a simple program 'tabtest.c' which checks
9320 this. New extensions are not added too often so this file can readily
9321 be maintained manually.
9323 There are two big advantages in doing things this way. The extensions
9324 can be looked up immediately and no longer need to be "added" using
9325 X509V3_add_standard_extensions(): this function now does nothing.
9326 [Side note: I get *lots* of email saying the extension code doesn't
9327 work because people forget to call this function]
9328 Also no dynamic allocation is done unless new extensions are added:
9329 so if we don't add custom extensions there is no need to call
9330 X509V3_EXT_cleanup().
9333 *) Modify enc utility's salting as follows: make salting the default. Add a
9334 magic header, so unsalted files fail gracefully instead of just decrypting
9335 to garbage. This is because not salting is a big security hole, so people
9336 should be discouraged from doing it.
9339 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9340 digest to be passed on the command line but it only used this
9341 parameter when signing a certificate. Modified so all relevant
9342 operations are affected by the digest parameter including the
9343 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9344 DSA key was used because it didn't fix the digest.
9347 *) Initial certificate chain verify code. Currently tests the untrusted
9348 certificates for consistency with the verify purpose (which is set
9349 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9351 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9352 this is because it will reject chains with invalid extensions whereas
9353 every previous version of OpenSSL and SSLeay made no checks at all.
9355 Trust code: checks the root CA for the relevant trust settings. Trust
9356 settings have an initial value consistent with the verify purpose: e.g.
9357 if the verify purpose is for SSL client use it expects the CA to be
9358 trusted for SSL client use. However the default value can be changed to
9359 permit custom trust settings: one example of this would be to only trust
9360 certificates from a specific "secure" set of CAs.
9362 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9363 which should be used for version portability: especially since the
9364 verify structure is likely to change more often now.
9366 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9367 to set them. If not set then assume SSL clients will verify SSL servers
9370 Two new options to the verify program: -untrusted allows a set of
9371 untrusted certificates to be passed in and -purpose which sets the
9372 intended purpose of the certificate. If a purpose is set then the
9373 new chain verify code is used to check extension consistency.
9376 *) Support for the authority information access extension.
9379 *) Modify RSA and DSA PEM read routines to transparently handle
9380 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9381 public keys in a format compatible with certificate
9382 SubjectPublicKeyInfo structures. Unfortunately there were already
9383 functions called *_PublicKey_* which used various odd formats so
9384 these are retained for compatibility: however the DSA variants were
9385 never in a public release so they have been deleted. Changed dsa/rsa
9386 utilities to handle the new format: note no releases ever handled public
9387 keys so we should be OK.
9389 The primary motivation for this change is to avoid the same fiasco
9390 that dogs private keys: there are several incompatible private key
9391 formats some of which are standard and some OpenSSL specific and
9392 require various evil hacks to allow partial transparent handling and
9393 even then it doesn't work with DER formats. Given the option anything
9394 other than PKCS#8 should be dumped: but the other formats have to
9395 stay in the name of compatibility.
9397 With public keys and the benefit of hindsight one standard format
9398 is used which works with EVP_PKEY, RSA or DSA structures: though
9399 it clearly returns an error if you try to read the wrong kind of key.
9401 Added a -pubkey option to the 'x509' utility to output the public key.
9402 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9403 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9404 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9405 that do the same as the EVP_PKEY_assign_*() except they up the
9406 reference count of the added key (they don't "swallow" the
9410 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9411 CRLs would fail if the file contained no certificates or no CRLs:
9412 added a new function to read in both types and return the number
9413 read: this means that if none are read it will be an error. The
9414 DER versions of the certificate and CRL reader would always fail
9415 because it isn't possible to mix certificates and CRLs in DER format
9416 without choking one or the other routine. Changed this to just read
9417 a certificate: this is the best we can do. Also modified the code
9418 in apps/verify.c to take notice of return codes: it was previously
9419 attempting to read in certificates from NULL pointers and ignoring
9420 any errors: this is one reason why the cert and CRL reader seemed
9421 to work. It doesn't check return codes from the default certificate
9422 routines: these may well fail if the certificates aren't installed.
9425 *) Code to support otherName option in GeneralName.
9428 *) First update to verify code. Change the verify utility
9429 so it warns if it is passed a self signed certificate:
9430 for consistency with the normal behaviour. X509_verify
9431 has been modified to it will now verify a self signed
9432 certificate if *exactly* the same certificate appears
9433 in the store: it was previously impossible to trust a
9434 single self signed certificate. This means that:
9435 openssl verify ss.pem
9436 now gives a warning about a self signed certificate but
9437 openssl verify -CAfile ss.pem ss.pem
9441 *) For servers, store verify_result in SSL_SESSION data structure
9442 (and add it to external session representation).
9443 This is needed when client certificate verifications fails,
9444 but an application-provided verification callback (set by
9445 SSL_CTX_set_cert_verify_callback) allows accepting the session
9446 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9447 but returns 1): When the session is reused, we have to set
9448 ssl->verify_result to the appropriate error code to avoid
9450 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9452 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9453 case in PKCS7_dataInit() where the signed PKCS7 structure
9454 didn't contain any existing data because it was being created.
9455 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9457 *) Add a salt to the key derivation routines in enc.c. This
9458 forms the first 8 bytes of the encrypted file. Also add a
9459 -S option to allow a salt to be input on the command line.
9462 *) New function X509_cmp(). Oddly enough there wasn't a function
9463 to compare two certificates. We do this by working out the SHA1
9464 hash and comparing that. X509_cmp() will be needed by the trust
9468 *) SSL_get1_session() is like SSL_get_session(), but increments
9469 the reference count in the SSL_SESSION returned.
9470 [Geoff Thorpe <geoff@eu.c2.net>]
9472 *) Fix for 'req': it was adding a null to request attributes.
9473 Also change the X509_LOOKUP and X509_INFO code to handle
9474 certificate auxiliary information.
9477 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9481 *) Add the possibility to add extra information to the memory leak
9482 detecting output, to form tracebacks, showing from where each
9483 allocation was originated: CRYPTO_push_info("constant string") adds
9484 the string plus current file name and line number to a per-thread
9485 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9486 is like calling CYRPTO_pop_info() until the stack is empty.
9487 Also updated memory leak detection code to be multi-thread-safe.
9490 *) Add options -text and -noout to pkcs7 utility and delete the
9491 encryption options which never did anything. Update docs.
9494 *) Add options to some of the utilities to allow the pass phrase
9495 to be included on either the command line (not recommended on
9496 OSes like Unix) or read from the environment. Update the
9497 manpages and fix a few bugs.
9500 *) Add a few manpages for some of the openssl commands.
9503 *) Fix the -revoke option in ca. It was freeing up memory twice,
9504 leaking and not finding already revoked certificates.
9507 *) Extensive changes to support certificate auxiliary information.
9508 This involves the use of X509_CERT_AUX structure and X509_AUX
9509 functions. An X509_AUX function such as PEM_read_X509_AUX()
9510 can still read in a certificate file in the usual way but it
9511 will also read in any additional "auxiliary information". By
9512 doing things this way a fair degree of compatibility can be
9513 retained: existing certificates can have this information added
9514 using the new 'x509' options.
9516 Current auxiliary information includes an "alias" and some trust
9517 settings. The trust settings will ultimately be used in enhanced
9518 certificate chain verification routines: currently a certificate
9519 can only be trusted if it is self signed and then it is trusted
9523 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9524 The problem was that one of the replacement routines had not been working
9525 since SSLeay releases. For now the offending routine has been replaced
9526 with non-optimised assembler. Even so, this now gives around 95%
9527 performance improvement for 1024 bit RSA signs.
9530 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9531 handling. Most clients have the effective key size in bits equal to
9532 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9533 A few however don't do this and instead use the size of the decrypted key
9534 to determine the RC2 key length and the AlgorithmIdentifier to determine
9535 the effective key length. In this case the effective key length can still
9536 be 40 bits but the key length can be 168 bits for example. This is fixed
9537 by manually forcing an RC2 key into the EVP_PKEY structure because the
9538 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9539 the key length and effective key length are equal.
9542 *) Add a bunch of functions that should simplify the creation of
9543 X509_NAME structures. Now you should be able to do:
9544 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9545 and have it automatically work out the correct field type and fill in
9546 the structures. The more adventurous can try:
9547 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9548 and it will (hopefully) work out the correct multibyte encoding.
9551 *) Change the 'req' utility to use the new field handling and multibyte
9552 copy routines. Before the DN field creation was handled in an ad hoc
9553 way in req, ca, and x509 which was rather broken and didn't support
9554 BMPStrings or UTF8Strings. Since some software doesn't implement
9555 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9556 using the dirstring_type option. See the new comment in the default
9557 openssl.cnf for more info.
9560 *) Make crypto/rand/md_rand.c more robust:
9561 - Assure unique random numbers after fork().
9562 - Make sure that concurrent threads access the global counter and
9563 md serializably so that we never lose entropy in them
9564 or use exactly the same state in multiple threads.
9565 Access to the large state is not always serializable because
9566 the additional locking could be a performance killer, and
9567 md should be large enough anyway.
9570 *) New file apps/app_rand.c with commonly needed functionality
9571 for handling the random seed file.
9573 Use the random seed file in some applications that previously did not:
9575 dsaparam -genkey (which also ignored its '-rand' option),
9578 x509 (when signing).
9579 Except on systems with /dev/urandom, it is crucial to have a random
9580 seed file at least for key creation, DSA signing, and for DH exchanges;
9581 for RSA signatures we could do without one.
9583 gendh and gendsa (unlike genrsa) used to read only the first byte
9584 of each file listed in the '-rand' option. The function as previously
9585 found in genrsa is now in app_rand.c and is used by all programs
9586 that support '-rand'.
9589 *) In RAND_write_file, use mode 0600 for creating files;
9590 don't just chmod when it may be too late.
9593 *) Report an error from X509_STORE_load_locations
9594 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9597 *) New function ASN1_mbstring_copy() this copies a string in either
9598 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9599 into an ASN1_STRING type. A mask of permissible types is passed
9600 and it chooses the "minimal" type to use or an error if not type
9604 *) Add function equivalents to the various macros in asn1.h. The old
9605 macros are retained with an M_ prefix. Code inside the library can
9606 use the M_ macros. External code (including the openssl utility)
9607 should *NOT* in order to be "shared library friendly".
9610 *) Add various functions that can check a certificate's extensions
9611 to see if it usable for various purposes such as SSL client,
9612 server or S/MIME and CAs of these types. This is currently
9613 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9614 verification. Also added a -purpose flag to x509 utility to
9615 print out all the purposes.
9618 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9622 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9623 for, obtain and decode and extension and obtain its critical flag.
9624 This allows all the necessary extension code to be handled in a
9625 single function call.
9628 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9629 platforms. See crypto/rc4/rc4_enc.c for further details.
9632 *) New -noout option to asn1parse. This causes no output to be produced
9633 its main use is when combined with -strparse and -out to extract data
9634 from a file (which may not be in ASN.1 format).
9637 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9638 when producing the local key id.
9639 [Richard Levitte <levitte@stacken.kth.se>]
9641 *) New option -dhparam in s_server. This allows a DH parameter file to be
9642 stated explicitly. If it is not stated then it tries the first server
9643 certificate file. The previous behaviour hard coded the filename
9647 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9648 a public key to be input or output. For example:
9649 openssl rsa -in key.pem -pubout -out pubkey.pem
9650 Also added necessary DSA public key functions to handle this.
9653 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9654 in the message. This was handled by allowing
9655 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9656 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9658 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9659 to the end of the strings whereas this didn't. This would cause problems
9660 if strings read with d2i_ASN1_bytes() were later modified.
9661 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9663 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9664 data and it contains EOF it will end up returning an error. This is
9665 caused by input 46 bytes long. The cause is due to the way base64
9666 BIOs find the start of base64 encoded data. They do this by trying a
9667 trial decode on each line until they find one that works. When they
9668 do a flag is set and it starts again knowing it can pass all the
9669 data directly through the decoder. Unfortunately it doesn't reset
9670 the context it uses. This means that if EOF is reached an attempt
9671 is made to pass two EOFs through the context and this causes the
9672 resulting error. This can also cause other problems as well. As is
9673 usual with these problems it takes *ages* to find and the fix is
9674 trivial: move one line.
9675 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9677 *) Ugly workaround to get s_client and s_server working under Windows. The
9678 old code wouldn't work because it needed to select() on sockets and the
9679 tty (for keypresses and to see if data could be written). Win32 only
9680 supports select() on sockets so we select() with a 1s timeout on the
9681 sockets and then see if any characters are waiting to be read, if none
9682 are present then we retry, we also assume we can always write data to
9683 the tty. This isn't nice because the code then blocks until we've
9684 received a complete line of data and it is effectively polling the
9685 keyboard at 1s intervals: however it's quite a bit better than not
9686 working at all :-) A dedicated Windows application might handle this
9687 with an event loop for example.
9690 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9691 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9692 will be called when RSA_sign() and RSA_verify() are used. This is useful
9693 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9694 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9695 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9696 This necessitated the support of an extra signature type NID_md5_sha1
9697 for SSL signatures and modifications to the SSL library to use it instead
9698 of calling RSA_public_decrypt() and RSA_private_encrypt().
9701 *) Add new -verify -CAfile and -CApath options to the crl program, these
9702 will lookup a CRL issuers certificate and verify the signature in a
9703 similar way to the verify program. Tidy up the crl program so it
9704 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9705 less strict. It will now permit CRL extensions even if it is not
9706 a V2 CRL: this will allow it to tolerate some broken CRLs.
9709 *) Initialize all non-automatic variables each time one of the openssl
9710 sub-programs is started (this is necessary as they may be started
9711 multiple times from the "OpenSSL>" prompt).
9712 [Lennart Bang, Bodo Moeller]
9714 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9715 removing all other RSA functionality (this is what NO_RSA does). This
9716 is so (for example) those in the US can disable those operations covered
9717 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9721 *) Non-copying interface to BIO pairs.
9722 (still largely untested)
9725 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9726 ASCII string. This was handled independently in various places before.
9729 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9730 UTF8 strings a character at a time.
9733 *) Use client_version from client hello to select the protocol
9734 (s23_srvr.c) and for RSA client key exchange verification
9735 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9738 *) Add various utility functions to handle SPKACs, these were previously
9739 handled by poking round in the structure internals. Added new function
9740 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9741 print, verify and generate SPKACs. Based on an original idea from
9742 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9745 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9748 *) Allow the config file extension section to be overwritten on the
9749 command line. Based on an original idea from Massimiliano Pala
9750 <madwolf@comune.modena.it>. The new option is called -extensions
9751 and can be applied to ca, req and x509. Also -reqexts to override
9752 the request extensions in req and -crlexts to override the crl extensions
9756 *) Add new feature to the SPKAC handling in ca. Now you can include
9757 the same field multiple times by preceding it by "XXXX." for example:
9760 this is the same syntax as used in the req config file.
9763 *) Allow certificate extensions to be added to certificate requests. These
9764 are specified in a 'req_extensions' option of the req section of the
9765 config file. They can be printed out with the -text option to req but
9766 are otherwise ignored at present.
9769 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9770 data read consists of only the final block it would not decrypted because
9771 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9772 A misplaced 'break' also meant the decrypted final block might not be
9773 copied until the next read.
9776 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9777 a few extra parameters to the DH structure: these will be useful if
9778 for example we want the value of 'q' or implement X9.42 DH.
9781 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9782 provides hooks that allow the default DSA functions or functions on a
9783 "per key" basis to be replaced. This allows hardware acceleration and
9784 hardware key storage to be handled without major modification to the
9785 library. Also added low level modexp hooks and CRYPTO_EX structure and
9786 associated functions.
9789 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9790 as "read only": it can't be written to and the buffer it points to will
9791 not be freed. Reading from a read only BIO is much more efficient than
9792 a normal memory BIO. This was added because there are several times when
9793 an area of memory needs to be read from a BIO. The previous method was
9794 to create a memory BIO and write the data to it, this results in two
9795 copies of the data and an O(n^2) reading algorithm. There is a new
9796 function BIO_new_mem_buf() which creates a read only memory BIO from
9797 an area of memory. Also modified the PKCS#7 routines to use read only
9801 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9802 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9803 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9804 but a retry condition occured while trying to read the rest.
9807 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9808 NID_pkcs7_encrypted by default: this was wrong since this should almost
9809 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9810 the encrypted data type: this is a more sensible place to put it and it
9811 allows the PKCS#12 code to be tidied up that duplicated this
9815 *) Changed obj_dat.pl script so it takes its input and output files on
9816 the command line. This should avoid shell escape redirection problems
9820 *) Initial support for certificate extension requests, these are included
9821 in things like Xenroll certificate requests. Included functions to allow
9822 extensions to be obtained and added.
9825 *) -crlf option to s_client and s_server for sending newlines as
9826 CRLF (as required by many protocols).
9829 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9831 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9832 [Ralf S. Engelschall]
9834 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9835 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9837 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9841 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9842 DH parameters/keys (q is lost during that conversion, but the resulting
9843 DH parameters contain its length).
9845 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9846 much faster than DH_generate_parameters (which creates parameters
9847 where p = 2*q + 1), and also the smaller q makes DH computations
9848 much more efficient (160-bit exponentiation instead of 1024-bit
9849 exponentiation); so this provides a convenient way to support DHE
9850 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9851 utter importance to use
9852 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9854 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9855 when such DH parameters are used, because otherwise small subgroup
9856 attacks may become possible!
9859 *) Avoid memory leak in i2d_DHparams.
9862 *) Allow the -k option to be used more than once in the enc program:
9863 this allows the same encrypted message to be read by multiple recipients.
9866 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9867 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9868 it will always use the numerical form of the OID, even if it has a short
9872 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9873 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9874 otherwise bn_mod_exp was called. In the case of hardware keys for example
9875 no private key components need be present and it might store extra data
9876 in the RSA structure, which cannot be accessed from bn_mod_exp.
9877 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9878 private key operations.
9881 *) Added support for SPARC Linux.
9884 *) pem_password_cb function type incompatibly changed from
9885 typedef int pem_password_cb(char *buf, int size, int rwflag);
9887 ....(char *buf, int size, int rwflag, void *userdata);
9888 so that applications can pass data to their callbacks:
9889 The PEM[_ASN1]_{read,write}... functions and macros now take an
9890 additional void * argument, which is just handed through whenever
9891 the password callback is called.
9892 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9894 New function SSL_CTX_set_default_passwd_cb_userdata.
9896 Compatibility note: As many C implementations push function arguments
9897 onto the stack in reverse order, the new library version is likely to
9898 interoperate with programs that have been compiled with the old
9899 pem_password_cb definition (PEM_whatever takes some data that
9900 happens to be on the stack as its last argument, and the callback
9901 just ignores this garbage); but there is no guarantee whatsoever that
9904 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9905 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9906 problems not only on Windows, but also on some Unix platforms.
9907 To avoid problematic command lines, these definitions are now in an
9908 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9909 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9912 *) MIPS III/IV assembler module is reimplemented.
9915 *) More DES library cleanups: remove references to srand/rand and
9916 delete an unused file.
9919 *) Add support for the the free Netwide assembler (NASM) under Win32,
9920 since not many people have MASM (ml) and it can be hard to obtain.
9921 This is currently experimental but it seems to work OK and pass all
9922 the tests. Check out INSTALL.W32 for info.
9925 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9926 without temporary keys kept an extra copy of the server key,
9927 and connections with temporary keys did not free everything in case
9931 *) New function RSA_check_key and new openssl rsa option -check
9932 for verifying the consistency of RSA keys.
9933 [Ulf Moeller, Bodo Moeller]
9935 *) Various changes to make Win32 compile work:
9936 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9937 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9938 comparison" warnings.
9939 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9942 *) Add a debugging option to PKCS#5 v2 key generation function: when
9943 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9944 derived keys are printed to stderr.
9947 *) Copy the flags in ASN1_STRING_dup().
9948 [Roman E. Pavlov <pre@mo.msk.ru>]
9950 *) The x509 application mishandled signing requests containing DSA
9951 keys when the signing key was also DSA and the parameters didn't match.
9953 It was supposed to omit the parameters when they matched the signing key:
9954 the verifying software was then supposed to automatically use the CA's
9955 parameters if they were absent from the end user certificate.
9957 Omitting parameters is no longer recommended. The test was also
9958 the wrong way round! This was probably due to unusual behaviour in
9959 EVP_cmp_parameters() which returns 1 if the parameters match.
9960 This meant that parameters were omitted when they *didn't* match and
9961 the certificate was useless. Certificates signed with 'ca' didn't have
9963 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9965 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9966 The interface is as follows:
9967 Applications can use
9968 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9969 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9970 "off" is now the default.
9971 The library internally uses
9972 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9973 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9974 to disable memory-checking temporarily.
9976 Some inconsistent states that previously were possible (and were
9977 even the default) are now avoided.
9979 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9980 with each memory chunk allocated; this is occasionally more helpful
9981 than just having a counter.
9983 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9985 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9989 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9990 which largely parallels "options", but is for changing API behaviour,
9991 whereas "options" are about protocol behaviour.
9992 Initial "mode" flags are:
9994 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9995 a single record has been written.
9996 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9997 retries use the same buffer location.
9998 (But all of the contents must be
10002 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10005 *) Fix problems with no-hmac etc.
10006 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10008 *) New functions RSA_get_default_method(), RSA_set_method() and
10009 RSA_get_method(). These allows replacement of RSA_METHODs without having
10010 to mess around with the internals of an RSA structure.
10013 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10014 Also really enable memory leak checks in openssl.c and in some
10016 [Chad C. Mulligan, Bodo Moeller]
10018 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10019 up the length of negative integers. This has now been simplified to just
10020 store the length when it is first determined and use it later, rather
10021 than trying to keep track of where data is copied and updating it to
10023 [Steve Henson, reported by Brien Wheeler
10024 <bwheeler@authentica-security.com>]
10026 *) Add a new function PKCS7_signatureVerify. This allows the verification
10027 of a PKCS#7 signature but with the signing certificate passed to the
10028 function itself. This contrasts with PKCS7_dataVerify which assumes the
10029 certificate is present in the PKCS#7 structure. This isn't always the
10030 case: certificates can be omitted from a PKCS#7 structure and be
10031 distributed by "out of band" means (such as a certificate database).
10034 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10035 function prototypes in pem.h, also change util/mkdef.pl to add the
10036 necessary function names.
10039 *) mk1mf.pl (used by Windows builds) did not properly read the
10040 options set by Configure in the top level Makefile, and Configure
10041 was not even able to write more than one option correctly.
10042 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10045 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10046 file to be loaded from a BIO or FILE pointer. The BIO version will
10047 for example allow memory BIOs to contain config info.
10050 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10051 Whoever hopes to achieve shared-library compatibility across versions
10052 must use this, not the compile-time macro.
10053 (Exercise 0.9.4: Which is the minimum library version required by
10055 Note: All this applies only to multi-threaded programs, others don't
10059 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10060 through a BIO pair triggered the default case, i.e.
10061 SSLerr(...,SSL_R_UNKNOWN_STATE).
10064 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10065 can use the SSL library even if none of the specific BIOs is
10069 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10070 for the encoded length.
10071 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10073 *) Add initial documentation of the X509V3 functions.
10076 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10077 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10078 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10079 secure PKCS#8 private key format with a high iteration count.
10082 *) Fix determination of Perl interpreter: A perl or perl5
10083 _directory_ in $PATH was also accepted as the interpreter.
10084 [Ralf S. Engelschall]
10086 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10087 wrong with it but it was very old and did things like calling
10088 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10089 unusual formatting.
10092 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10093 to use the new extension code.
10096 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10097 with macros. This should make it easier to change their form, add extra
10098 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10102 *) Add to configuration table a new entry that can specify an alternative
10103 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10104 according to Mark Crispin <MRC@Panda.COM>.
10108 *) DES CBC did not update the IV. Weird.
10111 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10112 Changing the behaviour of the former might break existing programs --
10113 where IV updating is needed, des_ncbc_encrypt can be used.
10116 *) When bntest is run from "make test" it drives bc to check its
10117 calculations, as well as internally checking them. If an internal check
10118 fails, it needs to cause bc to give a non-zero result or make test carries
10119 on without noticing the failure. Fixed.
10122 *) DES library cleanups.
10125 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10126 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10127 ciphers. NOTE: although the key derivation function has been verified
10128 against some published test vectors it has not been extensively tested
10129 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10133 *) Instead of "mkdir -p", which is not fully portable, use new
10134 Perl script "util/mkdir-p.pl".
10137 *) Rewrite the way password based encryption (PBE) is handled. It used to
10138 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10139 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10140 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10141 the 'parameter' field of the AlgorithmIdentifier is passed to the
10142 underlying key generation function so it must do its own ASN1 parsing.
10143 This has also changed the EVP_PBE_CipherInit() function which now has a
10144 'parameter' argument instead of literal salt and iteration count values
10145 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10148 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10149 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10150 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10151 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10152 value was just used as a "magic string" and not used directly its
10153 value doesn't matter.
10156 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10160 *) "linux-sparc64" configuration (ultrapenguin).
10161 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10162 "linux-sparc" configuration.
10163 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10165 *) config now generates no-xxx options for missing ciphers.
10168 *) Support the EBCDIC character set (work in progress).
10169 File ebcdic.c not yet included because it has a different license.
10170 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10172 *) Support BS2000/OSD-POSIX.
10173 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10175 *) Make callbacks for key generation use void * instead of char *.
10178 *) Make S/MIME samples compile (not yet tested).
10181 *) Additional typesafe stacks.
10184 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10188 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10190 *) New configuration variant "sco5-gcc".
10192 *) Updated some demos.
10193 [Sean O Riordain, Wade Scholine]
10195 *) Add missing BIO_free at exit of pkcs12 application.
10198 *) Fix memory leak in conf.c.
10201 *) Updates for Win32 to assembler version of MD5.
10204 *) Set #! path to perl in apps/der_chop to where we found it
10205 instead of using a fixed path.
10208 *) SHA library changes for irix64-mips4-cc.
10211 *) Improvements for VMS support.
10215 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10217 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10218 This also avoids the problems with SC4.2 and unpatched SC5.
10219 [Andy Polyakov <appro@fy.chalmers.se>]
10221 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10222 These are required because of the typesafe stack would otherwise break
10223 existing code. If old code used a structure member which used to be STACK
10224 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10225 sk_num or sk_value it would produce an error because the num, data members
10226 are not present in STACK_OF. Now it just produces a warning. sk_set
10227 replaces the old method of assigning a value to sk_value
10228 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10229 that does this will no longer work (and should use sk_set instead) but
10230 this could be regarded as a "questionable" behaviour anyway.
10233 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10234 correctly handle encrypted S/MIME data.
10237 *) Change type of various DES function arguments from des_cblock
10238 (which means, in function argument declarations, pointer to char)
10239 to des_cblock * (meaning pointer to array with 8 char elements),
10240 which allows the compiler to do more typechecking; it was like
10241 that back in SSLeay, but with lots of ugly casts.
10243 Introduce new type const_des_cblock.
10246 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10247 problems: find RecipientInfo structure that matches recipient certificate
10248 and initialise the ASN1 structures properly based on passed cipher.
10251 *) Belatedly make the BN tests actually check the results.
10254 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10255 to and from BNs: it was completely broken. New compilation option
10256 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10257 key elements as negative integers.
10260 *) Reorganize and speed up MD5.
10261 [Andy Polyakov <appro@fy.chalmers.se>]
10264 [Richard Levitte <richard@levitte.org>]
10266 *) New option -out to asn1parse to allow the parsed structure to be
10267 output to a file. This is most useful when combined with the -strparse
10268 option to examine the output of things like OCTET STRINGS.
10271 *) Make SSL library a little more fool-proof by not requiring any longer
10272 that SSL_set_{accept,connect}_state be called before
10273 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10274 in many applications because usually everything *appeared* to work as
10275 intended anyway -- now it really works as intended).
10278 *) Move openssl.cnf out of lib/.
10281 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10282 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10283 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10284 [Ralf S. Engelschall]
10286 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10287 handle PKCS#7 enveloped data properly.
10288 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10290 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10291 copying pointers. The cert_st handling is changed by this in
10292 various ways (and thus what used to be known as ctx->default_cert
10293 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10294 any longer when s->cert does not give us what we need).
10295 ssl_cert_instantiate becomes obsolete by this change.
10296 As soon as we've got the new code right (possibly it already is?),
10297 we have solved a couple of bugs of the earlier code where s->cert
10298 was used as if it could not have been shared with other SSL structures.
10300 Note that using the SSL API in certain dirty ways now will result
10301 in different behaviour than observed with earlier library versions:
10302 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10303 does not influence s as it used to.
10305 In order to clean up things more thoroughly, inside SSL_SESSION
10306 we don't use CERT any longer, but a new structure SESS_CERT
10307 that holds per-session data (if available); currently, this is
10308 the peer's certificate chain and, for clients, the server's certificate
10309 and temporary key. CERT holds only those values that can have
10310 meaningful defaults in an SSL_CTX.
10313 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10314 from the internal representation. Various PKCS#7 fixes: remove some
10315 evil casts and set the enc_dig_alg field properly based on the signing
10319 *) Allow PKCS#12 password to be set from the command line or the
10320 environment. Let 'ca' get its config file name from the environment
10321 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10325 *) Allow certificate policies extension to use an IA5STRING for the
10326 organization field. This is contrary to the PKIX definition but
10327 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10331 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10332 without disallowing inline assembler and the like for non-pedantic builds.
10335 *) Support Borland C++ builder.
10336 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10338 *) Support Mingw32.
10341 *) SHA-1 cleanups and performance enhancements.
10342 [Andy Polyakov <appro@fy.chalmers.se>]
10344 *) Sparc v8plus assembler for the bignum library.
10345 [Andy Polyakov <appro@fy.chalmers.se>]
10347 *) Accept any -xxx and +xxx compiler options in Configure.
10350 *) Update HPUX configuration.
10353 *) Add missing sk_<type>_unshift() function to safestack.h
10354 [Ralf S. Engelschall]
10356 *) New function SSL_CTX_use_certificate_chain_file that sets the
10357 "extra_cert"s in addition to the certificate. (This makes sense
10358 only for "PEM" format files, as chains as a whole are not
10362 *) Support verify_depth from the SSL API.
10363 x509_vfy.c had what can be considered an off-by-one-error:
10364 Its depth (which was not part of the external interface)
10365 was actually counting the number of certificates in a chain;
10366 now it really counts the depth.
10369 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10370 instead of X509err, which often resulted in confusing error
10371 messages since the error codes are not globally unique
10372 (e.g. an alleged error in ssl3_accept when a certificate
10373 didn't match the private key).
10375 *) New function SSL_CTX_set_session_id_context that allows to set a default
10376 value (so that you don't need SSL_set_session_id_context for each
10377 connection using the SSL_CTX).
10380 *) OAEP decoding bug fix.
10383 *) Support INSTALL_PREFIX for package builders, as proposed by
10387 *) New Configure options "threads" and "no-threads". For systems
10388 where the proper compiler options are known (currently Solaris
10389 and Linux), "threads" is the default.
10392 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10395 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10396 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10397 such as /usr/local/bin.
10400 *) "make linux-shared" to build shared libraries.
10401 [Niels Poppe <niels@netbox.org>]
10403 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10406 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10407 extension adding in x509 utility.
10410 *) Remove NOPROTO sections and error code comments.
10413 *) Partial rewrite of the DEF file generator to now parse the ANSI
10417 *) New Configure options --prefix=DIR and --openssldir=DIR.
10420 *) Complete rewrite of the error code script(s). It is all now handled
10421 by one script at the top level which handles error code gathering,
10422 header rewriting and C source file generation. It should be much better
10423 than the old method: it now uses a modified version of Ulf's parser to
10424 read the ANSI prototypes in all header files (thus the old K&R definitions
10425 aren't needed for error creation any more) and do a better job of
10426 translating function codes into names. The old 'ASN1 error code imbedded
10427 in a comment' is no longer necessary and it doesn't use .err files which
10428 have now been deleted. Also the error code call doesn't have to appear all
10429 on one line (which resulted in some large lines...).
10432 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10435 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10436 0 (which usually indicates a closed connection), but continue reading.
10439 *) Fix some race conditions.
10442 *) Add support for CRL distribution points extension. Add Certificate
10443 Policies and CRL distribution points documentation.
10446 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10449 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10450 8 of keying material. Merlin has also confirmed interop with this fix
10451 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10452 [Merlin Hughes <merlin@baltimore.ie>]
10454 *) Fix lots of warnings.
10455 [Richard Levitte <levitte@stacken.kth.se>]
10457 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10458 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10459 [Richard Levitte <levitte@stacken.kth.se>]
10461 *) Fix problems with sizeof(long) == 8.
10462 [Andy Polyakov <appro@fy.chalmers.se>]
10464 *) Change functions to ANSI C.
10467 *) Fix typos in error codes.
10468 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10470 *) Remove defunct assembler files from Configure.
10473 *) SPARC v8 assembler BIGNUM implementation.
10474 [Andy Polyakov <appro@fy.chalmers.se>]
10476 *) Support for Certificate Policies extension: both print and set.
10477 Various additions to support the r2i method this uses.
10480 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10481 return a const string when you are expecting an allocated buffer.
10484 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10485 types DirectoryString and DisplayText.
10488 *) Add code to allow r2i extensions to access the configuration database,
10489 add an LHASH database driver and add several ctx helper functions.
10492 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10493 fail when they extended the size of a BIGNUM.
10496 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10497 support typesafe stack.
10500 *) Fix typo in SSL_[gs]et_options().
10501 [Nils Frostberg <nils@medcom.se>]
10503 *) Delete various functions and files that belonged to the (now obsolete)
10504 old X509V3 handling code.
10507 *) New Configure option "rsaref".
10510 *) Don't auto-generate pem.h.
10513 *) Introduce type-safe ASN.1 SETs.
10516 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10517 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10519 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10520 that links with OpenSSL (well at least cause lots of warnings), but fear
10521 not: the conversion is trivial, and it eliminates loads of evil casts. A
10522 few STACKed things have been converted already. Feel free to convert more.
10523 In the fullness of time, I'll do away with the STACK type altogether.
10526 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10527 specified in <certfile> by updating the entry in the index.txt file.
10528 This way one no longer has to edit the index.txt file manually for
10529 revoking a certificate. The -revoke option does the gory details now.
10530 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10532 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10533 `-text' option at all and this way the `-noout -text' combination was
10534 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10535 [Ralf S. Engelschall]
10537 *) Make sure a corresponding plain text error message exists for the
10538 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10539 verify callback function determined that a certificate was revoked.
10540 [Ralf S. Engelschall]
10542 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10543 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10544 all available cipers including rc5, which was forgotten until now.
10545 In order to let the testing shell script know which algorithms
10546 are available, a new (up to now undocumented) command
10547 "openssl list-cipher-commands" is used.
10550 *) Bugfix: s_client occasionally would sleep in select() when
10551 it should have checked SSL_pending() first.
10554 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10555 the raw DSA values prior to ASN.1 encoding.
10558 *) Tweaks to Configure
10559 [Niels Poppe <niels@netbox.org>]
10561 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10565 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10568 *) New config option to avoid instructions that are illegal on the 80386.
10569 The default code is faster, but requires at least a 486.
10572 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10573 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10574 same as SSL2_VERSION anyway.
10577 *) New "-showcerts" option for s_client.
10580 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10581 application. Various cleanups and fixes.
10584 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10585 modify error routines to work internally. Add error codes and PBE init
10586 to library startup routines.
10589 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10590 packing functions to asn1 and evp. Changed function names and error
10591 codes along the way.
10594 *) PKCS12 integration: and so it begins... First of several patches to
10595 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10596 objects to objects.h
10599 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10600 and display support for Thawte strong extranet extension.
10603 *) Add LinuxPPC support.
10604 [Jeff Dubrule <igor@pobox.org>]
10606 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10607 bn_div_words in alpha.s.
10608 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10610 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10611 OAEP isn't supported when OpenSSL is built with RSAref.
10612 [Ulf Moeller <ulf@fitug.de>]
10614 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10615 so they no longer are missing under -DNOPROTO.
10616 [Soren S. Jorvang <soren@t.dk>]
10619 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10621 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10622 doesn't work when the session is reused. Coming soon!
10625 *) Fix a security hole, that allows sessions to be reused in the wrong
10626 context thus bypassing client cert protection! All software that uses
10627 client certs and session caches in multiple contexts NEEDS PATCHING to
10628 allow session reuse! A fuller solution is in the works.
10629 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10631 *) Some more source tree cleanups (removed obsolete files
10632 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10633 permission on "config" script to be executable) and a fix for the INSTALL
10635 [Ulf Moeller <ulf@fitug.de>]
10637 *) Remove some legacy and erroneous uses of malloc, free instead of
10639 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10641 *) Make rsa_oaep_test return non-zero on error.
10642 [Ulf Moeller <ulf@fitug.de>]
10644 *) Add support for native Solaris shared libraries. Configure
10645 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10646 if someone would make that last step automatic.
10647 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10649 *) ctx_size was not built with the right compiler during "make links". Fixed.
10652 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10653 except NULL ciphers". This means the default cipher list will no longer
10654 enable NULL ciphers. They need to be specifically enabled e.g. with
10655 the string "DEFAULT:eNULL".
10658 *) Fix to RSA private encryption routines: if p < q then it would
10659 occasionally produce an invalid result. This will only happen with
10660 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10663 *) Be less restrictive and allow also `perl util/perlpath.pl
10664 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10665 because this way one can also use an interpreter named `perl5' (which is
10666 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10667 installed as `perl').
10668 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10670 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10671 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10673 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10674 advapi32.lib to Win32 build and change the pem test comparision
10675 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10676 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10677 and crypto/des/ede_cbcm_enc.c.
10680 *) DES quad checksum was broken on big-endian architectures. Fixed.
10683 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10684 Win32 test batch file so it (might) work again. The Win32 test batch file
10685 is horrible: I feel ill....
10688 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10689 in e_os.h. Audit of header files to check ANSI and non ANSI
10690 sections: 10 functions were absent from non ANSI section and not exported
10691 from Windows DLLs. Fixed up libeay.num for new functions.
10694 *) Make `openssl version' output lines consistent.
10695 [Ralf S. Engelschall]
10697 *) Fix Win32 symbol export lists for BIO functions: Added
10698 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10699 to ms/libeay{16,32}.def.
10700 [Ralf S. Engelschall]
10702 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10703 fine under Unix and passes some trivial tests I've now added. But the
10704 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10705 added to make sure no one expects that this stuff really works in the
10706 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10707 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10709 [Ralf S. Engelschall]
10711 *) Fix the generation of two part addresses in perl.
10712 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10714 *) Add config entry for Linux on MIPS.
10715 [John Tobey <jtobey@channel1.com>]
10717 *) Make links whenever Configure is run, unless we are on Windoze.
10720 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10721 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10725 *) Add a useful kludge to allow package maintainers to specify compiler and
10726 other platforms details on the command line without having to patch the
10727 Configure script everytime: One now can use ``perl Configure
10728 <id>:<details>'', i.e. platform ids are allowed to have details appended
10729 to them (seperated by colons). This is treated as there would be a static
10730 pre-configured entry in Configure's %table under key <id> with value
10731 <details> and ``perl Configure <id>'' is called. So, when you want to
10732 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10733 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10734 now, which overrides the FreeBSD-elf entry on-the-fly.
10735 [Ralf S. Engelschall]
10737 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10740 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10741 on the `perl Configure ...' command line. This way one can compile
10742 OpenSSL libraries with Position Independent Code (PIC) which is needed
10743 for linking it into DSOs.
10744 [Ralf S. Engelschall]
10746 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10750 *) Cleaned up the LICENSE document: The official contact for any license
10751 questions now is the OpenSSL core team under openssl-core@openssl.org.
10752 And add a paragraph about the dual-license situation to make sure people
10753 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10754 to the OpenSSL toolkit.
10755 [Ralf S. Engelschall]
10757 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10758 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10759 Additonally cleaned up the `make links' target: Remove unnecessary
10760 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10761 to speed processing and no longer clutter the display with confusing
10762 stuff. Instead only the actually done links are displayed.
10763 [Ralf S. Engelschall]
10765 *) Permit null encryption ciphersuites, used for authentication only. It used
10766 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10767 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10771 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10772 signed attributes when verifying signatures (this would break them),
10773 the detached data encoding was wrong and public keys obtained using
10774 X509_get_pubkey() weren't freed.
10777 *) Add text documentation for the BUFFER functions. Also added a work around
10778 to a Win95 console bug. This was triggered by the password read stuff: the
10779 last character typed gets carried over to the next fread(). If you were
10780 generating a new cert request using 'req' for example then the last
10781 character of the passphrase would be CR which would then enter the first
10785 *) Added the new `Includes OpenSSL Cryptography Software' button as
10786 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10787 button and can be used by applications based on OpenSSL to show the
10788 relationship to the OpenSSL project.
10789 [Ralf S. Engelschall]
10791 *) Remove confusing variables in function signatures in files
10792 ssl/ssl_lib.c and ssl/ssl.h.
10793 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10795 *) Don't install bss_file.c under PREFIX/include/
10796 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10798 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10799 functions that return function pointers and has support for NT specific
10800 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10801 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10802 unsigned to signed types: this was killing the Win32 compile.
10805 *) Add new certificate file to stack functions,
10806 SSL_add_dir_cert_subjects_to_stack() and
10807 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10808 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10809 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10810 This means that Apache-SSL and similar packages don't have to mess around
10811 to add as many CAs as they want to the preferred list.
10814 *) Experiment with doxygen documentation. Currently only partially applied to
10816 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10817 openssl.doxy as the configuration file.
10820 *) Get rid of remaining C++-style comments which strict C compilers hate.
10821 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10823 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10824 compiled in by default: it has problems with large keys.
10827 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10828 DH private keys and/or callback functions which directly correspond to
10829 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10830 is needed for applications which have to configure certificates on a
10831 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10833 For the RSA certificate situation is makes no difference, but
10834 for the DSA certificate situation this fixes the "no shared cipher"
10835 problem where the OpenSSL cipher selection procedure failed because the
10836 temporary keys were not overtaken from the context and the API provided
10837 no way to reconfigure them.
10838 The new functions now let applications reconfigure the stuff and they
10839 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10840 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10841 non-public-API function ssl_cert_instantiate() is used as a helper
10842 function and also to reduce code redundancy inside ssl_rsa.c.
10843 [Ralf S. Engelschall]
10845 *) Move s_server -dcert and -dkey options out of the undocumented feature
10846 area because they are useful for the DSA situation and should be
10847 recognized by the users.
10848 [Ralf S. Engelschall]
10850 *) Fix the cipher decision scheme for export ciphers: the export bits are
10851 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10852 SSL_EXP_MASK. So, the original variable has to be used instead of the
10853 already masked variable.
10854 [Richard Levitte <levitte@stacken.kth.se>]
10856 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10857 [Richard Levitte <levitte@stacken.kth.se>]
10859 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10860 from `int' to `unsigned int' because it's a length and initialized by
10861 EVP_DigestFinal() which expects an `unsigned int *'.
10862 [Richard Levitte <levitte@stacken.kth.se>]
10864 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10865 script. Instead use the usual Shell->Perl transition trick.
10866 [Ralf S. Engelschall]
10868 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10869 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10870 -noout -modulus' as it's already the case for `openssl rsa -noout
10871 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10872 currently the public key is printed (a decision which was already done by
10873 `openssl dsa -modulus' in the past) which serves a similar purpose.
10874 Additionally the NO_RSA no longer completely removes the whole -modulus
10875 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10877 [Ralf S. Engelschall]
10879 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10880 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10881 [Arne Ansper <arne@ats.cyber.ee>]
10883 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10884 to be added. Now both 'req' and 'ca' can use new objects defined in the
10888 *) Add cool BIO that does syslog (or event log on NT).
10889 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10891 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10892 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10893 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10894 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10897 *) Add preliminary config info for new extension code.
10900 *) Make RSA_NO_PADDING really use no padding.
10901 [Ulf Moeller <ulf@fitug.de>]
10903 *) Generate errors when private/public key check is done.
10906 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10907 for some CRL extensions and new objects added.
10910 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10911 key usage extension and fuller support for authority key id.
10914 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10915 padding method for RSA, which is recommended for new applications in PKCS
10916 #1 v2.0 (RFC 2437, October 1998).
10917 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10918 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10919 against Bleichbacher's attack on RSA.
10920 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10923 *) Updates to the new SSL compression code
10924 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10926 *) Fix so that the version number in the master secret, when passed
10927 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10928 (because the server will not accept higher), that the version number
10929 is 0x03,0x01, not 0x03,0x00
10930 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10932 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10933 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10934 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10937 *) Support for RAW extensions where an arbitrary extension can be
10938 created by including its DER encoding. See apps/openssl.cnf for
10942 *) Make sure latest Perl versions don't interpret some generated C array
10943 code as Perl array code in the crypto/err/err_genc.pl script.
10944 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10946 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10947 not many people have the assembler. Various Win32 compilation fixes and
10948 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10949 build instructions.
10952 *) Modify configure script 'Configure' to automatically create crypto/date.h
10953 file under Win32 and also build pem.h from pem.org. New script
10954 util/mkfiles.pl to create the MINFO file on environments that can't do a
10955 'make files': perl util/mkfiles.pl >MINFO should work.
10958 *) Major rework of DES function declarations, in the pursuit of correctness
10959 and purity. As a result, many evil casts evaporated, and some weirdness,
10960 too. You may find this causes warnings in your code. Zapping your evil
10961 casts will probably fix them. Mostly.
10964 *) Fix for a typo in asn1.h. Bug fix to object creation script
10965 obj_dat.pl. It considered a zero in an object definition to mean
10966 "end of object": none of the objects in objects.h have any zeros
10967 so it wasn't spotted.
10968 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10970 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10971 Masking (CBCM). In the absence of test vectors, the best I have been able
10972 to do is check that the decrypt undoes the encrypt, so far. Send me test
10973 vectors if you have them.
10976 *) Correct calculation of key length for export ciphers (too much space was
10977 allocated for null ciphers). This has not been tested!
10980 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10981 message is now correct (it understands "crypto" and "ssl" on its
10982 command line). There is also now an "update" option. This will update
10983 the util/ssleay.num and util/libeay.num files with any new functions.
10985 perl util/mkdef.pl crypto ssl update
10986 it will update them.
10989 *) Overhauled the Perl interface (perl/*):
10990 - ported BN stuff to OpenSSL's different BN library
10991 - made the perl/ source tree CVS-aware
10992 - renamed the package from SSLeay to OpenSSL (the files still contain
10993 their history because I've copied them in the repository)
10994 - removed obsolete files (the test scripts will be replaced
10995 by better Test::Harness variants in the future)
10996 [Ralf S. Engelschall]
10998 *) First cut for a very conservative source tree cleanup:
10999 1. merge various obsolete readme texts into doc/ssleay.txt
11000 where we collect the old documents and readme texts.
11001 2. remove the first part of files where I'm already sure that we no
11002 longer need them because of three reasons: either they are just temporary
11003 files which were left by Eric or they are preserved original files where
11004 I've verified that the diff is also available in the CVS via "cvs diff
11005 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11006 the crypto/md/ stuff).
11007 [Ralf S. Engelschall]
11009 *) More extension code. Incomplete support for subject and issuer alt
11010 name, issuer and authority key id. Change the i2v function parameters
11011 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11012 what that's for :-) Fix to ASN1 macro which messed up
11013 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11016 *) Preliminary support for ENUMERATED type. This is largely copied from the
11020 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11021 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11023 *) Make sure `make rehash' target really finds the `openssl' program.
11024 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11026 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11027 like to hear about it if this slows down other processors.
11030 *) Add CygWin32 platform information to Configure script.
11031 [Alan Batie <batie@aahz.jf.intel.com>]
11033 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11034 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11036 *) New program nseq to manipulate netscape certificate sequences
11039 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11043 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11044 but the BN code had some problems that would cause failures when
11045 doing certificate verification and some other functions.
11046 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11048 *) Add ASN1 and PEM code to support netscape certificate sequences.
11051 *) Add ASN1 and PEM code to support netscape certificate sequences.
11054 *) Add several PKIX and private extended key usage OIDs.
11057 *) Modify the 'ca' program to handle the new extension code. Modify
11058 openssl.cnf for new extension format, add comments.
11061 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11062 and add a sample to openssl.cnf so req -x509 now adds appropriate
11066 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11067 error code, add initial support to X509_print() and x509 application.
11070 *) Takes a deep breath and start addding X509 V3 extension support code. Add
11071 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11072 stuff is currently isolated and isn't even compiled yet.
11075 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11076 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11077 Removed the versions check from X509 routines when loading extensions:
11078 this allows certain broken certificates that don't set the version
11079 properly to be processed.
11082 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11083 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11084 can still be regenerated with "make depend".
11087 *) Spelling mistake in C version of CAST-128.
11088 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11090 *) Changes to the error generation code. The perl script err-code.pl
11091 now reads in the old error codes and retains the old numbers, only
11092 adding new ones if necessary. It also only changes the .err files if new
11093 codes are added. The makefiles have been modified to only insert errors
11094 when needed (to avoid needlessly modifying header files). This is done
11095 by only inserting errors if the .err file is newer than the auto generated
11096 C file. To rebuild all the error codes from scratch (the old behaviour)
11097 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11098 or delete all the .err files.
11101 *) CAST-128 was incorrectly implemented for short keys. The C version has
11102 been fixed, but is untested. The assembler versions are also fixed, but
11103 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11104 to regenerate it if needed.
11105 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11106 Hagino <itojun@kame.net>]
11108 *) File was opened incorrectly in randfile.c.
11109 [Ulf Möller <ulf@fitug.de>]
11111 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11112 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11113 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11114 al: it's just almost always a UTCTime. Note this patch adds new error
11115 codes so do a "make errors" if there are problems.
11118 *) Correct Linux 1 recognition in config.
11119 [Ulf Möller <ulf@fitug.de>]
11121 *) Remove pointless MD5 hash when using DSA keys in ca.
11122 [Anonymous <nobody@replay.com>]
11124 *) Generate an error if given an empty string as a cert directory. Also
11125 generate an error if handed NULL (previously returned 0 to indicate an
11126 error, but didn't set one).
11127 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11129 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11132 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11133 parameters. This was causing a warning which killed off the Win32 compile.
11136 *) Remove C++ style comments from crypto/bn/bn_local.h.
11137 [Neil Costigan <neil.costigan@celocom.com>]
11139 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11140 based on a text string, looking up short and long names and finally
11141 "dot" format. The "dot" format stuff didn't work. Added new function
11142 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11143 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11144 OID is not part of the table.
11147 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11148 X509_LOOKUP_by_alias().
11151 *) Sort openssl functions by name.
11154 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11155 encryption from sample DSA keys (in case anyone is interested the password
11159 *) Make _all_ *_free functions accept a NULL pointer.
11160 [Frans Heymans <fheymans@isaserver.be>]
11162 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11164 [Anonymous <nobody@replay.com>]
11166 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11167 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11169 *) Don't blow it for numeric -newkey arguments to apps/req.
11170 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11172 *) Temp key "for export" tests were wrong in s3_srvr.c.
11173 [Anonymous <nobody@replay.com>]
11175 *) Add prototype for temp key callback functions
11176 SSL_CTX_set_tmp_{rsa,dh}_callback().
11179 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11180 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11183 *) X509_name_add_entry() freed the wrong thing after an error.
11184 [Arne Ansper <arne@ats.cyber.ee>]
11186 *) rsa_eay.c would attempt to free a NULL context.
11187 [Arne Ansper <arne@ats.cyber.ee>]
11189 *) BIO_s_socket() had a broken should_retry() on Windoze.
11190 [Arne Ansper <arne@ats.cyber.ee>]
11192 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11193 [Arne Ansper <arne@ats.cyber.ee>]
11195 *) Make sure the already existing X509_STORE->depth variable is initialized
11196 in X509_STORE_new(), but document the fact that this variable is still
11197 unused in the certificate verification process.
11198 [Ralf S. Engelschall]
11200 *) Fix the various library and apps files to free up pkeys obtained from
11201 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11204 *) Fix reference counting in X509_PUBKEY_get(). This makes
11205 demos/maurice/example2.c work, amongst others, probably.
11206 [Steve Henson and Ben Laurie]
11208 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11209 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11210 are no longer created. This way we have a single and consistent command
11211 line interface `openssl <command>', similar to `cvs <command>'.
11212 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11214 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11215 BIT STRING wrapper always have zero unused bits.
11218 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11221 *) Make the top-level INSTALL documentation easier to understand.
11224 *) Makefiles updated to exit if an error occurs in a sub-directory
11225 make (including if user presses ^C) [Paul Sutton]
11227 *) Make Montgomery context stuff explicit in RSA data structure.
11230 *) Fix build order of pem and err to allow for generated pem.h.
11233 *) Fix renumbering bug in X509_NAME_delete_entry().
11236 *) Enhanced the err-ins.pl script so it makes the error library number
11237 global and can add a library name. This is needed for external ASN1 and
11238 other error libraries.
11241 *) Fixed sk_insert which never worked properly.
11244 *) Fix ASN1 macros so they can handle indefinite length construted
11245 EXPLICIT tags. Some non standard certificates use these: they can now
11249 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11250 into a single doc/ssleay.txt bundle. This way the information is still
11251 preserved but no longer messes up this directory. Now it's new room for
11252 the new set of documenation files.
11253 [Ralf S. Engelschall]
11255 *) SETs were incorrectly DER encoded. This was a major pain, because they
11256 shared code with SEQUENCEs, which aren't coded the same. This means that
11257 almost everything to do with SETs or SEQUENCEs has either changed name or
11258 number of arguments.
11259 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11261 *) Fix test data to work with the above.
11264 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11265 was already fixed by Eric for 0.9.1 it seems.
11266 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11268 *) Autodetect FreeBSD3.
11271 *) Fix various bugs in Configure. This affects the following platforms:
11275 unixware-2.0-pentium
11279 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11280 before they are needed.
11283 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11287 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11289 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11290 changed SSLeay to OpenSSL in version strings.
11291 [Ralf S. Engelschall]
11293 *) Some fixups to the top-level documents.
11296 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11297 because the symlink to include/ was missing.
11298 [Ralf S. Engelschall]
11300 *) Incorporated the popular no-RSA/DSA-only patches
11301 which allow to compile a RSA-free SSLeay.
11302 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11304 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11305 when "ssleay" is still not found.
11306 [Ralf S. Engelschall]
11308 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11309 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11311 *) Updated the README file.
11312 [Ralf S. Engelschall]
11314 *) Added various .cvsignore files in the CVS repository subdirs
11315 to make a "cvs update" really silent.
11316 [Ralf S. Engelschall]
11318 *) Recompiled the error-definition header files and added
11319 missing symbols to the Win32 linker tables.
11320 [Ralf S. Engelschall]
11322 *) Cleaned up the top-level documents;
11323 o new files: CHANGES and LICENSE
11324 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11325 o merged COPYRIGHT into LICENSE
11326 o removed obsolete TODO file
11327 o renamed MICROSOFT to INSTALL.W32
11328 [Ralf S. Engelschall]
11330 *) Removed dummy files from the 0.9.1b source tree:
11331 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11332 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11333 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11334 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11335 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11336 [Ralf S. Engelschall]
11338 *) Added various platform portability fixes.
11341 *) The Genesis of the OpenSSL rpject:
11342 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11343 Young and Tim J. Hudson created while they were working for C2Net until
11345 [The OpenSSL Project]
11348 Changes between 0.9.0b and 0.9.1b [not released]
11350 *) Updated a few CA certificates under certs/
11353 *) Changed some BIGNUM api stuff.
11356 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11357 DGUX x86, Linux Alpha, etc.
11360 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11361 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11365 *) Add -strparse option to asn1pars program which parses nested
11367 [Dr Stephen Henson <shenson@bigfoot.com>]
11369 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11372 *) DSA fix for "ca" program.
11375 *) Added "-genkey" option to "dsaparam" program.
11378 *) Added RIPE MD160 (rmd160) message digest.
11381 *) Added -a (all) option to "ssleay version" command.
11384 *) Added PLATFORM define which is the id given to Configure.
11387 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11390 *) Extended the ASN.1 parser routines.
11393 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11396 *) Added a BN_CTX to the BN library.
11399 *) Fixed the weak key values in DES library
11402 *) Changed API in EVP library for cipher aliases.
11405 *) Added support for RC2/64bit cipher.
11408 *) Converted the lhash library to the crypto/mem.c functions.
11411 *) Added more recognized ASN.1 object ids.
11414 *) Added more RSA padding checks for SSL/TLS.
11417 *) Added BIO proxy/filter functionality.
11420 *) Added extra_certs to SSL_CTX which can be used
11421 send extra CA certificates to the client in the CA cert chain sending
11422 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11425 *) Now Fortezza is denied in the authentication phase because
11426 this is key exchange mechanism is not supported by SSLeay at all.
11429 *) Additional PKCS1 checks.
11432 *) Support the string "TLSv1" for all TLS v1 ciphers.
11435 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11436 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11439 *) Fixed a few memory leaks.
11442 *) Fixed various code and comment typos.
11445 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11446 bytes sent in the client random.
11447 [Edward Bishop <ebishop@spyglass.com>]