5 This is a high-level summary of the most important changes.
6 For a full list of changes, see the git commit log; for example,
7 https://github.com/openssl/openssl/commits/ and pick the appropriate
10 Changes between 1.0.2o and 1.0.2p [xx XXX xxxx]
12 *) Revert blinding in ECDSA sign and instead make problematic addition
13 length-invariant. Switch even to fixed-length Montgomery multiplication.
16 *) Change generating and checking of primes so that the error rate of not
17 being prime depends on the intended use based on the size of the input.
18 For larger primes this will result in more rounds of Miller-Rabin.
19 The maximal error rate for primes with more than 1080 bits is lowered
21 [Kurt Roeckx, Annie Yousar]
23 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
26 *) Add blinding to ECDSA and DSA signatures to protect against side channel
27 attacks discovered by Keegan Ryan (NCC Group).
30 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
31 now allow empty (zero character) pass phrases.
34 *) Certificate time validation (X509_cmp_time) enforces stricter
35 compliance with RFC 5280. Fractional seconds and timezone offsets
36 are no longer allowed.
39 Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
41 *) Constructed ASN.1 types with a recursive definition could exceed the stack
43 Constructed ASN.1 types with a recursive definition (such as can be found
44 in PKCS7) could eventually exceed the stack given malicious input with
45 excessive recursion. This could result in a Denial Of Service attack. There
46 are no such structures used within SSL/TLS that come from untrusted sources
47 so this is considered safe.
49 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
54 Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
56 *) Read/write after SSL object in error state
58 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
59 mechanism. The intent was that if a fatal error occurred during a handshake
60 then OpenSSL would move into the error state and would immediately fail if
61 you attempted to continue the handshake. This works as designed for the
62 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
63 SSL_connect()), however due to a bug it does not work correctly if
64 SSL_read() or SSL_write() is called directly. In that scenario, if the
65 handshake fails then a fatal error will be returned in the initial function
66 call. If SSL_read()/SSL_write() is subsequently called by the application
67 for the same SSL object then it will succeed and the data is passed without
68 being decrypted/encrypted directly from the SSL/TLS record layer.
70 In order to exploit this issue an application bug would have to be present
71 that resulted in a call to SSL_read()/SSL_write() being issued after having
72 already received a fatal error.
74 This issue was reported to OpenSSL by David Benjamin (Google).
78 *) rsaz_1024_mul_avx2 overflow bug on x86_64
80 There is an overflow bug in the AVX2 Montgomery multiplication procedure
81 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
82 Analysis suggests that attacks against RSA and DSA as a result of this
83 defect would be very difficult to perform and are not believed likely.
84 Attacks against DH1024 are considered just feasible, because most of the
85 work necessary to deduce information about a private key may be performed
86 offline. The amount of resources required for such an attack would be
87 significant. However, for an attack on TLS to be meaningful, the server
88 would have to share the DH1024 private key among multiple clients, which is
89 no longer an option since CVE-2016-0701.
91 This only affects processors that support the AVX2 but not ADX extensions
92 like Intel Haswell (4th generation).
94 This issue was reported to OpenSSL by David Benjamin (Google). The issue
95 was originally found via the OSS-Fuzz project.
99 Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
101 *) bn_sqrx8x_internal carry bug on x86_64
103 There is a carry propagating bug in the x86_64 Montgomery squaring
104 procedure. No EC algorithms are affected. Analysis suggests that attacks
105 against RSA and DSA as a result of this defect would be very difficult to
106 perform and are not believed likely. Attacks against DH are considered just
107 feasible (although very difficult) because most of the work necessary to
108 deduce information about a private key may be performed offline. The amount
109 of resources required for such an attack would be very significant and
110 likely only accessible to a limited number of attackers. An attacker would
111 additionally need online access to an unpatched system using the target
112 private key in a scenario with persistent DH parameters and a private
113 key that is shared between multiple clients.
115 This only affects processors that support the BMI1, BMI2 and ADX extensions
116 like Intel Broadwell (5th generation) and later or AMD Ryzen.
118 This issue was reported to OpenSSL by the OSS-Fuzz project.
122 *) Malformed X.509 IPAddressFamily could cause OOB read
124 If an X.509 certificate has a malformed IPAddressFamily extension,
125 OpenSSL could do a one-byte buffer overread. The most likely result
126 would be an erroneous display of the certificate in text format.
128 This issue was reported to OpenSSL by the OSS-Fuzz project.
132 Changes between 1.0.2k and 1.0.2l [25 May 2017]
134 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
135 platform rather than 'mingw'.
138 Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
140 *) Truncated packet could crash via OOB read
142 If one side of an SSL/TLS path is running on a 32-bit host and a specific
143 cipher is being used, then a truncated packet can cause that host to
144 perform an out-of-bounds read, usually resulting in a crash.
146 This issue was reported to OpenSSL by Robert Święcki of Google.
150 *) BN_mod_exp may produce incorrect results on x86_64
152 There is a carry propagating bug in the x86_64 Montgomery squaring
153 procedure. No EC algorithms are affected. Analysis suggests that attacks
154 against RSA and DSA as a result of this defect would be very difficult to
155 perform and are not believed likely. Attacks against DH are considered just
156 feasible (although very difficult) because most of the work necessary to
157 deduce information about a private key may be performed offline. The amount
158 of resources required for such an attack would be very significant and
159 likely only accessible to a limited number of attackers. An attacker would
160 additionally need online access to an unpatched system using the target
161 private key in a scenario with persistent DH parameters and a private
162 key that is shared between multiple clients. For example this can occur by
163 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
164 similar to CVE-2015-3193 but must be treated as a separate problem.
166 This issue was reported to OpenSSL by the OSS-Fuzz project.
170 *) Montgomery multiplication may produce incorrect results
172 There is a carry propagating bug in the Broadwell-specific Montgomery
173 multiplication procedure that handles input lengths divisible by, but
174 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
175 and DH private keys are impossible. This is because the subroutine in
176 question is not used in operations with the private key itself and an input
177 of the attacker's direct choice. Otherwise the bug can manifest itself as
178 transient authentication and key negotiation failures or reproducible
179 erroneous outcome of public-key operations with specially crafted input.
180 Among EC algorithms only Brainpool P-512 curves are affected and one
181 presumably can attack ECDH key negotiation. Impact was not analyzed in
182 detail, because pre-requisites for attack are considered unlikely. Namely
183 multiple clients have to choose the curve in question and the server has to
184 share the private key among them, neither of which is default behaviour.
185 Even then only clients that chose the curve will be affected.
187 This issue was publicly reported as transient failures and was not
188 initially recognized as a security issue. Thanks to Richard Morgan for
189 providing reproducible case.
193 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
194 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
195 prevent issues where no progress is being made and the peer continually
196 sends unrecognised record types, using up resources processing them.
199 Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
201 *) Missing CRL sanity check
203 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
204 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
205 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
207 This issue only affects the OpenSSL 1.0.2i
211 Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
213 *) OCSP Status Request extension unbounded memory growth
215 A malicious client can send an excessively large OCSP Status Request
216 extension. If that client continually requests renegotiation, sending a
217 large OCSP Status Request extension each time, then there will be unbounded
218 memory growth on the server. This will eventually lead to a Denial Of
219 Service attack through memory exhaustion. Servers with a default
220 configuration are vulnerable even if they do not support OCSP. Builds using
221 the "no-ocsp" build time option are not affected.
223 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
227 *) In order to mitigate the SWEET32 attack, the DES ciphers were moved from
230 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
235 *) OOB write in MDC2_Update()
237 An overflow can occur in MDC2_Update() either if called directly or
238 through the EVP_DigestUpdate() function using MDC2. If an attacker
239 is able to supply very large amounts of input data after a previous
240 call to EVP_EncryptUpdate() with a partial block then a length check
241 can overflow resulting in a heap corruption.
243 The amount of data needed is comparable to SIZE_MAX which is impractical
246 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
250 *) Malformed SHA512 ticket DoS
252 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
253 DoS attack where a malformed ticket will result in an OOB read which will
256 The use of SHA512 in TLS session tickets is comparatively rare as it requires
257 a custom server callback and ticket lookup mechanism.
259 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
263 *) OOB write in BN_bn2dec()
265 The function BN_bn2dec() does not check the return value of BN_div_word().
266 This can cause an OOB write if an application uses this function with an
267 overly large BIGNUM. This could be a problem if an overly large certificate
268 or CRL is printed out from an untrusted source. TLS is not affected because
269 record limits will reject an oversized certificate before it is parsed.
271 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
275 *) OOB read in TS_OBJ_print_bio()
277 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
278 the total length the OID text representation would use and not the amount
279 of data written. This will result in OOB reads when large OIDs are
282 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
286 *) Pointer arithmetic undefined behaviour
288 Avoid some undefined pointer arithmetic
290 A common idiom in the codebase is to check limits in the following manner:
293 Where "p" points to some malloc'd data of SIZE bytes and
296 "len" here could be from some externally supplied data (e.g. from a TLS
299 The rules of C pointer arithmetic are such that "p + len" is only well
300 defined where len <= SIZE. Therefore the above idiom is actually
303 For example this could cause problems if some malloc implementation
304 provides an address for "p" such that "p + len" actually overflows for
305 values of len that are too big and therefore p + len < limit.
307 This issue was reported to OpenSSL by Guido Vranken
311 *) Constant time flag not preserved in DSA signing
313 Operations in the DSA signing algorithm should run in constant time in
314 order to avoid side channel attacks. A flaw in the OpenSSL DSA
315 implementation means that a non-constant time codepath is followed for
316 certain operations. This has been demonstrated through a cache-timing
317 attack to be sufficient for an attacker to recover the private DSA key.
319 This issue was reported by César Pereida (Aalto University), Billy Brumley
320 (Tampere University of Technology), and Yuval Yarom (The University of
325 *) DTLS buffered message DoS
327 In a DTLS connection where handshake messages are delivered out-of-order
328 those messages that OpenSSL is not yet ready to process will be buffered
329 for later use. Under certain circumstances, a flaw in the logic means that
330 those messages do not get removed from the buffer even though the handshake
331 has been completed. An attacker could force up to approx. 15 messages to
332 remain in the buffer when they are no longer required. These messages will
333 be cleared when the DTLS connection is closed. The default maximum size for
334 a message is 100k. Therefore the attacker could force an additional 1500k
335 to be consumed per connection. By opening many simulataneous connections an
336 attacker could cause a DoS attack through memory exhaustion.
338 This issue was reported to OpenSSL by Quan Luo.
342 *) DTLS replay protection DoS
344 A flaw in the DTLS replay attack protection mechanism means that records
345 that arrive for future epochs update the replay protection "window" before
346 the MAC for the record has been validated. This could be exploited by an
347 attacker by sending a record for the next epoch (which does not have to
348 decrypt or have a valid MAC), with a very large sequence number. This means
349 that all subsequent legitimate packets are dropped causing a denial of
350 service for a specific DTLS connection.
352 This issue was reported to OpenSSL by the OCAP audit team.
356 *) Certificate message OOB reads
358 In OpenSSL 1.0.2 and earlier some missing message length checks can result
359 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
360 theoretical DoS risk but this has not been observed in practice on common
363 The messages affected are client certificate, client certificate request
364 and server certificate. As a result the attack can only be performed
365 against a client or a server which enables client authentication.
367 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
371 Changes between 1.0.2g and 1.0.2h [3 May 2016]
373 *) Prevent padding oracle in AES-NI CBC MAC check
375 A MITM attacker can use a padding oracle attack to decrypt traffic
376 when the connection uses an AES CBC cipher and the server support
379 This issue was introduced as part of the fix for Lucky 13 padding
380 attack (CVE-2013-0169). The padding check was rewritten to be in
381 constant time by making sure that always the same bytes are read and
382 compared against either the MAC or padding bytes. But it no longer
383 checked that there was enough data to have both the MAC and padding
386 This issue was reported by Juraj Somorovsky using TLS-Attacker.
390 *) Fix EVP_EncodeUpdate overflow
392 An overflow can occur in the EVP_EncodeUpdate() function which is used for
393 Base64 encoding of binary data. If an attacker is able to supply very large
394 amounts of input data then a length check can overflow resulting in a heap
397 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
398 the PEM_write_bio* family of functions. These are mainly used within the
399 OpenSSL command line applications, so any application which processes data
400 from an untrusted source and outputs it as a PEM file should be considered
401 vulnerable to this issue. User applications that call these APIs directly
402 with large amounts of untrusted data may also be vulnerable.
404 This issue was reported by Guido Vranken.
408 *) Fix EVP_EncryptUpdate overflow
410 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
411 is able to supply very large amounts of input data after a previous call to
412 EVP_EncryptUpdate() with a partial block then a length check can overflow
413 resulting in a heap corruption. Following an analysis of all OpenSSL
414 internal usage of the EVP_EncryptUpdate() function all usage is one of two
415 forms. The first form is where the EVP_EncryptUpdate() call is known to be
416 the first called function after an EVP_EncryptInit(), and therefore that
417 specific call must be safe. The second form is where the length passed to
418 EVP_EncryptUpdate() can be seen from the code to be some small value and
419 therefore there is no possibility of an overflow. Since all instances are
420 one of these two forms, it is believed that there can be no overflows in
421 internal code due to this problem. It should be noted that
422 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
423 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
424 of these calls have also been analysed too and it is believed there are no
425 instances in internal usage where an overflow could occur.
427 This issue was reported by Guido Vranken.
431 *) Prevent ASN.1 BIO excessive memory allocation
433 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
434 a short invalid encoding can casuse allocation of large amounts of memory
435 potentially consuming excessive resources or exhausting memory.
437 Any application parsing untrusted data through d2i BIO functions is
438 affected. The memory based functions such as d2i_X509() are *not* affected.
439 Since the memory based functions are used by the TLS library, TLS
440 applications are not affected.
442 This issue was reported by Brian Carpenter.
448 ASN1 Strings that are over 1024 bytes can cause an overread in applications
449 using the X509_NAME_oneline() function on EBCDIC systems. This could result
450 in arbitrary stack data being returned in the buffer.
452 This issue was reported by Guido Vranken.
456 *) Modify behavior of ALPN to invoke callback after SNI/servername
457 callback, such that updates to the SSL_CTX affect ALPN.
460 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
464 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
465 methods are enabled and ssl2 is disabled the methods return NULL.
468 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
470 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
471 Builds that are not configured with "enable-weak-ssl-ciphers" will not
472 provide any "EXPORT" or "LOW" strength ciphers.
475 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
476 is by default disabled at build-time. Builds that are not configured with
477 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
478 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
479 will need to explicitly call either of:
481 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
483 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
485 as appropriate. Even if either of those is used, or the application
486 explicitly uses the version-specific SSLv2_method() or its client and
487 server variants, SSLv2 ciphers vulnerable to exhaustive search key
488 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
489 ciphers, and SSLv2 56-bit DES are no longer available.
493 *) Fix a double-free in DSA code
495 A double free bug was discovered when OpenSSL parses malformed DSA private
496 keys and could lead to a DoS attack or memory corruption for applications
497 that receive DSA private keys from untrusted sources. This scenario is
500 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
505 *) Disable SRP fake user seed to address a server memory leak.
507 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
509 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
510 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
511 was changed to ignore the "fake user" SRP seed, even if the seed
514 Users should use SRP_VBASE_get1_by_user instead. Note that in
515 SRP_VBASE_get1_by_user, caller must free the returned value. Note
516 also that even though configuring the SRP seed attempts to hide
517 invalid usernames by continuing the handshake with fake
518 credentials, this behaviour is not constant time and no strong
519 guarantees are made that the handshake is indistinguishable from
520 that of a valid user.
524 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
526 In the BN_hex2bn function the number of hex digits is calculated using an
527 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
528 large values of |i| this can result in |bn_expand| not allocating any
529 memory because |i * 4| is negative. This can leave the internal BIGNUM data
530 field as NULL leading to a subsequent NULL ptr deref. For very large values
531 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
532 In this case memory is allocated to the internal BIGNUM data field, but it
533 is insufficiently sized leading to heap corruption. A similar issue exists
534 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
535 is ever called by user applications with very large untrusted hex/dec data.
536 This is anticipated to be a rare occurrence.
538 All OpenSSL internal usage of these functions use data that is not expected
539 to be untrusted, e.g. config file data or application command line
540 arguments. If user developed applications generate config file data based
541 on untrusted data then it is possible that this could also lead to security
542 consequences. This is also anticipated to be rare.
544 This issue was reported to OpenSSL by Guido Vranken.
548 *) Fix memory issues in BIO_*printf functions
550 The internal |fmtstr| function used in processing a "%s" format string in
551 the BIO_*printf functions could overflow while calculating the length of a
552 string and cause an OOB read when printing very long strings.
554 Additionally the internal |doapr_outch| function can attempt to write to an
555 OOB memory location (at an offset from the NULL pointer) in the event of a
556 memory allocation failure. In 1.0.2 and below this could be caused where
557 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
558 could be in processing a very long "%s" format string. Memory leaks can
561 The first issue may mask the second issue dependent on compiler behaviour.
562 These problems could enable attacks where large amounts of untrusted data
563 is passed to the BIO_*printf functions. If applications use these functions
564 in this way then they could be vulnerable. OpenSSL itself uses these
565 functions when printing out human-readable dumps of ASN.1 data. Therefore
566 applications that print this data could be vulnerable if the data is from
567 untrusted sources. OpenSSL command line applications could also be
568 vulnerable where they print out ASN.1 data, or if untrusted data is passed
569 as command line arguments.
571 Libssl is not considered directly vulnerable. Additionally certificates etc
572 received via remote connections via libssl are also unlikely to be able to
573 trigger these issues because of message size limits enforced within libssl.
575 This issue was reported to OpenSSL Guido Vranken.
579 *) Side channel attack on modular exponentiation
581 A side-channel attack was found which makes use of cache-bank conflicts on
582 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
583 of RSA keys. The ability to exploit this issue is limited as it relies on
584 an attacker who has control of code in a thread running on the same
585 hyper-threaded core as the victim thread which is performing decryptions.
587 This issue was reported to OpenSSL by Yuval Yarom, The University of
588 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
589 Nadia Heninger, University of Pennsylvania with more information at
590 http://cachebleed.info.
594 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
595 if no keysize is specified with default_bits. This fixes an
596 omission in an earlier change that changed all RSA/DSA key generation
597 apps to use 2048 bits by default.
600 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
602 *) DH small subgroups
604 Historically OpenSSL only ever generated DH parameters based on "safe"
605 primes. More recently (in version 1.0.2) support was provided for
606 generating X9.42 style parameter files such as those required for RFC 5114
607 support. The primes used in such files may not be "safe". Where an
608 application is using DH configured with parameters based on primes that are
609 not "safe" then an attacker could use this fact to find a peer's private
610 DH exponent. This attack requires that the attacker complete multiple
611 handshakes in which the peer uses the same private DH exponent. For example
612 this could be used to discover a TLS server's private DH exponent if it's
613 reusing the private DH exponent or it's using a static DH ciphersuite.
615 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
616 TLS. It is not on by default. If the option is not set then the server
617 reuses the same private DH exponent for the life of the server process and
618 would be vulnerable to this attack. It is believed that many popular
619 applications do set this option and would therefore not be at risk.
621 The fix for this issue adds an additional check where a "q" parameter is
622 available (as is the case in X9.42 based parameters). This detects the
623 only known attack, and is the only possible defense for static DH
624 ciphersuites. This could have some performance impact.
626 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
627 default and cannot be disabled. This could have some performance impact.
629 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
633 *) SSLv2 doesn't block disabled ciphers
635 A malicious client can negotiate SSLv2 ciphers that have been disabled on
636 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
637 been disabled, provided that the SSLv2 protocol was not also disabled via
640 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
641 and Sebastian Schinzel.
645 *) Reject DH handshakes with parameters shorter than 1024 bits.
648 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
650 *) BN_mod_exp may produce incorrect results on x86_64
652 There is a carry propagating bug in the x86_64 Montgomery squaring
653 procedure. No EC algorithms are affected. Analysis suggests that attacks
654 against RSA and DSA as a result of this defect would be very difficult to
655 perform and are not believed likely. Attacks against DH are considered just
656 feasible (although very difficult) because most of the work necessary to
657 deduce information about a private key may be performed offline. The amount
658 of resources required for such an attack would be very significant and
659 likely only accessible to a limited number of attackers. An attacker would
660 additionally need online access to an unpatched system using the target
661 private key in a scenario with persistent DH parameters and a private
662 key that is shared between multiple clients. For example this can occur by
663 default in OpenSSL DHE based SSL/TLS ciphersuites.
665 This issue was reported to OpenSSL by Hanno Böck.
669 *) Certificate verify crash with missing PSS parameter
671 The signature verification routines will crash with a NULL pointer
672 dereference if presented with an ASN.1 signature using the RSA PSS
673 algorithm and absent mask generation function parameter. Since these
674 routines are used to verify certificate signature algorithms this can be
675 used to crash any certificate verification operation and exploited in a
676 DoS attack. Any application which performs certificate verification is
677 vulnerable including OpenSSL clients and servers which enable client
680 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
684 *) X509_ATTRIBUTE memory leak
686 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
687 memory. This structure is used by the PKCS#7 and CMS routines so any
688 application which reads PKCS#7 or CMS data from untrusted sources is
689 affected. SSL/TLS is not affected.
691 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
696 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
697 This changes the decoding behaviour for some invalid messages,
698 though the change is mostly in the more lenient direction, and
699 legacy behaviour is preserved as much as possible.
702 *) In DSA_generate_parameters_ex, if the provided seed is too short,
703 use a random seed, as already documented.
704 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
706 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
708 *) Alternate chains certificate forgery
710 During certificate verfification, OpenSSL will attempt to find an
711 alternative certificate chain if the first attempt to build such a chain
712 fails. An error in the implementation of this logic can mean that an
713 attacker could cause certain checks on untrusted certificates to be
714 bypassed, such as the CA flag, enabling them to use a valid leaf
715 certificate to act as a CA and "issue" an invalid certificate.
717 This issue was reported to OpenSSL by Adam Langley/David Benjamin
722 *) Race condition handling PSK identify hint
724 If PSK identity hints are received by a multi-threaded client then
725 the values are wrongly updated in the parent SSL_CTX structure. This can
726 result in a race condition potentially leading to a double free of the
731 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
733 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
734 incompatibility in the handling of HMAC. The previous ABI has now been
737 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
739 *) Malformed ECParameters causes infinite loop
741 When processing an ECParameters structure OpenSSL enters an infinite loop
742 if the curve specified is over a specially malformed binary polynomial
745 This can be used to perform denial of service against any
746 system which processes public keys, certificate requests or
747 certificates. This includes TLS clients and TLS servers with
748 client authentication enabled.
750 This issue was reported to OpenSSL by Joseph Barr-Pixton.
754 *) Exploitable out-of-bounds read in X509_cmp_time
756 X509_cmp_time does not properly check the length of the ASN1_TIME
757 string and can read a few bytes out of bounds. In addition,
758 X509_cmp_time accepts an arbitrary number of fractional seconds in the
761 An attacker can use this to craft malformed certificates and CRLs of
762 various sizes and potentially cause a segmentation fault, resulting in
763 a DoS on applications that verify certificates or CRLs. TLS clients
764 that verify CRLs are affected. TLS clients and servers with client
765 authentication enabled may be affected if they use custom verification
768 This issue was reported to OpenSSL by Robert Swiecki (Google), and
769 independently by Hanno Böck.
773 *) PKCS7 crash with missing EnvelopedContent
775 The PKCS#7 parsing code does not handle missing inner EncryptedContent
776 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
777 with missing content and trigger a NULL pointer dereference on parsing.
779 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
780 structures from untrusted sources are affected. OpenSSL clients and
781 servers are not affected.
783 This issue was reported to OpenSSL by Michal Zalewski (Google).
787 *) CMS verify infinite loop with unknown hash function
789 When verifying a signedData message the CMS code can enter an infinite loop
790 if presented with an unknown hash function OID. This can be used to perform
791 denial of service against any system which verifies signedData messages using
793 This issue was reported to OpenSSL by Johannes Bauer.
797 *) Race condition handling NewSessionTicket
799 If a NewSessionTicket is received by a multi-threaded client when attempting to
800 reuse a previous ticket then a race condition can occur potentially leading to
801 a double free of the ticket data.
805 *) Removed support for the two export grade static DH ciphersuites
806 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
807 were newly added (along with a number of other static DH ciphersuites) to
808 1.0.2. However the two export ones have *never* worked since they were
809 introduced. It seems strange in any case to be adding new export
810 ciphersuites, and given "logjam" it also does not seem correct to fix them.
813 *) Only support 256-bit or stronger elliptic curves with the
814 'ecdh_auto' setting (server) or by default (client). Of supported
815 curves, prefer P-256 (both).
818 *) Reject DH handshakes with parameters shorter than 768 bits.
819 [Kurt Roeckx and Emilia Kasper]
821 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
823 *) ClientHello sigalgs DoS fix
825 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
826 invalid signature algorithms extension a NULL pointer dereference will
827 occur. This can be exploited in a DoS attack against the server.
829 This issue was was reported to OpenSSL by David Ramos of Stanford
832 [Stephen Henson and Matt Caswell]
834 *) Multiblock corrupted pointer fix
836 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
837 feature only applies on 64 bit x86 architecture platforms that support AES
838 NI instructions. A defect in the implementation of "multiblock" can cause
839 OpenSSL's internal write buffer to become incorrectly set to NULL when
840 using non-blocking IO. Typically, when the user application is using a
841 socket BIO for writing, this will only result in a failed connection.
842 However if some other BIO is used then it is likely that a segmentation
843 fault will be triggered, thus enabling a potential DoS attack.
845 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
849 *) Segmentation fault in DTLSv1_listen fix
851 The DTLSv1_listen function is intended to be stateless and processes the
852 initial ClientHello from many peers. It is common for user code to loop
853 over the call to DTLSv1_listen until a valid ClientHello is received with
854 an associated cookie. A defect in the implementation of DTLSv1_listen means
855 that state is preserved in the SSL object from one invocation to the next
856 that can lead to a segmentation fault. Errors processing the initial
857 ClientHello can trigger this scenario. An example of such an error could be
858 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
861 This issue was reported to OpenSSL by Per Allansson.
865 *) Segmentation fault in ASN1_TYPE_cmp fix
867 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
868 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
869 certificate signature algorithm consistency this can be used to crash any
870 certificate verification operation and exploited in a DoS attack. Any
871 application which performs certificate verification is vulnerable including
872 OpenSSL clients and servers which enable client authentication.
876 *) Segmentation fault for invalid PSS parameters fix
878 The signature verification routines will crash with a NULL pointer
879 dereference if presented with an ASN.1 signature using the RSA PSS
880 algorithm and invalid parameters. Since these routines are used to verify
881 certificate signature algorithms this can be used to crash any
882 certificate verification operation and exploited in a DoS attack. Any
883 application which performs certificate verification is vulnerable including
884 OpenSSL clients and servers which enable client authentication.
886 This issue was was reported to OpenSSL by Brian Carpenter.
890 *) ASN.1 structure reuse memory corruption fix
892 Reusing a structure in ASN.1 parsing may allow an attacker to cause
893 memory corruption via an invalid write. Such reuse is and has been
894 strongly discouraged and is believed to be rare.
896 Applications that parse structures containing CHOICE or ANY DEFINED BY
897 components may be affected. Certificate parsing (d2i_X509 and related
898 functions) are however not affected. OpenSSL clients and servers are
903 *) PKCS7 NULL pointer dereferences fix
905 The PKCS#7 parsing code does not handle missing outer ContentInfo
906 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
907 missing content and trigger a NULL pointer dereference on parsing.
909 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
910 otherwise parse PKCS#7 structures from untrusted sources are
911 affected. OpenSSL clients and servers are not affected.
913 This issue was reported to OpenSSL by Michal Zalewski (Google).
917 *) DoS via reachable assert in SSLv2 servers fix
919 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
920 servers that both support SSLv2 and enable export cipher suites by sending
921 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
923 This issue was discovered by Sean Burford (Google) and Emilia Käsper
924 (OpenSSL development team).
928 *) Empty CKE with client auth and DHE fix
930 If client auth is used then a server can seg fault in the event of a DHE
931 ciphersuite being selected and a zero length ClientKeyExchange message
932 being sent by the client. This could be exploited in a DoS attack.
936 *) Handshake with unseeded PRNG fix
938 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
939 with an unseeded PRNG. The conditions are:
940 - The client is on a platform where the PRNG has not been seeded
941 automatically, and the user has not seeded manually
942 - A protocol specific client method version has been used (i.e. not
943 SSL_client_methodv23)
944 - A ciphersuite is used that does not require additional random data from
945 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
947 If the handshake succeeds then the client random that has been used will
948 have been generated from a PRNG with insufficient entropy and therefore the
949 output may be predictable.
951 For example using the following command with an unseeded openssl will
952 succeed on an unpatched platform:
954 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
958 *) Use After Free following d2i_ECPrivatekey error fix
960 A malformed EC private key file consumed via the d2i_ECPrivateKey function
961 could cause a use after free condition. This, in turn, could cause a double
962 free in several private key parsing functions (such as d2i_PrivateKey
963 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
964 for applications that receive EC private keys from untrusted
965 sources. This scenario is considered rare.
967 This issue was discovered by the BoringSSL project and fixed in their
972 *) X509_to_X509_REQ NULL pointer deref fix
974 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
975 the certificate key is invalid. This function is rarely used in practice.
977 This issue was discovered by Brian Carpenter.
981 *) Removed the export ciphers from the DEFAULT ciphers
984 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
986 *) Change RSA and DH/DSA key generation apps to generate 2048-bit
990 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
991 ARMv5 through ARMv8, as opposite to "locking" it to single one.
992 So far those who have to target multiple plaforms would compromise
993 and argue that binary targeting say ARMv5 would still execute on
994 ARMv8. "Universal" build resolves this compromise by providing
995 near-optimal performance even on newer platforms.
998 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
999 (other platforms pending).
1000 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1002 *) Add support for the SignedCertificateTimestampList certificate and
1003 OCSP response extensions from RFC6962.
1006 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1007 for corner cases. (Certain input points at infinity could lead to
1008 bogus results, with non-infinity inputs mapped to infinity too.)
1011 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1012 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1013 common cases are optimized and there still is room for further
1014 improvements. Vector Permutation AES for Altivec is also added.
1017 *) Add support for little-endian ppc64 Linux target.
1018 [Marcelo Cerri (IBM)]
1020 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1021 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1022 are optimized and there still is room for further improvements.
1023 Both 32- and 64-bit modes are supported.
1024 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1026 *) Improved ARMv7 NEON support.
1029 *) Support for SPARC Architecture 2011 crypto extensions, first
1030 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1031 SHA256/512, MD5, GHASH and modular exponentiation.
1032 [Andy Polyakov, David Miller]
1034 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1036 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1038 *) Support for new and upcoming Intel processors, including AVX2,
1039 BMI and SHA ISA extensions. This includes additional "stitched"
1040 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1043 This work was sponsored by Intel Corp.
1046 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1047 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1048 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1051 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1052 this fixes a limiation in previous versions of OpenSSL.
1055 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1056 MGF1 digest and OAEP label.
1059 *) Add EVP support for key wrapping algorithms, to avoid problems with
1060 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1061 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1062 algorithms and include tests cases.
1065 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1067 [Douglas E. Engert, Steve Henson]
1069 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1070 difference in days and seconds between two tm or ASN1_TIME structures.
1073 *) Add -rev test option to s_server to just reverse order of characters
1074 received by client and send back to server. Also prints an abbreviated
1075 summary of the connection parameters.
1078 *) New option -brief for s_client and s_server to print out a brief summary
1079 of connection parameters.
1082 *) Add callbacks for arbitrary TLS extensions.
1083 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1085 *) New option -crl_download in several openssl utilities to download CRLs
1086 from CRLDP extension in certificates.
1089 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1092 *) New function X509_CRL_diff to generate a delta CRL from the difference
1093 of two full CRLs. Add support to "crl" utility.
1096 *) New functions to set lookup_crls function and to retrieve
1097 X509_STORE from X509_STORE_CTX.
1100 *) Print out deprecated issuer and subject unique ID fields in
1104 *) Extend OCSP I/O functions so they can be used for simple general purpose
1105 HTTP as well as OCSP. New wrapper function which can be used to download
1106 CRLs using the OCSP API.
1109 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1112 *) SSL_CONF* functions. These provide a common framework for application
1113 configuration using configuration files or command lines.
1116 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1117 message callback and prints the results. Needs compile time option
1118 "enable-ssl-trace". New options to s_client and s_server to enable
1122 *) New ctrl and macro to retrieve supported points extensions.
1123 Print out extension in s_server and s_client.
1126 *) New functions to retrieve certificate signature and signature
1130 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1134 *) New Suite B modes for TLS code. These use and enforce the requirements
1135 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1136 only use Suite B curves. The Suite B modes can be set by using the
1137 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1140 *) New chain verification flags for Suite B levels of security. Check
1141 algorithms are acceptable when flags are set in X509_verify_cert.
1144 *) Make tls1_check_chain return a set of flags indicating checks passed
1145 by a certificate chain. Add additional tests to handle client
1146 certificates: checks for matching certificate type and issuer name
1150 *) If an attempt is made to use a signature algorithm not in the peer
1151 preference list abort the handshake. If client has no suitable
1152 signature algorithms in response to a certificate request do not
1153 use the certificate.
1156 *) If server EC tmp key is not in client preference list abort handshake.
1159 *) Add support for certificate stores in CERT structure. This makes it
1160 possible to have different stores per SSL structure or one store in
1161 the parent SSL_CTX. Include distint stores for certificate chain
1162 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1163 to build and store a certificate chain in CERT structure: returing
1164 an error if the chain cannot be built: this will allow applications
1165 to test if a chain is correctly configured.
1167 Note: if the CERT based stores are not set then the parent SSL_CTX
1168 store is used to retain compatibility with existing behaviour.
1172 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1173 mask based on the current session, check mask when sending client
1174 hello and checking the requested ciphersuite.
1177 *) New ctrls to retrieve and set certificate types in a certificate
1178 request message. Print out received values in s_client. If certificate
1179 types is not set with custom values set sensible values based on
1180 supported signature algorithms.
1183 *) Support for distinct client and server supported signature algorithms.
1186 *) Add certificate callback. If set this is called whenever a certificate
1187 is required by client or server. An application can decide which
1188 certificate chain to present based on arbitrary criteria: for example
1189 supported signature algorithms. Add very simple example to s_server.
1190 This fixes many of the problems and restrictions of the existing client
1191 certificate callback: for example you can now clear an existing
1192 certificate and specify the whole chain.
1195 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1196 the certificate can be used for (if anything). Set valid_flags field
1197 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1198 to have similar checks in it.
1200 Add new "cert_flags" field to CERT structure and include a "strict mode".
1201 This enforces some TLS certificate requirements (such as only permitting
1202 certificate signature algorithms contained in the supported algorithms
1203 extension) which some implementations ignore: this option should be used
1204 with caution as it could cause interoperability issues.
1207 *) Update and tidy signature algorithm extension processing. Work out
1208 shared signature algorithms based on preferences and peer algorithms
1209 and print them out in s_client and s_server. Abort handshake if no
1210 shared signature algorithms.
1213 *) Add new functions to allow customised supported signature algorithms
1214 for SSL and SSL_CTX structures. Add options to s_client and s_server
1218 *) New function SSL_certs_clear() to delete all references to certificates
1219 from an SSL structure. Before this once a certificate had been added
1220 it couldn't be removed.
1223 *) Integrate hostname, email address and IP address checking with certificate
1224 verification. New verify options supporting checking in opensl utility.
1227 *) Fixes and wildcard matching support to hostname and email checking
1228 functions. Add manual page.
1229 [Florian Weimer (Red Hat Product Security Team)]
1231 *) New functions to check a hostname email or IP address against a
1232 certificate. Add options x509 utility to print results of checks against
1236 *) Fix OCSP checking.
1237 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1239 *) Initial experimental support for explicitly trusted non-root CAs.
1240 OpenSSL still tries to build a complete chain to a root but if an
1241 intermediate CA has a trust setting included that is used. The first
1242 setting is used: whether to trust (e.g., -addtrust option to the x509
1246 *) Add -trusted_first option which attempts to find certificates in the
1247 trusted store even if an untrusted chain is also supplied.
1250 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1251 platform support for Linux and Android.
1254 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1257 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1258 When in FIPS mode the approved implementations are used as normal,
1259 when not in FIPS mode the internal unapproved versions are used instead.
1260 This means that the FIPS capable OpenSSL isn't forced to use the
1261 (often lower perfomance) FIPS implementations outside FIPS mode.
1264 *) Transparently support X9.42 DH parameters when calling
1265 PEM_read_bio_DHparameters. This means existing applications can handle
1266 the new parameter format automatically.
1269 *) Initial experimental support for X9.42 DH parameter format: mainly
1270 to support use of 'q' parameter for RFC5114 parameters.
1273 *) Add DH parameters from RFC5114 including test data to dhtest.
1276 *) Support for automatic EC temporary key parameter selection. If enabled
1277 the most preferred EC parameters are automatically used instead of
1278 hardcoded fixed parameters. Now a server just has to call:
1279 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1280 support ECDH and use the most appropriate parameters.
1283 *) Enhance and tidy EC curve and point format TLS extension code. Use
1284 static structures instead of allocation if default values are used.
1285 New ctrls to set curves we wish to support and to retrieve shared curves.
1286 Print out shared curves in s_server. New options to s_server and s_client
1287 to set list of supported curves.
1290 *) New ctrls to retrieve supported signature algorithms and
1291 supported curve values as an array of NIDs. Extend openssl utility
1292 to print out received values.
1295 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1296 between NIDs and the more common NIST names such as "P-256". Enhance
1297 ecparam utility and ECC method to recognise the NIST names for curves.
1300 *) Enhance SSL/TLS certificate chain handling to support different
1301 chains for each certificate instead of one chain in the parent SSL_CTX.
1304 *) Support for fixed DH ciphersuite client authentication: where both
1305 server and client use DH certificates with common parameters.
1308 *) Support for fixed DH ciphersuites: those requiring DH server
1312 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1314 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1315 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1316 X509_CINF_get_signature were reverted post internal team review.
1318 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1320 *) Build fixes for the Windows and OpenVMS platforms
1321 [Matt Caswell and Richard Levitte]
1323 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1325 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1326 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1327 dereference. This could lead to a Denial Of Service attack. Thanks to
1328 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1332 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1333 dtls1_buffer_record function under certain conditions. In particular this
1334 could occur if an attacker sent repeated DTLS records with the same
1335 sequence number but for the next epoch. The memory leak could be exploited
1336 by an attacker in a Denial of Service attack through memory exhaustion.
1337 Thanks to Chris Mueller for reporting this issue.
1341 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1342 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1343 method would be set to NULL which could later result in a NULL pointer
1344 dereference. Thanks to Frank Schmirler for reporting this issue.
1348 *) Abort handshake if server key exchange message is omitted for ephemeral
1351 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1352 reporting this issue.
1356 *) Remove non-export ephemeral RSA code on client and server. This code
1357 violated the TLS standard by allowing the use of temporary RSA keys in
1358 non-export ciphersuites and could be used by a server to effectively
1359 downgrade the RSA key length used to a value smaller than the server
1360 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1361 INRIA or reporting this issue.
1365 *) Fixed issue where DH client certificates are accepted without verification.
1366 An OpenSSL server will accept a DH certificate for client authentication
1367 without the certificate verify message. This effectively allows a client to
1368 authenticate without the use of a private key. This only affects servers
1369 which trust a client certificate authority which issues certificates
1370 containing DH keys: these are extremely rare and hardly ever encountered.
1371 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1376 *) Ensure that the session ID context of an SSL is updated when its
1377 SSL_CTX is updated via SSL_set_SSL_CTX.
1379 The session ID context is typically set from the parent SSL_CTX,
1380 and can vary with the CTX.
1383 *) Fix various certificate fingerprint issues.
1385 By using non-DER or invalid encodings outside the signed portion of a
1386 certificate the fingerprint can be changed without breaking the signature.
1387 Although no details of the signed portion of the certificate can be changed
1388 this can cause problems with some applications: e.g. those using the
1389 certificate fingerprint for blacklists.
1391 1. Reject signatures with non zero unused bits.
1393 If the BIT STRING containing the signature has non zero unused bits reject
1394 the signature. All current signature algorithms require zero unused bits.
1396 2. Check certificate algorithm consistency.
1398 Check the AlgorithmIdentifier inside TBS matches the one in the
1399 certificate signature. NB: this will result in signature failure
1400 errors for some broken certificates.
1402 Thanks to Konrad Kraszewski from Google for reporting this issue.
1404 3. Check DSA/ECDSA signatures use DER.
1406 Reencode DSA/ECDSA signatures and compare with the original received
1407 signature. Return an error if there is a mismatch.
1409 This will reject various cases including garbage after signature
1410 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1411 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1412 (negative or with leading zeroes).
1414 Further analysis was conducted and fixes were developed by Stephen Henson
1415 of the OpenSSL core team.
1420 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1421 results on some platforms, including x86_64. This bug occurs at random
1422 with a very low probability, and is not known to be exploitable in any
1423 way, though its exact impact is difficult to determine. Thanks to Pieter
1424 Wuille (Blockstream) who reported this issue and also suggested an initial
1425 fix. Further analysis was conducted by the OpenSSL development team and
1426 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1427 the OpenSSL core team.
1431 *) Do not resume sessions on the server if the negotiated protocol
1432 version does not match the session's version. Resuming with a different
1433 version, while not strictly forbidden by the RFC, is of questionable
1434 sanity and breaks all known clients.
1435 [David Benjamin, Emilia Käsper]
1437 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1438 early CCS messages during renegotiation. (Note that because
1439 renegotiation is encrypted, this early CCS was not exploitable.)
1442 *) Tighten client-side session ticket handling during renegotiation:
1443 ensure that the client only accepts a session ticket if the server sends
1444 the extension anew in the ServerHello. Previously, a TLS client would
1445 reuse the old extension state and thus accept a session ticket if one was
1446 announced in the initial ServerHello.
1448 Similarly, ensure that the client requires a session ticket if one
1449 was advertised in the ServerHello. Previously, a TLS client would
1450 ignore a missing NewSessionTicket message.
1453 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1455 *) SRTP Memory Leak.
1457 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1458 sends a carefully crafted handshake message, to cause OpenSSL to fail
1459 to free up to 64k of memory causing a memory leak. This could be
1460 exploited in a Denial Of Service attack. This issue affects OpenSSL
1461 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1462 whether SRTP is used or configured. Implementations of OpenSSL that
1463 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1465 The fix was developed by the OpenSSL team.
1469 *) Session Ticket Memory Leak.
1471 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1472 integrity of that ticket is first verified. In the event of a session
1473 ticket integrity check failing, OpenSSL will fail to free memory
1474 causing a memory leak. By sending a large number of invalid session
1475 tickets an attacker could exploit this issue in a Denial Of Service
1480 *) Build option no-ssl3 is incomplete.
1482 When OpenSSL is configured with "no-ssl3" as a build option, servers
1483 could accept and complete a SSL 3.0 handshake, and clients could be
1484 configured to send them.
1486 [Akamai and the OpenSSL team]
1488 *) Add support for TLS_FALLBACK_SCSV.
1489 Client applications doing fallback retries should call
1490 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1492 [Adam Langley, Bodo Moeller]
1494 *) Add additional DigestInfo checks.
1496 Reencode DigestInto in DER and check against the original when
1497 verifying RSA signature: this will reject any improperly encoded
1498 DigestInfo structures.
1500 Note: this is a precautionary measure and no attacks are currently known.
1504 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1506 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1507 SRP code can be overrun an internal buffer. Add sanity check that
1508 g, A, B < N to SRP code.
1510 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1511 Group for discovering this issue.
1515 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1516 TLS 1.0 instead of higher protocol versions when the ClientHello message
1517 is badly fragmented. This allows a man-in-the-middle attacker to force a
1518 downgrade to TLS 1.0 even if both the server and the client support a
1519 higher protocol version, by modifying the client's TLS records.
1521 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1522 researching this issue.
1526 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1527 to a denial of service attack. A malicious server can crash the client
1528 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1529 ciphersuite and sending carefully crafted handshake messages.
1531 Thanks to Felix Gröbert (Google) for discovering and researching this
1536 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1537 to leak memory. This can be exploited through a Denial of Service attack.
1538 Thanks to Adam Langley for discovering and researching this issue.
1542 *) An attacker can force openssl to consume large amounts of memory whilst
1543 processing DTLS handshake messages. This can be exploited through a
1544 Denial of Service attack.
1545 Thanks to Adam Langley for discovering and researching this issue.
1549 *) An attacker can force an error condition which causes openssl to crash
1550 whilst processing DTLS packets due to memory being freed twice. This
1551 can be exploited through a Denial of Service attack.
1552 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1557 *) If a multithreaded client connects to a malicious server using a resumed
1558 session and the server sends an ec point format extension it could write
1559 up to 255 bytes to freed memory.
1561 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1566 *) A malicious server can crash an OpenSSL client with a null pointer
1567 dereference (read) by specifying an SRP ciphersuite even though it was not
1568 properly negotiated with the client. This can be exploited through a
1569 Denial of Service attack.
1571 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1572 discovering and researching this issue.
1576 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1577 X509_name_oneline, X509_name_print_ex et al. to leak some information
1578 from the stack. Applications may be affected if they echo pretty printing
1579 output to the attacker.
1581 Thanks to Ivan Fratric (Google) for discovering this issue.
1583 [Emilia Käsper, and Steve Henson]
1585 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1586 for corner cases. (Certain input points at infinity could lead to
1587 bogus results, with non-infinity inputs mapped to infinity too.)
1590 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1592 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1593 handshake can force the use of weak keying material in OpenSSL
1594 SSL/TLS clients and servers.
1596 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1597 researching this issue. (CVE-2014-0224)
1598 [KIKUCHI Masashi, Steve Henson]
1600 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1601 OpenSSL DTLS client the code can be made to recurse eventually crashing
1604 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1606 [Imre Rad, Steve Henson]
1608 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1609 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1610 client or server. This is potentially exploitable to run arbitrary
1611 code on a vulnerable client or server.
1613 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1614 [Jüri Aedla, Steve Henson]
1616 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1617 are subject to a denial of service attack.
1619 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1620 this issue. (CVE-2014-3470)
1621 [Felix Gröbert, Ivan Fratric, Steve Henson]
1623 *) Harmonize version and its documentation. -f flag is used to display
1625 [mancha <mancha1@zoho.com>]
1627 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1628 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1629 [mancha <mancha1@zoho.com>]
1631 *) Fix some double frees. These are not thought to be exploitable.
1632 [mancha <mancha1@zoho.com>]
1634 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1636 *) A missing bounds check in the handling of the TLS heartbeat extension
1637 can be used to reveal up to 64k of memory to a connected client or
1640 Thanks for Neel Mehta of Google Security for discovering this bug and to
1641 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1642 preparing the fix (CVE-2014-0160)
1643 [Adam Langley, Bodo Moeller]
1645 *) Fix for the attack described in the paper "Recovering OpenSSL
1646 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1647 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1648 http://eprint.iacr.org/2014/140
1650 Thanks to Yuval Yarom and Naomi Benger for discovering this
1651 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1652 [Yuval Yarom and Naomi Benger]
1654 *) TLS pad extension: draft-agl-tls-padding-03
1656 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1657 TLS client Hello record length value would otherwise be > 255 and
1658 less that 512 pad with a dummy extension containing zeroes so it
1659 is at least 512 bytes long.
1661 [Adam Langley, Steve Henson]
1663 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1665 *) Fix for TLS record tampering bug. A carefully crafted invalid
1666 handshake could crash OpenSSL with a NULL pointer exception.
1667 Thanks to Anton Johansson for reporting this issues.
1670 *) Keep original DTLS digest and encryption contexts in retransmission
1671 structures so we can use the previous session parameters if they need
1672 to be resent. (CVE-2013-6450)
1675 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1676 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1677 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1678 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1679 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1680 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1681 [Rob Stradling, Adam Langley]
1683 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1685 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1686 supporting platforms or when small records were transferred.
1687 [Andy Polyakov, Steve Henson]
1689 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1691 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1693 This addresses the flaw in CBC record processing discovered by
1694 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1695 at: http://www.isg.rhul.ac.uk/tls/
1697 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1698 Security Group at Royal Holloway, University of London
1699 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1700 Emilia Käsper for the initial patch.
1702 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1704 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1705 ciphersuites which can be exploited in a denial of service attack.
1706 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1707 and detecting this bug and to Wolfgang Ettlinger
1708 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1712 *) Return an error when checking OCSP signatures when key is NULL.
1713 This fixes a DoS attack. (CVE-2013-0166)
1716 *) Make openssl verify return errors.
1717 [Chris Palmer <palmer@google.com> and Ben Laurie]
1719 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1720 the right response is stapled. Also change SSL_get_certificate()
1721 so it returns the certificate actually sent.
1722 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1723 [Rob Stradling <rob.stradling@comodo.com>]
1725 *) Fix possible deadlock when decoding public keys.
1728 *) Don't use TLS 1.0 record version number in initial client hello
1732 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1734 *) Sanity check record length before skipping explicit IV in TLS
1735 1.2, 1.1 and DTLS to fix DoS attack.
1737 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1738 fuzzing as a service testing platform.
1742 *) Initialise tkeylen properly when encrypting CMS messages.
1743 Thanks to Solar Designer of Openwall for reporting this issue.
1746 *) In FIPS mode don't try to use composite ciphers as they are not
1750 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1752 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1753 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1754 mean any application compiled against OpenSSL 1.0.0 headers setting
1755 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1756 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1757 0x10000000L Any application which was previously compiled against
1758 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1759 will need to be recompiled as a result. Letting be results in
1760 inability to disable specifically TLS 1.1 and in client context,
1761 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1764 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1765 disable just protocol X, but all protocols above X *if* there are
1766 protocols *below* X still enabled. In more practical terms it means
1767 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1768 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1769 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1773 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1775 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1776 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1777 in CRYPTO_realloc_clean.
1779 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1780 issue and to Adam Langley <agl@chromium.org> for fixing it.
1782 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1784 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1787 *) Workarounds for some broken servers that "hang" if a client hello
1788 record length exceeds 255 bytes.
1790 1. Do not use record version number > TLS 1.0 in initial client
1791 hello: some (but not all) hanging servers will now work.
1792 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1793 the number of ciphers sent in the client hello. This should be
1794 set to an even number, such as 50, for example by passing:
1795 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1796 Most broken servers should now work.
1797 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1798 TLS 1.2 client support entirely.
1801 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1804 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1806 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1807 STRING form instead of a DigestInfo.
1810 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1811 and the RSA_sign/RSA_verify functions. This was made more apparent when
1812 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1813 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1814 the correct format in RSA_verify so both forms transparently work.
1817 *) Some servers which support TLS 1.0 can choke if we initially indicate
1818 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1819 encrypted premaster secret. As a workaround use the maximum pemitted
1820 client version in client hello, this should keep such servers happy
1821 and still work with previous versions of OpenSSL.
1824 *) Add support for TLS/DTLS heartbeats.
1825 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1827 *) Add support for SCTP.
1828 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1830 *) Improved PRNG seeding for VOS.
1831 [Paul Green <Paul.Green@stratus.com>]
1833 *) Extensive assembler packs updates, most notably:
1835 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1836 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1837 - x86_64: bit-sliced AES implementation;
1838 - ARM: NEON support, contemporary platforms optimizations;
1839 - s390x: z196 support;
1840 - *: GHASH and GF(2^m) multiplication implementations;
1844 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1845 (removal of unnecessary code)
1846 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1848 *) Add TLS key material exporter from RFC 5705.
1851 *) Add DTLS-SRTP negotiation from RFC 5764.
1854 *) Add Next Protocol Negotiation,
1855 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1856 disabled with a no-npn flag to config or Configure. Code donated
1858 [Adam Langley <agl@google.com> and Ben Laurie]
1860 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1861 NIST-P256, NIST-P521, with constant-time single point multiplication on
1862 typical inputs. Compiler support for the nonstandard type __uint128_t is
1863 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1864 Code made available under Apache License version 2.0.
1866 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1867 line to include this in your build of OpenSSL, and run "make depend" (or
1868 "make update"). This enables the following EC_METHODs:
1870 EC_GFp_nistp224_method()
1871 EC_GFp_nistp256_method()
1872 EC_GFp_nistp521_method()
1874 EC_GROUP_new_by_curve_name() will automatically use these (while
1875 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1877 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1879 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1880 all platforms. Move ssize_t definition from e_os.h to the public
1881 header file e_os2.h as it now appears in public header file cms.h
1884 *) New -sigopt option to the ca, req and x509 utilities. Additional
1885 signature parameters can be passed using this option and in
1889 *) Add RSA PSS signing function. This will generate and set the
1890 appropriate AlgorithmIdentifiers for PSS based on those in the
1891 corresponding EVP_MD_CTX structure. No application support yet.
1894 *) Support for companion algorithm specific ASN1 signing routines.
1895 New function ASN1_item_sign_ctx() signs a pre-initialised
1896 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1897 the appropriate parameters.
1900 *) Add new algorithm specific ASN1 verification initialisation function
1901 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1902 handling will be the same no matter what EVP_PKEY_METHOD is used.
1903 Add a PSS handler to support verification of PSS signatures: checked
1904 against a number of sample certificates.
1907 *) Add signature printing for PSS. Add PSS OIDs.
1908 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1910 *) Add algorithm specific signature printing. An individual ASN1 method
1911 can now print out signatures instead of the standard hex dump.
1913 More complex signatures (e.g. PSS) can print out more meaningful
1914 information. Include DSA version that prints out the signature
1918 *) Password based recipient info support for CMS library: implementing
1922 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1923 neatly separates the code into cipher and PBE sections and is required
1924 for some algorithms that split PBES2 into separate pieces (such as
1925 password based CMS).
1928 *) Session-handling fixes:
1929 - Fix handling of connections that are resuming with a session ID,
1930 but also support Session Tickets.
1931 - Fix a bug that suppressed issuing of a new ticket if the client
1932 presented a ticket with an expired session.
1933 - Try to set the ticket lifetime hint to something reasonable.
1934 - Make tickets shorter by excluding irrelevant information.
1935 - On the client side, don't ignore renewed tickets.
1936 [Adam Langley, Bodo Moeller (Google)]
1938 *) Fix PSK session representation.
1941 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1943 This work was sponsored by Intel.
1946 *) Add GCM support to TLS library. Some custom code is needed to split
1947 the IV between the fixed (from PRF) and explicit (from TLS record)
1948 portions. This adds all GCM ciphersuites supported by RFC5288 and
1949 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
1950 add a special AESGCM string for GCM only.
1953 *) Expand range of ctrls for AES GCM. Permit setting invocation
1954 field on decrypt and retrieval of invocation field only on encrypt.
1957 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1958 As required by RFC5289 these ciphersuites cannot be used if for
1959 versions of TLS earlier than 1.2.
1962 *) For FIPS capable OpenSSL interpret a NULL default public key method
1963 as unset and return the appopriate default but do *not* set the default.
1964 This means we can return the appopriate method in applications that
1965 swicth between FIPS and non-FIPS modes.
1968 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1969 ENGINE is used then we cannot handle that in the FIPS module so we
1970 keep original code iff non-FIPS operations are allowed.
1973 *) Add -attime option to openssl utilities.
1974 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1976 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1979 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1980 FIPS EC methods unconditionally for now.
1983 *) New build option no-ec2m to disable characteristic 2 code.
1986 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1987 all cases can be covered as some introduce binary incompatibilities.
1990 *) Redirect RSA operations to FIPS module including keygen,
1991 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1994 *) Add similar low level API blocking to ciphers.
1997 *) Low level digest APIs are not approved in FIPS mode: any attempt
1998 to use these will cause a fatal error. Applications that *really* want
1999 to use them can use the private_* version instead.
2002 *) Redirect cipher operations to FIPS module for FIPS builds.
2005 *) Redirect digest operations to FIPS module for FIPS builds.
2008 *) Update build system to add "fips" flag which will link in fipscanister.o
2009 for static and shared library builds embedding a signature if needed.
2012 *) Output TLS supported curves in preference order instead of numerical
2013 order. This is currently hardcoded for the highest order curves first.
2014 This should be configurable so applications can judge speed vs strength.
2017 *) Add TLS v1.2 server support for client authentication.
2020 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2024 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2025 FIPS modules versions.
2028 *) Add TLS v1.2 client side support for client authentication. Keep cache
2029 of handshake records longer as we don't know the hash algorithm to use
2030 until after the certificate request message is received.
2033 *) Initial TLS v1.2 client support. Add a default signature algorithms
2034 extension including all the algorithms we support. Parse new signature
2035 format in client key exchange. Relax some ECC signing restrictions for
2036 TLS v1.2 as indicated in RFC5246.
2039 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2040 to new signature format when needed using client digest preference.
2041 All server ciphersuites should now work correctly in TLS v1.2. No client
2042 support yet and no support for client certificates.
2045 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2046 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2047 ciphersuites. At present only RSA key exchange ciphersuites work with
2048 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2049 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2050 and version checking.
2053 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2054 with this defined it will not be affected by any changes to ssl internal
2055 structures. Add several utility functions to allow openssl application
2056 to work with OPENSSL_NO_SSL_INTERN defined.
2059 *) A long standing patch to add support for SRP from EdelWeb (Peter
2060 Sylvester and Christophe Renou) was integrated.
2061 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
2062 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
2065 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2068 *) Permit abbreviated handshakes when renegotiating using the function
2069 SSL_renegotiate_abbreviated().
2070 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2072 *) Add call to ENGINE_register_all_complete() to
2073 ENGINE_load_builtin_engines(), so some implementations get used
2074 automatically instead of needing explicit application support.
2077 *) Add support for TLS key exporter as described in RFC5705.
2078 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2080 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2081 a few changes are required:
2083 Add SSL_OP_NO_TLSv1_1 flag.
2084 Add TLSv1_1 methods.
2085 Update version checking logic to handle version 1.1.
2086 Add explicit IV handling (ported from DTLS code).
2087 Add command line options to s_client/s_server.
2090 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2092 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2093 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2094 content decryption and always return the same error. Note: this attack
2095 needs on average 2^20 messages so it only affects automated senders. The
2096 old behaviour can be reenabled in the CMS code by setting the
2097 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2098 an MMA defence is not necessary.
2099 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2100 this issue. (CVE-2012-0884)
2103 *) Fix CVE-2011-4619: make sure we really are receiving a
2104 client hello before rejecting multiple SGC restarts. Thanks to
2105 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2108 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2110 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2111 Thanks to Antonio Martin, Enterprise Secure Access Research and
2112 Development, Cisco Systems, Inc. for discovering this bug and
2113 preparing a fix. (CVE-2012-0050)
2116 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2118 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2119 of the Vaudenay padding oracle attack on CBC mode encryption
2120 which enables an efficient plaintext recovery attack against
2121 the OpenSSL implementation of DTLS. Their attack exploits timing
2122 differences arising during decryption processing. A research
2123 paper describing this attack can be found at:
2124 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2125 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2126 Security Group at Royal Holloway, University of London
2127 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2128 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2129 for preparing the fix. (CVE-2011-4108)
2130 [Robin Seggelmann, Michael Tuexen]
2132 *) Clear bytes used for block padding of SSL 3.0 records.
2134 [Adam Langley (Google)]
2136 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2137 Kadianakis <desnacked@gmail.com> for discovering this issue and
2138 Adam Langley for preparing the fix. (CVE-2011-4619)
2139 [Adam Langley (Google)]
2141 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2142 [Andrey Kulikov <amdeich@gmail.com>]
2144 *) Prevent malformed RFC3779 data triggering an assertion failure.
2145 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2146 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2147 [Rob Austein <sra@hactrn.net>]
2149 *) Improved PRNG seeding for VOS.
2150 [Paul Green <Paul.Green@stratus.com>]
2152 *) Fix ssl_ciph.c set-up race.
2153 [Adam Langley (Google)]
2155 *) Fix spurious failures in ecdsatest.c.
2156 [Emilia Käsper (Google)]
2158 *) Fix the BIO_f_buffer() implementation (which was mixing different
2159 interpretations of the '..._len' fields).
2160 [Adam Langley (Google)]
2162 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2163 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2164 threads won't reuse the same blinding coefficients.
2166 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2167 lock to call BN_BLINDING_invert_ex, and avoids one use of
2168 BN_BLINDING_update for each BN_BLINDING structure (previously,
2169 the last update always remained unused).
2170 [Emilia Käsper (Google)]
2172 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2173 [Bob Buckholz (Google)]
2175 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2177 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2178 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2179 [Kaspar Brand <ossl@velox.ch>]
2181 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2182 for multi-threaded use of ECDH. (CVE-2011-3210)
2183 [Adam Langley (Google)]
2185 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2188 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2189 signature public key algorithm by using OID xref utilities instead.
2190 Before this you could only use some ECC ciphersuites with SHA1 only.
2193 *) Add protection against ECDSA timing attacks as mentioned in the paper
2194 by Billy Bob Brumley and Nicola Tuveri, see:
2196 http://eprint.iacr.org/2011/232.pdf
2198 [Billy Bob Brumley and Nicola Tuveri]
2200 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2202 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2203 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2205 *) Fix bug in string printing code: if *any* escaping is enabled we must
2206 escape the escape character (backslash) or the resulting string is
2210 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2212 *) Disable code workaround for ancient and obsolete Netscape browsers
2213 and servers: an attacker can use it in a ciphersuite downgrade attack.
2214 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2217 *) Fixed J-PAKE implementation error, originally discovered by
2218 Sebastien Martini, further info and confirmation from Stefan
2219 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2222 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2224 *) Fix extension code to avoid race conditions which can result in a buffer
2225 overrun vulnerability: resumed sessions must not be modified as they can
2226 be shared by multiple threads. CVE-2010-3864
2229 *) Fix WIN32 build system to correctly link an ENGINE directory into
2233 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2235 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2237 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2239 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2241 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2242 context. The operation can be customised via the ctrl mechanism in
2243 case ENGINEs want to include additional functionality.
2246 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2249 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2250 output hashes compatible with older versions of OpenSSL.
2251 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2253 *) Fix compression algorithm handling: if resuming a session use the
2254 compression algorithm of the resumed session instead of determining
2255 it from client hello again. Don't allow server to change algorithm.
2258 *) Add load_crls() function to apps tidying load_certs() too. Add option
2259 to verify utility to allow additional CRLs to be included.
2262 *) Update OCSP request code to permit adding custom headers to the request:
2263 some responders need this.
2266 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2268 [Julia Lawall <julia@diku.dk>]
2270 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2271 needlessly dereferenced structures, used obsolete functions and
2272 didn't handle all updated verify codes correctly.
2275 *) Disable MD2 in the default configuration.
2278 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2279 indicate the initial BIO being pushed or popped. This makes it possible
2280 to determine whether the BIO is the one explicitly called or as a result
2281 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2282 it handles reference counts correctly and doesn't zero out the I/O bio
2283 when it is not being explicitly popped. WARNING: applications which
2284 included workarounds for the old buggy behaviour will need to be modified
2285 or they could free up already freed BIOs.
2288 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2289 renaming to all platforms (within the 0.9.8 branch, this was
2290 done conditionally on Netware platforms to avoid a name clash).
2291 [Guenter <lists@gknw.net>]
2293 *) Add ECDHE and PSK support to DTLS.
2294 [Michael Tuexen <tuexen@fh-muenster.de>]
2296 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2300 *) Add "missing" function EVP_MD_flags() (without this the only way to
2301 retrieve a digest flags is by accessing the structure directly. Update
2302 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2303 or cipher is registered as in the "from" argument. Print out all
2304 registered digests in the dgst usage message instead of manually
2305 attempting to work them out.
2308 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2309 this allows the use of compression and extensions. Change default cipher
2310 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2311 by default unless an application cipher string requests it.
2314 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2315 key ids to find matching certificates and keys but some PKCS#12 files
2316 don't follow the (somewhat unwritten) rules and this strategy fails.
2317 Now just gather all certificates together and the first private key
2318 then look for the first certificate that matches the key.
2321 *) Support use of registered digest and cipher names for dgst and cipher
2322 commands instead of having to add each one as a special case. So now
2329 openssl dgst -sha256 foo
2331 and this works for ENGINE based algorithms too.
2335 *) Update Gost ENGINE to support parameter files.
2336 [Victor B. Wagner <vitus@cryptocom.ru>]
2338 *) Support GeneralizedTime in ca utility.
2339 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2341 *) Enhance the hash format used for certificate directory links. The new
2342 form uses the canonical encoding (meaning equivalent names will work
2343 even if they aren't identical) and uses SHA1 instead of MD5. This form
2344 is incompatible with the older format and as a result c_rehash should
2345 be used to rebuild symbolic links.
2348 *) Make PKCS#8 the default write format for private keys, replacing the
2349 traditional format. This form is standardised, more secure and doesn't
2350 include an implicit MD5 dependency.
2353 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2354 committed to OpenSSL should pass this lot as a minimum.
2357 *) Add session ticket override functionality for use by EAP-FAST.
2358 [Jouni Malinen <j@w1.fi>]
2360 *) Modify HMAC functions to return a value. Since these can be implemented
2361 in an ENGINE errors can occur.
2364 *) Type-checked OBJ_bsearch_ex.
2367 *) Type-checked OBJ_bsearch. Also some constification necessitated
2368 by type-checking. Still to come: TXT_DB, bsearch(?),
2369 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2373 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2374 seconds to a tm structure directly, instead of going through OS
2375 specific date routines. This avoids any issues with OS routines such
2376 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2377 and X509_time_adj_ex() to cover the extended range. The existing
2378 X509_time_adj() is still usable and will no longer have any date issues.
2381 *) Delta CRL support. New use deltas option which will attempt to locate
2382 and search any appropriate delta CRLs available.
2384 This work was sponsored by Google.
2387 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2388 code and add additional score elements. Validate alternate CRL paths
2389 as part of the CRL checking and indicate a new error "CRL path validation
2390 error" in this case. Applications wanting additional details can use
2391 the verify callback and check the new "parent" field. If this is not
2392 NULL CRL path validation is taking place. Existing applications wont
2393 see this because it requires extended CRL support which is off by
2396 This work was sponsored by Google.
2399 *) Support for freshest CRL extension.
2401 This work was sponsored by Google.
2404 *) Initial indirect CRL support. Currently only supported in the CRLs
2405 passed directly and not via lookup. Process certificate issuer
2406 CRL entry extension and lookup CRL entries by bother issuer name
2407 and serial number. Check and process CRL issuer entry in IDP extension.
2409 This work was sponsored by Google.
2412 *) Add support for distinct certificate and CRL paths. The CRL issuer
2413 certificate is validated separately in this case. Only enabled if
2414 an extended CRL support flag is set: this flag will enable additional
2415 CRL functionality in future.
2417 This work was sponsored by Google.
2420 *) Add support for policy mappings extension.
2422 This work was sponsored by Google.
2425 *) Fixes to pathlength constraint, self issued certificate handling,
2426 policy processing to align with RFC3280 and PKITS tests.
2428 This work was sponsored by Google.
2431 *) Support for name constraints certificate extension. DN, email, DNS
2432 and URI types are currently supported.
2434 This work was sponsored by Google.
2437 *) To cater for systems that provide a pointer-based thread ID rather
2438 than numeric, deprecate the current numeric thread ID mechanism and
2439 replace it with a structure and associated callback type. This
2440 mechanism allows a numeric "hash" to be extracted from a thread ID in
2441 either case, and on platforms where pointers are larger than 'long',
2442 mixing is done to help ensure the numeric 'hash' is usable even if it
2443 can't be guaranteed unique. The default mechanism is to use "&errno"
2444 as a pointer-based thread ID to distinguish between threads.
2446 Applications that want to provide their own thread IDs should now use
2447 CRYPTO_THREADID_set_callback() to register a callback that will call
2448 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2450 Note that ERR_remove_state() is now deprecated, because it is tied
2451 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2452 to free the current thread's error state should be replaced by
2453 ERR_remove_thread_state(NULL).
2455 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2456 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2457 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2458 application was previously providing a numeric thread callback that
2459 was inappropriate for distinguishing threads, then uniqueness might
2460 have been obtained with &errno that happened immediately in the
2461 intermediate development versions of OpenSSL; this is no longer the
2462 case, the numeric thread callback will now override the automatic use
2464 [Geoff Thorpe, with help from Bodo Moeller]
2466 *) Initial support for different CRL issuing certificates. This covers a
2467 simple case where the self issued certificates in the chain exist and
2468 the real CRL issuer is higher in the existing chain.
2470 This work was sponsored by Google.
2473 *) Removed effectively defunct crypto/store from the build.
2476 *) Revamp of STACK to provide stronger type-checking. Still to come:
2477 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2478 ASN1_STRING, CONF_VALUE.
2481 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2482 RAM on SSL connections. This option can save about 34k per idle SSL.
2485 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2486 STACK, TXT_DB, bsearch, qsort.
2489 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2490 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2491 support for data, signedData, compressedData, digestedData and
2492 encryptedData, envelopedData types included. Scripts to check against
2493 RFC4134 examples draft and interop and consistency checks of many
2494 content types and variants.
2497 *) Add options to enc utility to support use of zlib compression BIO.
2500 *) Extend mk1mf to support importing of options and assembly language
2501 files from Configure script, currently only included in VC-WIN32.
2502 The assembly language rules can now optionally generate the source
2503 files from the associated perl scripts.
2506 *) Implement remaining functionality needed to support GOST ciphersuites.
2507 Interop testing has been performed using CryptoPro implementations.
2508 [Victor B. Wagner <vitus@cryptocom.ru>]
2510 *) s390x assembler pack.
2513 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2517 *) Implement Opaque PRF Input TLS extension as specified in
2518 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2519 official specification yet and no extension type assignment by
2520 IANA exists, this extension (for now) will have to be explicitly
2521 enabled when building OpenSSL by providing the extension number
2522 to use. For example, specify an option
2524 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2526 to the "config" or "Configure" script to enable the extension,
2527 assuming extension number 0x9527 (which is a completely arbitrary
2528 and unofficial assignment based on the MD5 hash of the Internet
2529 Draft). Note that by doing so, you potentially lose
2530 interoperability with other TLS implementations since these might
2531 be using the same extension number for other purposes.
2533 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2534 opaque PRF input value to use in the handshake. This will create
2535 an interal copy of the length-'len' string at 'src', and will
2536 return non-zero for success.
2538 To get more control and flexibility, provide a callback function
2541 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2542 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2546 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2549 Callback function 'cb' will be called in handshakes, and is
2550 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2551 Argument 'arg' is for application purposes (the value as given to
2552 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2553 be provided to the callback function). The callback function
2554 has to return non-zero to report success: usually 1 to use opaque
2555 PRF input just if possible, or 2 to enforce use of the opaque PRF
2556 input. In the latter case, the library will abort the handshake
2557 if opaque PRF input is not successfully negotiated.
2559 Arguments 'peerinput' and 'len' given to the callback function
2560 will always be NULL and 0 in the case of a client. A server will
2561 see the client's opaque PRF input through these variables if
2562 available (NULL and 0 otherwise). Note that if the server
2563 provides an opaque PRF input, the length must be the same as the
2564 length of the client's opaque PRF input.
2566 Note that the callback function will only be called when creating
2567 a new session (session resumption can resume whatever was
2568 previously negotiated), and will not be called in SSL 2.0
2569 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2570 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2571 for applications that need to enforce opaque PRF input.
2575 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2578 [Victor B. Wagner <vitus@cryptocom.ru>]
2580 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2581 RFC4507bis. The encrypted ticket format is an encrypted encoded
2582 SSL_SESSION structure, that way new session features are automatically
2585 If a client application caches session in an SSL_SESSION structure
2586 support is transparent because tickets are now stored in the encoded
2589 The SSL_CTX structure automatically generates keys for ticket
2590 protection in servers so again support should be possible
2591 with no application modification.
2593 If a client or server wishes to disable RFC4507 support then the option
2594 SSL_OP_NO_TICKET can be set.
2596 Add a TLS extension debugging callback to allow the contents of any client
2597 or server extensions to be examined.
2599 This work was sponsored by Google.
2602 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2603 OpenSSL should now compile cleanly on gcc 4.2
2604 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2606 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2607 support including streaming MAC support: this is required for GOST
2608 ciphersuite support.
2609 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2611 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2612 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2613 to output in BER and PEM format.
2616 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2617 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2618 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2619 ENGINE support for HMAC keys which are unextractable. New -mac and
2620 -macopt options to dgst utility.
2623 *) New option -sigopt to dgst utility. Update dgst to use
2624 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2625 alternative signing paramaters such as X9.31 or PSS in the dgst
2629 *) Change ssl_cipher_apply_rule(), the internal function that does
2630 the work each time a ciphersuite string requests enabling
2631 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2632 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2633 the order of disabled ciphersuites such that those ciphersuites
2634 that most recently went from enabled to disabled not only stay
2635 in order with respect to each other, but also have higher priority
2636 than other disabled ciphersuites the next time ciphersuites are
2639 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2640 the same ciphersuites as with "HIGH" alone, but in a specific
2641 order where the PSK ciphersuites come first (since they are the
2642 most recently disabled ciphersuites when "HIGH" is parsed).
2644 Also, change ssl_create_cipher_list() (using this new
2645 funcionality) such that between otherwise identical
2646 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2650 *) Change ssl_create_cipher_list() so that it automatically
2651 arranges the ciphersuites in reasonable order before starting
2652 to process the rule string. Thus, the definition for "DEFAULT"
2653 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2654 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2655 This makes it much easier to arrive at a reasonable default order
2656 in applications for which anonymous ciphers are OK (meaning
2657 that you can't actually use DEFAULT).
2658 [Bodo Moeller; suggested by Victor Duchovni]
2660 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2661 processing) into multiple integers instead of setting
2662 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2663 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2664 (These masks as well as the individual bit definitions are hidden
2665 away into the non-exported interface ssl/ssl_locl.h, so this
2666 change to the definition of the SSL_CIPHER structure shouldn't
2667 affect applications.) This give us more bits for each of these
2668 categories, so there is no longer a need to coagulate AES128 and
2669 AES256 into a single algorithm bit, and to coagulate Camellia128
2670 and Camellia256 into a single algorithm bit, which has led to all
2673 Thus, among other things, the kludge introduced in 0.9.7m and
2674 0.9.8e for masking out AES256 independently of AES128 or masking
2675 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2677 With the change, we also introduce new ciphersuite aliases that
2678 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2682 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2683 Use the leftmost N bytes of the signature input if the input is
2684 larger than the prime q (with N being the size in bytes of q).
2687 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2688 it yet and it is largely untested.
2691 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2694 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2695 some compilers (gcc 4.2 and later) reject their use. Safestack is
2696 reimplemented. Update ASN1 to avoid use of legacy functions.
2699 *) Win32/64 targets are linked with Winsock2.
2702 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2703 to external functions. This can be used to increase CRL handling
2704 efficiency especially when CRLs are very large by (for example) storing
2705 the CRL revoked certificates in a database.
2708 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2709 new CRLs added to a directory can be used. New command line option
2710 -verify_return_error to s_client and s_server. This causes real errors
2711 to be returned by the verify callback instead of carrying on no matter
2712 what. This reflects the way a "real world" verify callback would behave.
2715 *) GOST engine, supporting several GOST algorithms and public key formats.
2716 Kindly donated by Cryptocom.
2719 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2720 partitioned by DP are handled but no indirect CRL or reason partitioning
2721 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2722 selected via a scoring technique which handles IDP and AKID in CRLs.
2725 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2726 will ultimately be used for all verify operations: this will remove the
2727 X509_STORE dependency on certificate verification and allow alternative
2728 lookup methods. X509_STORE based implementations of these two callbacks.
2731 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2732 Modify get_crl() to find a valid (unexpired) CRL if possible.
2735 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2736 this would be called X509_CRL_cmp() but that name is already used by
2737 a function that just compares CRL issuer names. Cache several CRL
2738 extensions in X509_CRL structure and cache CRLDP in X509.
2741 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2742 this maps equivalent X509_NAME structures into a consistent structure.
2743 Name comparison can then be performed rapidly using memcmp().
2746 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2750 *) Allow digests to supply their own micalg string for S/MIME type using
2751 the ctrl EVP_MD_CTRL_MICALG.
2754 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2755 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2756 ctrl. It can then customise the structure before and/or after signing
2760 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2761 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2762 to free up any added signature OIDs.
2765 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2766 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2767 digest and cipher tables. New options added to openssl utility:
2768 list-message-digest-algorithms and list-cipher-algorithms.
2771 *) Change the array representation of binary polynomials: the list
2772 of degrees of non-zero coefficients is now terminated with -1.
2773 Previously it was terminated with 0, which was also part of the
2774 value; thus, the array representation was not applicable to
2775 polynomials where t^0 has coefficient zero. This change makes
2776 the array representation useful in a more general context.
2779 *) Various modifications and fixes to SSL/TLS cipher string
2780 handling. For ECC, the code now distinguishes between fixed ECDH
2781 with RSA certificates on the one hand and with ECDSA certificates
2782 on the other hand, since these are separate ciphersuites. The
2783 unused code for Fortezza ciphersuites has been removed.
2785 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2786 (not "ECDHE"). For consistency with the code for DH
2787 certificates, use of ECDH certificates is now considered ECDH
2788 authentication, not RSA or ECDSA authentication (the latter is
2789 merely the CA's signing algorithm and not actively used in the
2792 The temporary ciphersuite alias "ECCdraft" is no longer
2793 available, and ECC ciphersuites are no longer excluded from "ALL"
2794 and "DEFAULT". The following aliases now exist for RFC 4492
2795 ciphersuites, most of these by analogy with the DH case:
2797 kECDHr - ECDH cert, signed with RSA
2798 kECDHe - ECDH cert, signed with ECDSA
2799 kECDH - ECDH cert (signed with either RSA or ECDSA)
2800 kEECDH - ephemeral ECDH
2801 ECDH - ECDH cert or ephemeral ECDH
2807 AECDH - anonymous ECDH
2808 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2812 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2813 Use correct micalg parameters depending on digest(s) in signed message.
2816 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2817 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2820 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2821 an engine to register a method. Add ENGINE lookups for methods and
2822 functional reference processing.
2825 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2826 EVP_{Sign,Verify}* which allow an application to customise the signature
2830 *) New -resign option to smime utility. This adds one or more signers
2831 to an existing PKCS#7 signedData structure. Also -md option to use an
2832 alternative message digest algorithm for signing.
2835 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2836 create PKCS7 structures containing multiple signers. Update smime
2837 application to support multiple signers.
2840 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2844 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2845 Reorganize PBE internals to lookup from a static table using NIDs,
2846 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2847 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2848 PRF which will be automatically used with PBES2.
2851 *) Replace the algorithm specific calls to generate keys in "req" with the
2855 *) Update PKCS#7 enveloped data routines to use new API. This is now
2856 supported by any public key method supporting the encrypt operation. A
2857 ctrl is added to allow the public key algorithm to examine or modify
2858 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2862 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2863 a default digest type to use. In most cases this will be SHA1 but some
2864 algorithms (such as GOST) need to specify an alternative digest. The
2865 return value indicates how strong the prefernce is 1 means optional and
2866 2 is mandatory (that is it is the only supported type). Modify
2867 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2868 use the default md. Update openssl utilities to use the default digest
2869 type for signing if it is not explicitly indicated.
2872 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2873 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2874 signing method from the key type. This effectively removes the link
2875 between digests and public key types.
2878 *) Add an OID cross reference table and utility functions. Its purpose is to
2879 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2880 rsaEncryption. This will allow some of the algorithm specific hackery
2881 needed to use the correct OID to be removed.
2884 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2885 structures for PKCS7_sign(). They are now set up by the relevant public
2889 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2892 *) Add support for key derivation (agreement) in the API, DH method and
2896 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2897 public and private key formats. As a side effect these add additional
2898 command line functionality not previously available: DSA signatures can be
2899 generated and verified using pkeyutl and DH key support and generation in
2904 [Oliver Tappe <zooey@hirschkaefer.de>]
2906 *) New make target "install_html_docs" installs HTML renditions of the
2908 [Oliver Tappe <zooey@hirschkaefer.de>]
2910 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
2911 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2912 support key and parameter generation and add initial key generation
2913 functionality for RSA.
2916 *) Add functions for main EVP_PKEY_method operations. The undocumented
2917 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2918 EVP_PKEY_{encrypt,decrypt}_old.
2921 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2922 key API, doesn't do much yet.
2925 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2926 public key algorithms. New option to openssl utility:
2927 "list-public-key-algorithms" to print out info.
2930 *) Implement the Supported Elliptic Curves Extension for
2931 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2934 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2935 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2938 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2939 utilities such as rsa, dsa, dsaparam etc except they process any key
2943 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2944 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2945 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2949 *) Initial support for pluggable public key ASN1.
2950 De-spaghettify the public key ASN1 handling. Move public and private
2951 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2952 algorithm specific handling to a single module within the relevant
2953 algorithm directory. Add functions to allow (near) opaque processing
2954 of public and private key structures.
2957 *) Implement the Supported Point Formats Extension for
2958 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2961 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2962 for the psk identity [hint] and the psk callback functions to the
2963 SSL_SESSION, SSL and SSL_CTX structure.
2966 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2970 SSL_CTX_use_psk_identity_hint
2971 SSL_get_psk_identity_hint
2972 SSL_get_psk_identity
2973 SSL_use_psk_identity_hint
2975 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2977 *) Add RFC 3161 compliant time stamp request creation, response generation
2978 and response verification functionality.
2979 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2981 *) Add initial support for TLS extensions, specifically for the server_name
2982 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2983 have new members for a host name. The SSL data structure has an
2984 additional member SSL_CTX *initial_ctx so that new sessions can be
2985 stored in that context to allow for session resumption, even after the
2986 SSL has been switched to a new SSL_CTX in reaction to a client's
2987 server_name extension.
2989 New functions (subject to change):
2991 SSL_get_servername()
2992 SSL_get_servername_type()
2995 New CTRL codes and macros (subject to change):
2997 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2998 - SSL_CTX_set_tlsext_servername_callback()
2999 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3000 - SSL_CTX_set_tlsext_servername_arg()
3001 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3003 openssl s_client has a new '-servername ...' option.
3005 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3006 '-key2 ...', '-servername_fatal' (subject to change). This allows
3007 testing the HostName extension for a specific single host name ('-cert'
3008 and '-key' remain fallbacks for handshakes without HostName
3009 negotiation). If the unrecogninzed_name alert has to be sent, this by
3010 default is a warning; it becomes fatal with the '-servername_fatal'
3013 [Peter Sylvester, Remy Allais, Christophe Renou]
3015 *) Whirlpool hash implementation is added.
3018 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3019 bn(64,32). Because of instruction set limitations it doesn't have
3020 any negative impact on performance. This was done mostly in order
3021 to make it possible to share assembler modules, such as bn_mul_mont
3022 implementations, between 32- and 64-bit builds without hassle.
3025 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3026 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3030 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3031 dedicated Montgomery multiplication procedure, is introduced.
3032 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3033 "64-bit" performance on certain 32-bit targets.
3036 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3037 in SSL structures. New SSL ctrl to set maximum send fragment size.
3038 Save memory by seeting the I/O buffer sizes dynamically instead of
3039 using the maximum available value.
3042 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3043 in addition to the text details.
3046 *) Very, very preliminary EXPERIMENTAL support for printing of general
3047 ASN1 structures. This currently produces rather ugly output and doesn't
3048 handle several customised structures at all.
3051 *) Integrated support for PVK file format and some related formats such
3052 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3053 these in the 'rsa' and 'dsa' utilities.
3056 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3059 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3060 place for the (very old) "NETSCAPE" format certificates which are now
3061 handled using new ASN1 code equivalents.
3064 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3065 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3066 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3069 *) Modify CRL distribution points extension code to print out previously
3070 unsupported fields. Enhance extension setting code to allow setting of
3074 *) Add print and set support for Issuing Distribution Point CRL extension.
3077 *) Change 'Configure' script to enable Camellia by default.
3080 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3082 *) When rejecting SSL/TLS records due to an incorrect version number, never
3083 update s->server with a new major version number. As of
3084 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3085 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3086 the previous behavior could result in a read attempt at NULL when
3087 receiving specific incorrect SSL/TLS records once record payload
3088 protection is active. (CVE-2010-0740)
3089 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3091 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3092 could be crashed if the relevant tables were not present (e.g. chrooted).
3093 [Tomas Hoger <thoger@redhat.com>]
3095 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3097 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3098 [Martin Olsson, Neel Mehta]
3100 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3101 accommodate for stack sorting, always a write lock!).
3104 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3105 excessive delays in the RAND_poll(): over a minute. As a workaround
3106 include a time check in the inner Heap32Next loop too.
3109 *) The code that handled flushing of data in SSL/TLS originally used the
3110 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3111 the problem outlined in PR#1949. The fix suggested there however can
3112 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3113 of Apache). So instead simplify the code to flush unconditionally.
3114 This should be fine since flushing with no data to flush is a no op.
3117 *) Handle TLS versions 2.0 and later properly and correctly use the
3118 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3119 off ancient servers have a habit of sticking around for a while...
3122 *) Modify compression code so it frees up structures without using the
3123 ex_data callbacks. This works around a problem where some applications
3124 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3125 restarting) then use compression (e.g. SSL with compression) later.
3126 This results in significant per-connection memory leaks and
3127 has caused some security issues including CVE-2008-1678 and
3131 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3132 change when encrypting or decrypting.
3135 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3136 connect and renegotiate with servers which do not support RI.
3137 Until RI is more widely deployed this option is enabled by default.
3140 *) Add "missing" ssl ctrls to clear options and mode.
3143 *) If client attempts to renegotiate and doesn't support RI respond with
3144 a no_renegotiation alert as required by RFC5746. Some renegotiating
3145 TLS clients will continue a connection gracefully when they receive
3146 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3147 waiting for a server hello which it will never receive. Now we treat a
3148 received no_renegotiation alert as a fatal error. This is because
3149 applications requesting a renegotiation might well expect it to succeed
3150 and would have no code in place to handle the server denying it so the
3151 only safe thing to do is to terminate the connection.
3154 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3155 peer supports secure renegotiation and 0 otherwise. Print out peer
3156 renegotiation support in s_client/s_server.
3159 *) Replace the highly broken and deprecated SPKAC certification method with
3160 the updated NID creation version. This should correctly handle UTF8.
3163 *) Implement RFC5746. Re-enable renegotiation but require the extension
3164 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3165 turns out to be a bad idea. It has been replaced by
3166 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3167 SSL_CTX_set_options(). This is really not recommended unless you
3168 know what you are doing.
3169 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3171 *) Fixes to stateless session resumption handling. Use initial_ctx when
3172 issuing and attempting to decrypt tickets in case it has changed during
3173 servername handling. Use a non-zero length session ID when attempting
3174 stateless session resumption: this makes it possible to determine if
3175 a resumption has occurred immediately after receiving server hello
3176 (several places in OpenSSL subtly assume this) instead of later in
3180 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3181 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3182 fixes for a few places where the return code is not checked
3184 [Julia Lawall <julia@diku.dk>]
3186 *) Add --strict-warnings option to Configure script to include devteam
3187 warnings in other configurations.
3190 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3191 makes it possible to install openssl libraries in locations which
3192 have names other than "lib", for example "/usr/lib64" which some
3194 [Steve Henson, based on patch from Jeremy Utley]
3196 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3197 X690 8.9.12 and can produce some misleading textual output of OIDs.
3198 [Steve Henson, reported by Dan Kaminsky]
3200 *) Delete MD2 from algorithm tables. This follows the recommendation in
3201 several standards that it is not used in new applications due to
3202 several cryptographic weaknesses. For binary compatibility reasons
3203 the MD2 API is still compiled in by default.
3206 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3210 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3211 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3213 [Guenter <lists@gknw.net>]
3215 *) Fix the server certificate chain building code to use X509_verify_cert(),
3216 it used to have an ad-hoc builder which was unable to cope with anything
3217 other than a simple chain.
3218 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3220 *) Don't check self signed certificate signatures in X509_verify_cert()
3221 by default (a flag can override this): it just wastes time without
3222 adding any security. As a useful side effect self signed root CAs
3223 with non-FIPS digests are now usable in FIPS mode.
3226 *) In dtls1_process_out_of_seq_message() the check if the current message
3227 is already buffered was missing. For every new message was memory
3228 allocated, allowing an attacker to perform an denial of service attack
3229 with sending out of seq handshake messages until there is no memory
3230 left. Additionally every future messege was buffered, even if the
3231 sequence number made no sense and would be part of another handshake.
3232 So only messages with sequence numbers less than 10 in advance will be
3233 buffered. (CVE-2009-1378)
3234 [Robin Seggelmann, discovered by Daniel Mentz]
3236 *) Records are buffered if they arrive with a future epoch to be
3237 processed after finishing the corresponding handshake. There is
3238 currently no limitation to this buffer allowing an attacker to perform
3239 a DOS attack with sending records with future epochs until there is no
3240 memory left. This patch adds the pqueue_size() function to detemine
3241 the size of a buffer and limits the record buffer to 100 entries.
3243 [Robin Seggelmann, discovered by Daniel Mentz]
3245 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3246 parent structure is freed. (CVE-2009-1379)
3249 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3250 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3253 [Ilya O. <vrghost@gmail.com>]
3255 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3257 *) Disable renegotiation completely - this fixes a severe security
3258 problem (CVE-2009-3555) at the cost of breaking all
3259 renegotiation. Renegotiation can be re-enabled by setting
3260 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3261 run-time. This is really not recommended unless you know what
3265 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3267 *) Don't set val to NULL when freeing up structures, it is freed up by
3268 underlying code. If sizeof(void *) > sizeof(long) this can result in
3269 zeroing past the valid field. (CVE-2009-0789)
3270 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3272 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3273 checked correctly. This would allow some invalid signed attributes to
3274 appear to verify correctly. (CVE-2009-0591)
3275 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3277 *) Reject UniversalString and BMPString types with invalid lengths. This
3278 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3279 a legal length. (CVE-2009-0590)
3282 *) Set S/MIME signing as the default purpose rather than setting it
3283 unconditionally. This allows applications to override it at the store
3287 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3288 to handle some structures.
3291 *) Improve efficiency of mem_gets: don't search whole buffer each time
3293 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3295 *) New -hex option for openssl rand.
3298 *) Print out UTF8String and NumericString when parsing ASN1.
3301 *) Support NumericString type for name components.
3304 *) Allow CC in the environment to override the automatically chosen
3305 compiler. Note that nothing is done to ensure flags work with the
3309 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3311 *) Properly check EVP_VerifyFinal() and similar return values
3313 [Ben Laurie, Bodo Moeller, Google Security Team]
3315 *) Enable TLS extensions by default.
3318 *) Allow the CHIL engine to be loaded, whether the application is
3319 multithreaded or not. (This does not release the developer from the
3320 obligation to set up the dynamic locking callbacks.)
3321 [Sander Temme <sander@temme.net>]
3323 *) Use correct exit code if there is an error in dgst command.
3324 [Steve Henson; problem pointed out by Roland Dirlewanger]
3326 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3327 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3330 *) Add experimental JPAKE support, including demo authentication in
3331 s_client and s_server.
3334 *) Set the comparison function in v3_addr_canonize().
3335 [Rob Austein <sra@hactrn.net>]
3337 *) Add support for XMPP STARTTLS in s_client.
3338 [Philip Paeps <philip@freebsd.org>]
3340 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3341 to ensure that even with this option, only ciphersuites in the
3342 server's preference list will be accepted. (Note that the option
3343 applies only when resuming a session, so the earlier behavior was
3344 just about the algorithm choice for symmetric cryptography.)
3347 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3349 *) Fix NULL pointer dereference if a DTLS server received
3350 ChangeCipherSpec as first record (CVE-2009-1386).
3353 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
3354 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3357 *) The fix in 0.9.8c that supposedly got rid of unsafe
3358 double-checked locking was incomplete for RSA blinding,
3359 addressing just one layer of what turns out to have been
3360 doubly unsafe triple-checked locking.
3362 So now fix this for real by retiring the MONT_HELPER macro
3363 in crypto/rsa/rsa_eay.c.
3365 [Bodo Moeller; problem pointed out by Marius Schilder]
3367 *) Various precautionary measures:
3369 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3371 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3372 (NB: This would require knowledge of the secret session ticket key
3373 to exploit, in which case you'd be SOL either way.)
3375 - Change bn_nist.c so that it will properly handle input BIGNUMs
3376 outside the expected range.
3378 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3381 [Neel Mehta, Bodo Moeller]
3383 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3384 the load fails. Useful for distros.
3385 [Ben Laurie and the FreeBSD team]
3387 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3390 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3393 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3395 This work was sponsored by Logica.
3398 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3399 keystores. Support for SSL/TLS client authentication too.
3400 Not compiled unless enable-capieng specified to Configure.
3402 This work was sponsored by Logica.
3405 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3406 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3407 attribute creation routines such as certifcate requests and PKCS#12
3411 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3413 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3414 handshake which could lead to a cilent crash as found using the
3415 Codenomicon TLS test suite (CVE-2008-1672)
3416 [Steve Henson, Mark Cox]
3418 *) Fix double free in TLS server name extensions which could lead to
3419 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3422 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3424 Clear the error queue to ensure that error entries left from
3425 older function calls do not interfere with the correct operation.
3426 [Lutz Jaenicke, Erik de Castro Lopo]
3428 *) Remove root CA certificates of commercial CAs:
3430 The OpenSSL project does not recommend any specific CA and does not
3431 have any policy with respect to including or excluding any CA.
3432 Therefore it does not make any sense to ship an arbitrary selection
3433 of root CA certificates with the OpenSSL software.
3436 *) RSA OAEP patches to fix two separate invalid memory reads.
3437 The first one involves inputs when 'lzero' is greater than
3438 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3439 before the beginning of from). The second one involves inputs where
3440 the 'db' section contains nothing but zeroes (there is a one-byte
3441 invalid read after the end of 'db').
3442 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3444 *) Partial backport from 0.9.9-dev:
3446 Introduce bn_mul_mont (dedicated Montgomery multiplication
3447 procedure) as a candidate for BIGNUM assembler implementation.
3448 While 0.9.9-dev uses assembler for various architectures, only
3449 x86_64 is available by default here in the 0.9.8 branch, and
3450 32-bit x86 is available through a compile-time setting.
3452 To try the 32-bit x86 assembler implementation, use Configure
3453 option "enable-montasm" (which exists only for this backport).
3455 As "enable-montasm" for 32-bit x86 disclaims code stability
3456 anyway, in this constellation we activate additional code
3457 backported from 0.9.9-dev for further performance improvements,
3458 namely BN_from_montgomery_word. (To enable this otherwise,
3459 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3461 [Andy Polyakov (backport partially by Bodo Moeller)]
3463 *) Add TLS session ticket callback. This allows an application to set
3464 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3465 values. This is useful for key rollover for example where several key
3466 sets may exist with different names.
3469 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3470 This was broken until now in 0.9.8 releases, such that the only way
3471 a registered ENGINE could be used (assuming it initialises
3472 successfully on the host) was to explicitly set it as the default
3473 for the relevant algorithms. This is in contradiction with 0.9.7
3474 behaviour and the documentation. With this fix, when an ENGINE is
3475 registered into a given algorithm's table of implementations, the
3476 'uptodate' flag is reset so that auto-discovery will be used next
3477 time a new context for that algorithm attempts to select an
3479 [Ian Lister (tweaked by Geoff Thorpe)]
3481 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3482 implemention in the following ways:
3484 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3487 Lack of BER streaming support means one pass streaming processing is
3488 only supported if data is detached: setting the streaming flag is
3489 ignored for embedded content.
3491 CMS support is disabled by default and must be explicitly enabled
3492 with the enable-cms configuration option.
3495 *) Update the GMP engine glue to do direct copies between BIGNUM and
3496 mpz_t when openssl and GMP use the same limb size. Otherwise the
3497 existing "conversion via a text string export" trick is still used.
3498 [Paul Sheer <paulsheer@gmail.com>]
3500 *) Zlib compression BIO. This is a filter BIO which compressed and
3501 uncompresses any data passed through it.
3504 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3505 RFC3394 compatible AES key wrapping.
3508 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3509 sets string data without copying. X509_ALGOR_set0() and
3510 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3511 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3512 from an X509_ATTRIBUTE structure optionally checking it occurs only
3513 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3517 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3518 to get the expected BN_FLG_CONSTTIME behavior.
3519 [Bodo Moeller (Google)]
3523 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3524 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3525 - added some more tests to do_tests.pl
3526 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3527 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3528 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3529 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3530 - various changes to netware.pl to enable gcc-cross builds on Win32
3532 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3533 - various changes to fix missing prototype warnings
3534 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3535 - added AES, WHIRLPOOL and CPUID assembler code to build files
3536 - added missing AES assembler make rules to mk1mf.pl
3537 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3538 [Guenter Knauf <eflash@gmx.net>]
3540 *) Implement certificate status request TLS extension defined in RFC3546.
3541 A client can set the appropriate parameters and receive the encoded
3542 OCSP response via a callback. A server can query the supplied parameters
3543 and set the encoded OCSP response in the callback. Add simplified examples
3544 to s_client and s_server.
3547 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3549 *) Fix various bugs:
3550 + Binary incompatibility of ssl_ctx_st structure
3551 + DTLS interoperation with non-compliant servers
3552 + Don't call get_session_cb() without proposed session
3553 + Fix ia64 assembler code
3554 [Andy Polyakov, Steve Henson]
3556 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3558 *) DTLS Handshake overhaul. There were longstanding issues with
3559 OpenSSL DTLS implementation, which were making it impossible for
3560 RFC 4347 compliant client to communicate with OpenSSL server.
3561 Unfortunately just fixing these incompatibilities would "cut off"
3562 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3563 server keeps tolerating non RFC compliant syntax. The opposite is
3564 not true, 0.9.8f client can not communicate with earlier server.
3565 This update even addresses CVE-2007-4995.
3568 *) Changes to avoid need for function casts in OpenSSL: some compilers
3569 (gcc 4.2 and later) reject their use.
3570 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3573 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3574 RFC4507bis. The encrypted ticket format is an encrypted encoded
3575 SSL_SESSION structure, that way new session features are automatically
3578 If a client application caches session in an SSL_SESSION structure
3579 support is transparent because tickets are now stored in the encoded
3582 The SSL_CTX structure automatically generates keys for ticket
3583 protection in servers so again support should be possible
3584 with no application modification.
3586 If a client or server wishes to disable RFC4507 support then the option
3587 SSL_OP_NO_TICKET can be set.
3589 Add a TLS extension debugging callback to allow the contents of any client
3590 or server extensions to be examined.
3592 This work was sponsored by Google.
3595 *) Add initial support for TLS extensions, specifically for the server_name
3596 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3597 have new members for a host name. The SSL data structure has an
3598 additional member SSL_CTX *initial_ctx so that new sessions can be
3599 stored in that context to allow for session resumption, even after the
3600 SSL has been switched to a new SSL_CTX in reaction to a client's
3601 server_name extension.
3603 New functions (subject to change):
3605 SSL_get_servername()
3606 SSL_get_servername_type()
3609 New CTRL codes and macros (subject to change):
3611 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3612 - SSL_CTX_set_tlsext_servername_callback()
3613 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3614 - SSL_CTX_set_tlsext_servername_arg()
3615 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3617 openssl s_client has a new '-servername ...' option.
3619 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3620 '-key2 ...', '-servername_fatal' (subject to change). This allows
3621 testing the HostName extension for a specific single host name ('-cert'
3622 and '-key' remain fallbacks for handshakes without HostName
3623 negotiation). If the unrecogninzed_name alert has to be sent, this by
3624 default is a warning; it becomes fatal with the '-servername_fatal'
3627 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3629 *) Add AES and SSE2 assembly language support to VC++ build.
3632 *) Mitigate attack on final subtraction in Montgomery reduction.
3635 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3636 (which previously caused an internal error).
3639 *) Squeeze another 10% out of IGE mode when in != out.
3642 *) AES IGE mode speedup.
3643 [Dean Gaudet (Google)]
3645 *) Add the Korean symmetric 128-bit cipher SEED (see
3646 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3647 add SEED ciphersuites from RFC 4162:
3649 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3650 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3651 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3652 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3654 To minimize changes between patchlevels in the OpenSSL 0.9.8
3655 series, SEED remains excluded from compilation unless OpenSSL
3656 is configured with 'enable-seed'.
3657 [KISA, Bodo Moeller]
3659 *) Mitigate branch prediction attacks, which can be practical if a
3660 single processor is shared, allowing a spy process to extract
3661 information. For detailed background information, see
3662 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3663 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3664 and Necessary Software Countermeasures"). The core of the change
3665 are new versions BN_div_no_branch() and
3666 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3667 respectively, which are slower, but avoid the security-relevant
3668 conditional branches. These are automatically called by BN_div()
3669 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3670 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3671 remove a conditional branch.
3673 BN_FLG_CONSTTIME is the new name for the previous
3674 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3675 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3676 in the exponent causes BN_mod_exp_mont() to use the alternative
3677 implementation in BN_mod_exp_mont_consttime().) The old name
3678 remains as a deprecated alias.
3680 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3681 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3682 constant-time implementations for more than just exponentiation.
3683 Here too the old name is kept as a deprecated alias.
3685 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3686 the BN_BLINDING structure gets an independent copy of the
3687 modulus. This means that the previous "BIGNUM *m" argument to
3688 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3689 essentially becomes "const BIGNUM *m", although we can't actually
3690 change this in the header file before 0.9.9. It allows
3691 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3692 enable BN_FLG_CONSTTIME.
3694 [Matthew D Wood (Intel Corp)]
3696 *) In the SSL/TLS server implementation, be strict about session ID
3697 context matching (which matters if an application uses a single
3698 external cache for different purposes). Previously,
3699 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3700 set. This did ensure strict client verification, but meant that,
3701 with applications using a single external cache for quite
3702 different requirements, clients could circumvent ciphersuite
3703 restrictions for a given session ID context by starting a session
3704 in a different context.
3707 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3708 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3709 authentication-only ciphersuites.
3712 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3713 not complete and could lead to a possible single byte overflow
3714 (CVE-2007-5135) [Ben Laurie]
3716 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3718 *) Since AES128 and AES256 (and similarly Camellia128 and
3719 Camellia256) share a single mask bit in the logic of
3720 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3721 kludge to work properly if AES128 is available and AES256 isn't
3722 (or if Camellia128 is available and Camellia256 isn't).
3725 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3726 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3727 When a point or a seed is encoded in a BIT STRING, we need to
3728 prevent the removal of trailing zero bits to get the proper DER
3729 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3730 of a NamedBitList, for which trailing 0 bits need to be removed.)
3733 *) Have SSL/TLS server implementation tolerate "mismatched" record
3734 protocol version while receiving ClientHello even if the
3735 ClientHello is fragmented. (The server can't insist on the
3736 particular protocol version it has chosen before the ServerHello
3737 message has informed the client about his choice.)
3740 *) Add RFC 3779 support.
3741 [Rob Austein for ARIN, Ben Laurie]
3743 *) Load error codes if they are not already present instead of using a
3744 static variable. This allows them to be cleanly unloaded and reloaded.
3745 Improve header file function name parsing.
3748 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3749 or CAPABILITY handshake as required by RFCs.
3752 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3754 *) Introduce limits to prevent malicious keys being able to
3755 cause a denial of service. (CVE-2006-2940)
3756 [Steve Henson, Bodo Moeller]
3758 *) Fix ASN.1 parsing of certain invalid structures that can result
3759 in a denial of service. (CVE-2006-2937) [Steve Henson]
3761 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3762 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3764 *) Fix SSL client code which could crash if connecting to a
3765 malicious SSLv2 server. (CVE-2006-4343)
3766 [Tavis Ormandy and Will Drewry, Google Security Team]
3768 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3769 match only those. Before that, "AES256-SHA" would be interpreted
3770 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3771 the same strength classification in 0.9.7h) as we currently only
3772 have a single AES bit in the ciphersuite description bitmap.
3773 That change, however, also applied to ciphersuite strings such as
3774 "RC4-MD5" that intentionally matched multiple ciphersuites --
3775 namely, SSL 2.0 ciphersuites in addition to the more common ones
3776 from SSL 3.0/TLS 1.0.
3778 So we change the selection algorithm again: Naming an explicit
3779 ciphersuite selects this one ciphersuite, and any other similar
3780 ciphersuite (same bitmap) from *other* protocol versions.
3781 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3782 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3784 Since SSL 2.0 does not have any ciphersuites for which the
3785 128/256 bit distinction would be relevant, this works for now.
3786 The proper fix will be to use different bits for AES128 and
3787 AES256, which would have avoided the problems from the beginning;
3788 however, bits are scarce, so we can only do this in a new release
3789 (not just a patchlevel) when we can change the SSL_CIPHER
3790 definition to split the single 'unsigned long mask' bitmap into
3791 multiple values to extend the available space.
3795 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3797 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3798 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3800 *) Add AES IGE and biIGE modes.
3803 *) Change the Unix randomness entropy gathering to use poll() when
3804 possible instead of select(), since the latter has some
3805 undesirable limitations.
3806 [Darryl Miles via Richard Levitte and Bodo Moeller]
3808 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3809 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3810 cannot be implicitly activated as part of, e.g., the "AES" alias.
3811 However, please upgrade to OpenSSL 0.9.9[-dev] for
3812 non-experimental use of the ECC ciphersuites to get TLS extension
3813 support, which is required for curve and point format negotiation
3814 to avoid potential handshake problems.
3817 *) Disable rogue ciphersuites:
3819 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3820 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3821 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3823 The latter two were purportedly from
3824 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3827 Also deactivate the remaining ciphersuites from
3828 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3829 unofficial, and the ID has long expired.
3832 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3833 dual-core machines) and other potential thread-safety issues.
3836 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3837 versions), which is now available for royalty-free use
3838 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3839 Also, add Camellia TLS ciphersuites from RFC 4132.
3841 To minimize changes between patchlevels in the OpenSSL 0.9.8
3842 series, Camellia remains excluded from compilation unless OpenSSL
3843 is configured with 'enable-camellia'.
3846 *) Disable the padding bug check when compression is in use. The padding
3847 bug check assumes the first packet is of even length, this is not
3848 necessarily true if compresssion is enabled and can result in false
3849 positives causing handshake failure. The actual bug test is ancient
3850 code so it is hoped that implementations will either have fixed it by
3851 now or any which still have the bug do not support compression.
3854 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3856 *) When applying a cipher rule check to see if string match is an explicit
3857 cipher suite and only match that one cipher suite if it is.
3860 *) Link in manifests for VC++ if needed.
3861 [Austin Ziegler <halostatue@gmail.com>]
3863 *) Update support for ECC-based TLS ciphersuites according to
3864 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3865 TLS extensions, which are supported starting with the 0.9.9
3866 branch, not in the OpenSSL 0.9.8 branch).
3869 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3870 opaque EVP_CIPHER_CTX handling.
3873 *) Fixes and enhancements to zlib compression code. We now only use
3874 "zlib1.dll" and use the default __cdecl calling convention on Win32
3875 to conform with the standards mentioned here:
3876 http://www.zlib.net/DLL_FAQ.txt
3877 Static zlib linking now works on Windows and the new --with-zlib-include
3878 --with-zlib-lib options to Configure can be used to supply the location
3879 of the headers and library. Gracefully handle case where zlib library
3883 *) Several fixes and enhancements to the OID generation code. The old code
3884 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3885 handle numbers larger than ULONG_MAX, truncated printing and had a
3886 non standard OBJ_obj2txt() behaviour.
3889 *) Add support for building of engines under engine/ as shared libraries
3890 under VC++ build system.
3893 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3894 Hopefully, we will not see any false combination of paths any more.
3897 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3899 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3900 (part of SSL_OP_ALL). This option used to disable the
3901 countermeasure against man-in-the-middle protocol-version
3902 rollback in the SSL 2.0 server implementation, which is a bad
3903 idea. (CVE-2005-2969)
3905 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3906 for Information Security, National Institute of Advanced Industrial
3907 Science and Technology [AIST], Japan)]
3909 *) Add two function to clear and return the verify parameter flags.
3912 *) Keep cipherlists sorted in the source instead of sorting them at
3913 runtime, thus removing the need for a lock.
3916 *) Avoid some small subgroup attacks in Diffie-Hellman.
3917 [Nick Mathewson and Ben Laurie]
3919 *) Add functions for well-known primes.
3922 *) Extended Windows CE support.
3923 [Satoshi Nakamura and Andy Polyakov]
3925 *) Initialize SSL_METHOD structures at compile time instead of during
3926 runtime, thus removing the need for a lock.
3929 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3930 attempting to decrypt each encrypted key in turn. Add support to
3934 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3936 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3939 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3942 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3943 key into the same file any more.
3946 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3949 *) Add -utf8 command line and config file option to 'ca'.
3950 [Stefan <stf@udoma.org]
3952 *) Removed the macro des_crypt(), as it seems to conflict with some
3953 libraries. Use DES_crypt().
3956 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3957 involves renaming the source and generated shared-libs for
3958 both. The engines will accept the corrected or legacy ids
3959 ('ncipher' and '4758_cca' respectively) when binding. NB,
3960 this only applies when building 'shared'.
3961 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3963 *) Add attribute functions to EVP_PKEY structure. Modify
3964 PKCS12_create() to recognize a CSP name attribute and
3965 use it. Make -CSP option work again in pkcs12 utility.
3968 *) Add new functionality to the bn blinding code:
3969 - automatic re-creation of the BN_BLINDING parameters after
3970 a fixed number of uses (currently 32)
3971 - add new function for parameter creation
3972 - introduce flags to control the update behaviour of the
3973 BN_BLINDING parameters
3974 - hide BN_BLINDING structure
3975 Add a second BN_BLINDING slot to the RSA structure to improve
3976 performance when a single RSA object is shared among several
3980 *) Add support for DTLS.
3981 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3983 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3984 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3987 *) Remove buggy and incompletet DH cert support from
3988 ssl/ssl_rsa.c and ssl/s3_both.c
3991 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3992 the apps/openssl applications.
3995 *) Compile clean with "-Wall -Wmissing-prototypes
3996 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3997 DEBUG_SAFESTACK must also be set.
4000 *) Change ./Configure so that certain algorithms can be disabled by default.
4001 The new counterpiece to "no-xxx" is "enable-xxx".
4003 The patented RC5 and MDC2 algorithms will now be disabled unless
4004 "enable-rc5" and "enable-mdc2", respectively, are specified.
4006 (IDEA remains enabled despite being patented. This is because IDEA
4007 is frequently required for interoperability, and there is no license
4008 fee for non-commercial use. As before, "no-idea" can be used to
4009 avoid this algorithm.)
4013 *) Add processing of proxy certificates (see RFC 3820). This work was
4014 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
4015 EGEE (Enabling Grids for E-science in Europe).
4018 *) RC4 performance overhaul on modern architectures/implementations, such
4019 as Intel P4, IA-64 and AMD64.
4022 *) New utility extract-section.pl. This can be used specify an alternative
4023 section number in a pod file instead of having to treat each file as
4024 a separate case in Makefile. This can be done by adding two lines to the
4027 =for comment openssl_section:XXX
4029 The blank line is mandatory.
4033 *) New arguments -certform, -keyform and -pass for s_client and s_server
4034 to allow alternative format key and certificate files and passphrase
4038 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4039 update associated structures and add various utility functions.
4041 Add new policy related verify parameters, include policy checking in
4042 standard verify code. Enhance 'smime' application with extra parameters
4043 to support policy checking and print out.
4046 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4047 Nehemiah processors. These extensions support AES encryption in hardware
4048 as well as RNG (though RNG support is currently disabled).
4049 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4051 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4054 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4055 [Andy Polyakov and a number of other people]
4057 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4058 implementation contributed by IBM.
4059 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4061 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4062 exponent rather than 'unsigned long'. There is a corresponding change to
4063 the new 'rsa_keygen' element of the RSA_METHOD structure.
4064 [Jelte Jansen, Geoff Thorpe]
4066 *) Functionality for creating the initial serial number file is now
4067 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4069 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4070 number file to 1, which is bound to cause problems. To avoid
4071 the problems while respecting compatibility between different 0.9.7
4072 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4073 CA.pl for serial number initialization. With the new release 0.9.8,
4074 we can fix the problem directly in the 'ca' utility.)
4077 *) Reduced header interdepencies by declaring more opaque objects in
4078 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4079 give fewer recursive includes, which could break lazy source code - so
4080 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4081 developers should define this symbol when building and using openssl to
4082 ensure they track the recommended behaviour, interfaces, [etc], but
4083 backwards-compatible behaviour prevails when this isn't defined.
4086 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4089 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4090 This will generate a random key of the appropriate length based on the
4091 cipher context. The EVP_CIPHER can provide its own random key generation
4092 routine to support keys of a specific form. This is used in the des and
4093 3des routines to generate a key of the correct parity. Update S/MIME
4094 code to use new functions and hence generate correct parity DES keys.
4095 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4096 valid (weak or incorrect parity).
4099 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4100 as looking them up. This is useful when the verified structure may contain
4101 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4102 present unless the new PKCS7_NO_CRL flag is asserted.
4105 *) Extend ASN1 oid configuration module. It now additionally accepts the
4108 shortName = some long name, 1.2.3.4
4111 *) Reimplemented the BN_CTX implementation. There is now no more static
4112 limitation on the number of variables it can handle nor the depth of the
4113 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4114 information can now expand as required, and rather than having a single
4115 static array of bignums, BN_CTX now uses a linked-list of such arrays
4116 allowing it to expand on demand whilst maintaining the usefulness of
4117 BN_CTX's "bundling".
4120 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4121 to allow all RSA operations to function using a single BN_CTX.
4124 *) Preliminary support for certificate policy evaluation and checking. This
4125 is initially intended to pass the tests outlined in "Conformance Testing
4126 of Relying Party Client Certificate Path Processing Logic" v1.07.
4129 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4130 remained unused and not that useful. A variety of other little bignum
4131 tweaks and fixes have also been made continuing on from the audit (see
4135 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4136 associated ASN1, EVP and SSL functions and old ASN1 macros.
4139 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4140 and this should never fail. So the return value from the use of
4141 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4142 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4145 *) BN_CTX_get() should return zero-valued bignums, providing the same
4146 initialised value as BN_new().
4147 [Geoff Thorpe, suggested by Ulf Möller]
4149 *) Support for inhibitAnyPolicy certificate extension.
4152 *) An audit of the BIGNUM code is underway, for which debugging code is
4153 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4154 is considered valid when processing BIGNUMs, and causes execution to
4155 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4156 further steps are taken to deliberately pollute unused data in BIGNUM
4157 structures to try and expose faulty code further on. For now, openssl will
4158 (in its default mode of operation) continue to tolerate the inconsistent
4159 forms that it has tolerated in the past, but authors and packagers should
4160 consider trying openssl and their own applications when compiled with
4161 these debugging symbols defined. It will help highlight potential bugs in
4162 their own code, and will improve the test coverage for OpenSSL itself. At
4163 some point, these tighter rules will become openssl's default to improve
4164 maintainability, though the assert()s and other overheads will remain only
4165 in debugging configurations. See bn.h for more details.
4166 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4168 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4169 that can only be obtained through BN_CTX_new() (which implicitly
4170 initialises it). The presence of this function only made it possible
4171 to overwrite an existing structure (and cause memory leaks).
4174 *) Because of the callback-based approach for implementing LHASH as a
4175 template type, lh_insert() adds opaque objects to hash-tables and
4176 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4177 to clean up those corresponding objects before destroying the hash table
4178 (and losing the object pointers). So some over-zealous constifications in
4179 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4180 objects as "const" and the lh_doall[_arg] callback wrappers are not
4181 prototyped to have "const" restrictions on the object pointers they are
4182 given (and so aren't required to cast them away any more).
4185 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4186 (speed) prefers to use its own implementation. The two implementations
4187 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4188 its object type properly exposed (MS_TM) instead of casting to/from "char
4189 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4190 aren't necessarily the greatest nomenclatures - but this is what was used
4191 internally to the implementation so I've used that for now.
4194 *) Ensure that deprecated functions do not get compiled when
4195 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4196 the self-tests were still using deprecated key-generation functions so
4197 these have been updated also.
4200 *) Reorganise PKCS#7 code to separate the digest location functionality
4201 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
4202 New function PKCS7_set_digest() to set the digest type for PKCS#7
4203 digestedData type. Add additional code to correctly generate the
4204 digestedData type and add support for this type in PKCS7 initialization
4208 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4209 structure of type "other".
4212 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4213 sure the loop does correctly stop and breaking ("division by zero")
4214 modulus operations are not performed. The (pre-generated) prime
4215 table crypto/bn/bn_prime.h was already correct, but it could not be
4216 re-generated on some platforms because of the "division by zero"
4217 situation in the script.
4218 [Ralf S. Engelschall]
4220 *) Update support for ECC-based TLS ciphersuites according to
4221 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4222 SHA-1 now is only used for "small" curves (where the
4223 representation of a field element takes up to 24 bytes); for
4224 larger curves, the field element resulting from ECDH is directly
4225 used as premaster secret.
4226 [Douglas Stebila (Sun Microsystems Laboratories)]
4228 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4229 curve secp160r1 to the tests.
4230 [Douglas Stebila (Sun Microsystems Laboratories)]
4232 *) Add the possibility to load symbols globally with DSO.
4233 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4235 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4236 control of the error stack.
4239 *) Add support for STORE in ENGINE.
4242 *) Add the STORE type. The intention is to provide a common interface
4243 to certificate and key stores, be they simple file-based stores, or
4244 HSM-type store, or LDAP stores, or...
4245 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4248 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4249 pass a list of arguments to any function as well as provide a way
4250 for a function to pass data back to the caller.
4253 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4254 works like BUF_strdup() but can be used to duplicate a portion of
4255 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4259 *) Add the function sk_find_ex() which works like sk_find(), but will
4260 return an index to an element even if an exact match couldn't be
4261 found. The index is guaranteed to point at the element where the
4262 searched-for key would be inserted to preserve sorting order.
4265 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4266 takes an extra flags argument for optional functionality. Currently,
4267 the following flags are defined:
4269 OBJ_BSEARCH_VALUE_ON_NOMATCH
4270 This one gets OBJ_bsearch_ex() to return a pointer to the first
4271 element where the comparing function returns a negative or zero
4274 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4275 This one gets OBJ_bsearch_ex() to return a pointer to the first
4276 element where the comparing function returns zero. This is useful
4277 if there are more than one element where the comparing function
4281 *) Make it possible to create self-signed certificates with 'openssl ca'
4282 in such a way that the self-signed certificate becomes part of the
4283 CA database and uses the same mechanisms for serial number generation
4284 as all other certificate signing. The new flag '-selfsign' enables
4285 this functionality. Adapt CA.sh and CA.pl.in.
4288 *) Add functionality to check the public key of a certificate request
4289 against a given private. This is useful to check that a certificate
4290 request can be signed by that key (self-signing).
4293 *) Make it possible to have multiple active certificates with the same
4294 subject in the CA index file. This is done only if the keyword
4295 'unique_subject' is set to 'no' in the main CA section (default
4296 if 'CA_default') of the configuration file. The value is saved
4297 with the database itself in a separate index attribute file,
4298 named like the index file with '.attr' appended to the name.
4301 *) Generate muti valued AVAs using '+' notation in config files for
4305 *) Support for nameConstraints certificate extension.
4308 *) Support for policyConstraints certificate extension.
4311 *) Support for policyMappings certificate extension.
4314 *) Make sure the default DSA_METHOD implementation only uses its
4315 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4316 and change its own handlers to be NULL so as to remove unnecessary
4317 indirection. This lets alternative implementations fallback to the
4318 default implementation more easily.
4321 *) Support for directoryName in GeneralName related extensions
4325 *) Make it possible to link applications using Makefile.shared.
4326 Make that possible even when linking against static libraries!
4329 *) Support for single pass processing for S/MIME signing. This now
4330 means that S/MIME signing can be done from a pipe, in addition
4331 cleartext signing (multipart/signed type) is effectively streaming
4332 and the signed data does not need to be all held in memory.
4334 This is done with a new flag PKCS7_STREAM. When this flag is set
4335 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4336 is done after the data is output (and digests calculated) in
4337 SMIME_write_PKCS7().
4340 *) Add full support for -rpath/-R, both in shared libraries and
4341 applications, at least on the platforms where it's known how
4345 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4346 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4347 will now compute a table of multiples of the generator that
4348 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4349 faster (notably in the case of a single point multiplication,
4350 scalar * generator).
4351 [Nils Larsch, Bodo Moeller]
4353 *) IPv6 support for certificate extensions. The various extensions
4354 which use the IP:a.b.c.d can now take IPv6 addresses using the
4355 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4359 *) Added an ENGINE that implements RSA by performing private key
4360 exponentiations with the GMP library. The conversions to and from
4361 GMP's mpz_t format aren't optimised nor are any montgomery forms
4362 cached, and on x86 it appears OpenSSL's own performance has caught up.
4363 However there are likely to be other architectures where GMP could
4364 provide a boost. This ENGINE is not built in by default, but it can be
4365 specified at Configure time and should be accompanied by the necessary
4366 linker additions, eg;
4367 ./config -DOPENSSL_USE_GMP -lgmp
4370 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4371 testing availability of engines with "-t" - the old behaviour is
4372 produced by increasing the feature's verbosity with "-tt".
4375 *) ECDSA routines: under certain error conditions uninitialized BN objects
4376 could be freed. Solution: make sure initialization is performed early
4377 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4381 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4382 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4383 software implementations. For DSA and DH, parameter generation can
4384 also be overriden by providing the appropriate method callbacks.
4387 *) Change the "progress" mechanism used in key-generation and
4388 primality testing to functions that take a new BN_GENCB pointer in
4389 place of callback/argument pairs. The new API functions have "_ex"
4390 postfixes and the older functions are reimplemented as wrappers for
4391 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4392 declarations of the old functions to help (graceful) attempts to
4393 migrate to the new functions. Also, the new key-generation API
4394 functions operate on a caller-supplied key-structure and return
4395 success/failure rather than returning a key or NULL - this is to
4396 help make "keygen" another member function of RSA_METHOD etc.
4398 Example for using the new callback interface:
4400 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4404 BN_GENCB_set(&my_cb, my_callback, my_arg);
4406 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4407 /* For the meaning of a, b in calls to my_callback(), see the
4408 * documentation of the function that calls the callback.
4409 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4410 * my_callback should return 1 if it wants BN_is_prime_ex()
4411 * to continue, or 0 to stop.
4416 *) Change the ZLIB compression method to be stateful, and make it
4417 available to TLS with the number defined in
4418 draft-ietf-tls-compression-04.txt.
4421 *) Add the ASN.1 structures and functions for CertificatePair, which
4422 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4424 CertificatePair ::= SEQUENCE {
4425 forward [0] Certificate OPTIONAL,
4426 reverse [1] Certificate OPTIONAL,
4427 -- at least one of the pair shall be present -- }
4429 Also implement the PEM functions to read and write certificate
4430 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4432 This needed to be defined, mostly for the sake of the LDAP
4433 attribute crossCertificatePair, but may prove useful elsewhere as
4437 *) Make it possible to inhibit symlinking of shared libraries in
4438 Makefile.shared, for Cygwin's sake.
4441 *) Extend the BIGNUM API by creating a function
4442 void BN_set_negative(BIGNUM *a, int neg);
4443 and a macro that behave like
4444 int BN_is_negative(const BIGNUM *a);
4446 to avoid the need to access 'a->neg' directly in applications.
4449 *) Implement fast modular reduction for pseudo-Mersenne primes
4450 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4451 EC_GROUP_new_curve_GFp() will now automatically use this
4453 [Nils Larsch <nla@trustcenter.de>]
4455 *) Add new lock type (CRYPTO_LOCK_BN).
4458 *) Change the ENGINE framework to automatically load engines
4459 dynamically from specific directories unless they could be
4460 found to already be built in or loaded. Move all the
4461 current engines except for the cryptodev one to a new
4463 The engines in engines/ are built as shared libraries if
4464 the "shared" options was given to ./Configure or ./config.
4465 Otherwise, they are inserted in libcrypto.a.
4466 /usr/local/ssl/engines is the default directory for dynamic
4467 engines, but that can be overriden at configure time through
4468 the usual use of --prefix and/or --openssldir, and at run
4469 time with the environment variable OPENSSL_ENGINES.
4470 [Geoff Thorpe and Richard Levitte]
4472 *) Add Makefile.shared, a helper makefile to build shared
4473 libraries. Addapt Makefile.org.
4476 *) Add version info to Win32 DLLs.
4477 [Peter 'Luna' Runestig" <peter@runestig.com>]
4479 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4480 can be added using this API to created arbitrary PKCS#12
4481 files while avoiding the low level API.
4483 New options to PKCS12_create(), key or cert can be NULL and
4484 will then be omitted from the output file. The encryption
4485 algorithm NIDs can be set to -1 for no encryption, the mac
4486 iteration count can be set to 0 to omit the mac.
4488 Enhance pkcs12 utility by making the -nokeys and -nocerts
4489 options work when creating a PKCS#12 file. New option -nomac
4490 to omit the mac, NONE can be set for an encryption algorithm.
4491 New code is modified to use the enhanced PKCS12_create()
4492 instead of the low level API.
4495 *) Extend ASN1 encoder to support indefinite length constructed
4496 encoding. This can output sequences tags and octet strings in
4497 this form. Modify pk7_asn1.c to support indefinite length
4498 encoding. This is experimental and needs additional code to
4499 be useful, such as an ASN1 bio and some enhanced streaming
4502 Extend template encode functionality so that tagging is passed
4503 down to the template encoder.
4506 *) Let 'openssl req' fail if an argument to '-newkey' is not
4507 recognized instead of using RSA as a default.
4510 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4511 As these are not official, they are not included in "ALL";
4512 the "ECCdraft" ciphersuite group alias can be used to select them.
4513 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4515 *) Add ECDH engine support.
4516 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4518 *) Add ECDH in new directory crypto/ecdh/.
4519 [Douglas Stebila (Sun Microsystems Laboratories)]
4521 *) Let BN_rand_range() abort with an error after 100 iterations
4522 without success (which indicates a broken PRNG).
4525 *) Change BN_mod_sqrt() so that it verifies that the input value
4526 is really the square of the return value. (Previously,
4527 BN_mod_sqrt would show GIGO behaviour.)
4530 *) Add named elliptic curves over binary fields from X9.62, SECG,
4531 and WAP/WTLS; add OIDs that were still missing.
4533 [Sheueling Chang Shantz and Douglas Stebila
4534 (Sun Microsystems Laboratories)]
4536 *) Extend the EC library for elliptic curves over binary fields
4537 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4540 EC_GF2m_simple_method
4544 EC_GROUP_new_curve_GF2m
4545 EC_GROUP_set_curve_GF2m
4546 EC_GROUP_get_curve_GF2m
4547 EC_POINT_set_affine_coordinates_GF2m
4548 EC_POINT_get_affine_coordinates_GF2m
4549 EC_POINT_set_compressed_coordinates_GF2m
4551 Point compression for binary fields is disabled by default for
4552 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4555 As binary polynomials are represented as BIGNUMs, various members
4556 of the EC_GROUP and EC_POINT data structures can be shared
4557 between the implementations for prime fields and binary fields;
4558 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4559 are essentially identical to their ..._GFp counterparts.
4560 (For simplicity, the '..._GFp' prefix has been dropped from
4561 various internal method names.)
4563 An internal 'field_div' method (similar to 'field_mul' and
4564 'field_sqr') has been added; this is used only for binary fields.
4566 [Sheueling Chang Shantz and Douglas Stebila
4567 (Sun Microsystems Laboratories)]
4569 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4570 through methods ('mul', 'precompute_mult').
4572 The generic implementations (now internally called 'ec_wNAF_mul'
4573 and 'ec_wNAF_precomputed_mult') remain the default if these
4574 methods are undefined.
4576 [Sheueling Chang Shantz and Douglas Stebila
4577 (Sun Microsystems Laboratories)]
4579 *) New function EC_GROUP_get_degree, which is defined through
4580 EC_METHOD. For curves over prime fields, this returns the bit
4581 length of the modulus.
4583 [Sheueling Chang Shantz and Douglas Stebila
4584 (Sun Microsystems Laboratories)]
4586 *) New functions EC_GROUP_dup, EC_POINT_dup.
4587 (These simply call ..._new and ..._copy).
4589 [Sheueling Chang Shantz and Douglas Stebila
4590 (Sun Microsystems Laboratories)]
4592 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4593 Polynomials are represented as BIGNUMs (where the sign bit is not
4594 used) in the following functions [macros]:
4597 BN_GF2m_sub [= BN_GF2m_add]
4598 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4599 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4600 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4602 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4603 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4604 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4605 BN_GF2m_cmp [= BN_ucmp]
4607 (Note that only the 'mod' functions are actually for fields GF(2^m).
4608 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4610 For some functions, an the irreducible polynomial defining a
4611 field can be given as an 'unsigned int[]' with strictly
4612 decreasing elements giving the indices of those bits that are set;
4613 i.e., p[] represents the polynomial
4614 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4616 p[0] > p[1] > ... > p[k] = 0.
4617 This applies to the following functions:
4622 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4623 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4625 BN_GF2m_mod_sqrt_arr
4626 BN_GF2m_mod_solve_quad_arr
4630 Conversion can be performed by the following functions:
4635 bntest.c has additional tests for binary polynomial arithmetic.
4637 Two implementations for BN_GF2m_mod_div() are available.
4638 The default algorithm simply uses BN_GF2m_mod_inv() and
4639 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4640 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4641 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4643 [Sheueling Chang Shantz and Douglas Stebila
4644 (Sun Microsystems Laboratories)]
4646 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4647 functionality is disabled at compile-time.
4648 [Douglas Stebila <douglas.stebila@sun.com>]
4650 *) Change default behaviour of 'openssl asn1parse' so that more
4651 information is visible when viewing, e.g., a certificate:
4653 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4654 mode the content of non-printable OCTET STRINGs is output in a
4655 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4656 avoid the appearance of a printable string.
4657 [Nils Larsch <nla@trustcenter.de>]
4659 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4661 EC_GROUP_set_asn1_flag()
4662 EC_GROUP_get_asn1_flag()
4663 EC_GROUP_set_point_conversion_form()
4664 EC_GROUP_get_point_conversion_form()
4665 These control ASN1 encoding details:
4666 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4667 has been set to OPENSSL_EC_NAMED_CURVE.
4668 - Points are encoded in uncompressed form by default; options for
4669 asn1_for are as for point2oct, namely
4670 POINT_CONVERSION_COMPRESSED
4671 POINT_CONVERSION_UNCOMPRESSED
4672 POINT_CONVERSION_HYBRID
4674 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4677 EC_GROUP_get0_seed()
4678 EC_GROUP_get_seed_len()
4679 This is used only for ASN1 purposes (so far).
4680 [Nils Larsch <nla@trustcenter.de>]
4682 *) Add 'field_type' member to EC_METHOD, which holds the NID
4683 of the appropriate field type OID. The new function
4684 EC_METHOD_get_field_type() returns this value.
4685 [Nils Larsch <nla@trustcenter.de>]
4690 EC_POINT_point2hex()
4691 EC_POINT_hex2point()
4692 providing useful interfaces to EC_POINT_point2oct() and
4693 EC_POINT_oct2point().
4694 [Nils Larsch <nla@trustcenter.de>]
4696 *) Change internals of the EC library so that the functions
4697 EC_GROUP_set_generator()
4698 EC_GROUP_get_generator()
4699 EC_GROUP_get_order()
4700 EC_GROUP_get_cofactor()
4701 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4702 to methods, which would lead to unnecessary code duplication when
4703 adding different types of curves.
4704 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4706 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4707 arithmetic, and such that modified wNAFs are generated
4708 (which avoid length expansion in many cases).
4711 *) Add a function EC_GROUP_check_discriminant() (defined via
4712 EC_METHOD) that verifies that the curve discriminant is non-zero.
4714 Add a function EC_GROUP_check() that makes some sanity tests
4715 on a EC_GROUP, its generator and order. This includes
4716 EC_GROUP_check_discriminant().
4717 [Nils Larsch <nla@trustcenter.de>]
4719 *) Add ECDSA in new directory crypto/ecdsa/.
4721 Add applications 'openssl ecparam' and 'openssl ecdsa'
4722 (these are based on 'openssl dsaparam' and 'openssl dsa').
4724 ECDSA support is also included in various other files across the
4725 library. Most notably,
4726 - 'openssl req' now has a '-newkey ecdsa:file' option;
4727 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4728 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4729 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4730 them suitable for ECDSA where domain parameters must be
4731 extracted before the specific public key;
4732 - ECDSA engine support has been added.
4733 [Nils Larsch <nla@trustcenter.de>]
4735 *) Include some named elliptic curves, and add OIDs from X9.62,
4736 SECG, and WAP/WTLS. Each curve can be obtained from the new
4738 EC_GROUP_new_by_curve_name(),
4739 and the list of available named curves can be obtained with
4740 EC_get_builtin_curves().
4741 Also add a 'curve_name' member to EC_GROUP objects, which can be
4743 EC_GROUP_set_curve_name()
4744 EC_GROUP_get_curve_name()
4745 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4747 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4748 was actually never needed) and in BN_mul(). The removal in BN_mul()
4749 required a small change in bn_mul_part_recursive() and the addition
4750 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4751 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4752 bn_sub_words() and bn_add_words() except they take arrays with
4756 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4758 *) Cleanse PEM buffers before freeing them since they may contain
4760 [Benjamin Bennett <ben@psc.edu>]
4762 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4763 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4764 authentication-only ciphersuites.
4767 *) Since AES128 and AES256 share a single mask bit in the logic of
4768 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4769 kludge to work properly if AES128 is available and AES256 isn't.
4772 *) Expand security boundary to match 1.1.1 module.
4775 *) Remove redundant features: hash file source, editing of test vectors
4776 modify fipsld to use external fips_premain.c signature.
4779 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4780 run algorithm test programs.
4783 *) Make algorithm test programs more tolerant of whitespace.
4786 *) Have SSL/TLS server implementation tolerate "mismatched" record
4787 protocol version while receiving ClientHello even if the
4788 ClientHello is fragmented. (The server can't insist on the
4789 particular protocol version it has chosen before the ServerHello
4790 message has informed the client about his choice.)
4793 *) Load error codes if they are not already present instead of using a
4794 static variable. This allows them to be cleanly unloaded and reloaded.
4797 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4799 *) Introduce limits to prevent malicious keys being able to
4800 cause a denial of service. (CVE-2006-2940)
4801 [Steve Henson, Bodo Moeller]
4803 *) Fix ASN.1 parsing of certain invalid structures that can result
4804 in a denial of service. (CVE-2006-2937) [Steve Henson]
4806 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4807 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4809 *) Fix SSL client code which could crash if connecting to a
4810 malicious SSLv2 server. (CVE-2006-4343)
4811 [Tavis Ormandy and Will Drewry, Google Security Team]
4813 *) Change ciphersuite string processing so that an explicit
4814 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4815 will no longer include "AES128-SHA"), and any other similar
4816 ciphersuite (same bitmap) from *other* protocol versions (so that
4817 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4818 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4819 changes from 0.9.8b and 0.9.8d.
4822 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4824 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4825 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4827 *) Change the Unix randomness entropy gathering to use poll() when
4828 possible instead of select(), since the latter has some
4829 undesirable limitations.
4830 [Darryl Miles via Richard Levitte and Bodo Moeller]
4832 *) Disable rogue ciphersuites:
4834 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4835 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4836 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4838 The latter two were purportedly from
4839 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4842 Also deactive the remaining ciphersuites from
4843 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4844 unofficial, and the ID has long expired.
4847 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4848 dual-core machines) and other potential thread-safety issues.
4851 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4853 *) Adapt fipsld and the build system to link against the validated FIPS
4854 module in FIPS mode.
4857 *) Fixes for VC++ 2005 build under Windows.
4860 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4861 from a Windows bash shell such as MSYS. It is autodetected from the
4862 "config" script when run from a VC++ environment. Modify standard VC++
4863 build to use fipscanister.o from the GNU make build.
4866 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4868 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4869 The value now differs depending on if you build for FIPS or not.
4870 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4871 safely run with a non-FIPSed libcrypto, as it may crash because of
4872 the difference induced by this change.
4875 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4877 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4878 (part of SSL_OP_ALL). This option used to disable the
4879 countermeasure against man-in-the-middle protocol-version
4880 rollback in the SSL 2.0 server implementation, which is a bad
4881 idea. (CVE-2005-2969)
4883 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4884 for Information Security, National Institute of Advanced Industrial
4885 Science and Technology [AIST], Japan)]
4887 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4888 mainly for FIPS compliance and not fully integrated at this stage.
4891 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4892 the exponentiation using a fixed-length exponent. (Otherwise,
4893 the information leaked through timing could expose the secret key
4894 after many signatures; cf. Bleichenbacher's attack on DSA with
4898 *) Make a new fixed-window mod_exp implementation the default for
4899 RSA, DSA, and DH private-key operations so that the sequence of
4900 squares and multiplies and the memory access pattern are
4901 independent of the particular secret key. This will mitigate
4902 cache-timing and potential related attacks.
4904 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4905 and this is automatically used by BN_mod_exp_mont() if the new flag
4906 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4907 will use this BN flag for private exponents unless the flag
4908 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4909 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4911 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4913 *) Change the client implementation for SSLv23_method() and
4914 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4915 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4916 (Previously, the SSL 2.0 backwards compatible Client Hello
4917 message format would be used even with SSL_OP_NO_SSLv2.)
4920 *) Add support for smime-type MIME parameter in S/MIME messages which some
4924 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4925 a threadsafe manner. Modify rsa code to use new function and add calls
4926 to dsa and dh code (which had race conditions before).
4929 *) Include the fixed error library code in the C error file definitions
4930 instead of fixing them up at runtime. This keeps the error code
4931 structures constant.
4934 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4936 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4939 *) Fixes for newer kerberos headers. NB: the casts are needed because
4940 the 'length' field is signed on one version and unsigned on another
4941 with no (?) obvious way to tell the difference, without these VC++
4942 complains. Also the "definition" of FAR (blank) is no longer included
4943 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4944 some needed definitions.
4947 *) Undo Cygwin change.
4950 *) Added support for proxy certificates according to RFC 3820.
4951 Because they may be a security thread to unaware applications,
4952 they must be explicitely allowed in run-time. See
4953 docs/HOWTO/proxy_certificates.txt for further information.
4956 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4958 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4959 server and client random values. Previously
4960 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4961 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4963 This change has negligible security impact because:
4965 1. Server and client random values still have 24 bytes of pseudo random
4968 2. Server and client random values are sent in the clear in the initial
4971 3. The master secret is derived using the premaster secret (48 bytes in
4972 size for static RSA ciphersuites) as well as client server and random
4975 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4978 [Stephen Henson, reported by UK NISCC]
4980 *) Use Windows randomness collection on Cygwin.
4983 *) Fix hang in EGD/PRNGD query when communication socket is closed
4984 prematurely by EGD/PRNGD.
4985 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4987 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4990 *) Back-port of selected performance improvements from development
4991 branch, as well as improved support for PowerPC platforms.
4994 *) Add lots of checks for memory allocation failure, error codes to indicate
4995 failure and freeing up memory if a failure occurs.
4996 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4998 *) Add new -passin argument to dgst.
5001 *) Perform some character comparisons of different types in X509_NAME_cmp:
5002 this is needed for some certificates that reencode DNs into UTF8Strings
5003 (in violation of RFC3280) and can't or wont issue name rollover
5007 *) Make an explicit check during certificate validation to see that
5008 the CA setting in each certificate on the chain is correct. As a
5009 side effect always do the following basic checks on extensions,
5010 not just when there's an associated purpose to the check:
5012 - if there is an unhandled critical extension (unless the user
5013 has chosen to ignore this fault)
5014 - if the path length has been exceeded (if one is set at all)
5015 - that certain extensions fit the associated purpose (if one has
5019 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5021 *) Avoid a race condition when CRLs are checked in a multi threaded
5022 environment. This would happen due to the reordering of the revoked
5023 entries during signature checking and serial number lookup. Now the
5024 encoding is cached and the serial number sort performed under a lock.
5025 Add new STACK function sk_is_sorted().
5028 *) Add Delta CRL to the extension code.
5031 *) Various fixes to s3_pkt.c so alerts are sent properly.
5032 [David Holmes <d.holmes@f5.com>]
5034 *) Reduce the chances of duplicate issuer name and serial numbers (in
5035 violation of RFC3280) using the OpenSSL certificate creation utilities.
5036 This is done by creating a random 64 bit value for the initial serial
5037 number when a serial number file is created or when a self signed
5038 certificate is created using 'openssl req -x509'. The initial serial
5039 number file is created using 'openssl x509 -next_serial' in CA.pl
5040 rather than being initialized to 1.
5043 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5045 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5046 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5047 [Joe Orton, Steve Henson]
5049 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5051 [Joe Orton, Steve Henson]
5053 *) Make it possible to have multiple active certificates with the same
5054 subject in the CA index file. This is done only if the keyword
5055 'unique_subject' is set to 'no' in the main CA section (default
5056 if 'CA_default') of the configuration file. The value is saved
5057 with the database itself in a separate index attribute file,
5058 named like the index file with '.attr' appended to the name.
5061 *) X509 verify fixes. Disable broken certificate workarounds when
5062 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5063 keyUsage extension present. Don't accept CRLs with unhandled critical
5064 extensions: since verify currently doesn't process CRL extensions this
5065 rejects a CRL with *any* critical extensions. Add new verify error codes
5069 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5070 A clarification of RFC2560 will require the use of OCTET STRINGs and
5071 some implementations cannot handle the current raw format. Since OpenSSL
5072 copies and compares OCSP nonces as opaque blobs without any attempt at
5073 parsing them this should not create any compatibility issues.
5076 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5077 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5078 this HMAC (and other) operations are several times slower than OpenSSL
5082 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5083 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5085 *) Use the correct content when signing type "other".
5088 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5090 *) Fix various bugs revealed by running the NISCC test suite:
5092 Stop out of bounds reads in the ASN1 code when presented with
5093 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5095 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5097 If verify callback ignores invalid public key errors don't try to check
5098 certificate signature with the NULL public key.
5102 *) New -ignore_err option in ocsp application to stop the server
5103 exiting on the first error in a request.
5106 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5107 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5111 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5112 extra data after the compression methods not only for TLS 1.0
5113 but also for SSL 3.0 (as required by the specification).
5114 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5116 *) Change X509_certificate_type() to mark the key as exported/exportable
5117 when it's 512 *bits* long, not 512 bytes.
5120 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5121 blocks during encryption.
5124 *) Various fixes to base64 BIO and non blocking I/O. On write
5125 flushes were not handled properly if the BIO retried. On read
5126 data was not being buffered properly and had various logic bugs.
5127 This also affects blocking I/O when the data being decoded is a
5131 *) Various S/MIME bugfixes and compatibility changes:
5132 output correct application/pkcs7 MIME type if
5133 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5134 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5135 of files as .eml work). Correctly handle very long lines in MIME
5139 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5141 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5142 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5143 a protocol version number mismatch like a decryption error
5144 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5147 *) Turn on RSA blinding by default in the default implementation
5148 to avoid a timing attack. Applications that don't want it can call
5149 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5150 They would be ill-advised to do so in most cases.
5151 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5153 *) Change RSA blinding code so that it works when the PRNG is not
5154 seeded (in this case, the secret RSA exponent is abused as
5155 an unpredictable seed -- if it is not unpredictable, there
5156 is no point in blinding anyway). Make RSA blinding thread-safe
5157 by remembering the creator's thread ID in rsa->blinding and
5158 having all other threads use local one-time blinding factors
5159 (this requires more computation than sharing rsa->blinding, but
5160 avoids excessive locking; and if an RSA object is not shared
5161 between threads, blinding will still be very fast).
5164 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5165 ENGINE as defaults for all supported algorithms irrespective of
5166 the 'flags' parameter. 'flags' is now honoured, so applications
5167 should make sure they are passing it correctly.
5170 *) Target "mingw" now allows native Windows code to be generated in
5171 the Cygwin environment as well as with the MinGW compiler.
5174 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5176 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5177 via timing by performing a MAC computation even if incorrrect
5178 block cipher padding has been found. This is a countermeasure
5179 against active attacks where the attacker has to distinguish
5180 between bad padding and a MAC verification error. (CVE-2003-0078)
5182 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5183 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5184 Martin Vuagnoux (EPFL, Ilion)]
5186 *) Make the no-err option work as intended. The intention with no-err
5187 is not to have the whole error stack handling routines removed from
5188 libcrypto, it's only intended to remove all the function name and
5189 reason texts, thereby removing some of the footprint that may not
5190 be interesting if those errors aren't displayed anyway.
5192 NOTE: it's still possible for any application or module to have it's
5193 own set of error texts inserted. The routines are there, just not
5194 used by default when no-err is given.
5197 *) Add support for FreeBSD on IA64.
5198 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5200 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5201 Kerberos function mit_des_cbc_cksum(). Before this change,
5202 the value returned by DES_cbc_cksum() was like the one from
5203 mit_des_cbc_cksum(), except the bytes were swapped.
5204 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5206 *) Allow an application to disable the automatic SSL chain building.
5207 Before this a rather primitive chain build was always performed in
5208 ssl3_output_cert_chain(): an application had no way to send the
5209 correct chain if the automatic operation produced an incorrect result.
5211 Now the chain builder is disabled if either:
5213 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5215 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5217 The reasoning behind this is that an application would not want the
5218 auto chain building to take place if extra chain certificates are
5219 present and it might also want a means of sending no additional
5220 certificates (for example the chain has two certificates and the
5224 *) Add the possibility to build without the ENGINE framework.
5225 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5227 *) Under Win32 gmtime() can return NULL: check return value in
5228 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5231 *) DSA routines: under certain error conditions uninitialized BN objects
5232 could be freed. Solution: make sure initialization is performed early
5233 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5234 Nils Larsch <nla@trustcenter.de> via PR#459)
5237 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5238 checked on reconnect on the client side, therefore session resumption
5239 could still fail with a "ssl session id is different" error. This
5240 behaviour is masked when SSL_OP_ALL is used due to
5241 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5242 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5243 followup to PR #377.
5246 *) IA-32 assembler support enhancements: unified ELF targets, support
5247 for SCO/Caldera platforms, fix for Cygwin shared build.
5250 *) Add support for FreeBSD on sparc64. As a consequence, support for
5251 FreeBSD on non-x86 processors is separate from x86 processors on
5252 the config script, much like the NetBSD support.
5253 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5255 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5257 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5260 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5261 code (06) was taken as the first octet of the session ID and the last
5262 octet was ignored consequently. As a result SSLv2 client side session
5263 caching could not have worked due to the session ID mismatch between
5265 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5269 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5270 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5274 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5275 seems that in spite of existing for more than a year, many application
5276 author have done nothing to provide the necessary callbacks, which
5277 means that this particular engine will not work properly anywhere.
5278 This is a very unfortunate situation which forces us, in the name
5279 of usability, to give the hw_ncipher.c a static lock, which is part
5281 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5282 appear in 0.9.8 or later. We EXPECT application authors to have
5283 dealt properly with this when 0.9.8 is released (unless we actually
5284 make such changes in the libcrypto locking code that changes will
5285 have to be made anyway).
5288 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5289 octets have been read, EOF or an error occurs. Without this change
5290 some truncated ASN1 structures will not produce an error.
5293 *) Disable Heimdal support, since it hasn't been fully implemented.
5294 Still give the possibility to force the use of Heimdal, but with
5295 warnings and a request that patches get sent to openssl-dev.
5298 *) Add the VC-CE target, introduce the WINCE sysname, and add
5299 INSTALL.WCE and appropriate conditionals to make it build.
5300 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5302 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5303 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5304 edit numbers of the version.
5305 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5307 *) Introduce safe string copy and catenation functions
5308 (BUF_strlcpy() and BUF_strlcat()).
5309 [Ben Laurie (CHATS) and Richard Levitte]
5311 *) Avoid using fixed-size buffers for one-line DNs.
5312 [Ben Laurie (CHATS)]
5314 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5315 resizing buffers containing secrets, and use where appropriate.
5316 [Ben Laurie (CHATS)]
5318 *) Avoid using fixed size buffers for configuration file location.
5319 [Ben Laurie (CHATS)]
5321 *) Avoid filename truncation for various CA files.
5322 [Ben Laurie (CHATS)]
5324 *) Use sizeof in preference to magic numbers.
5325 [Ben Laurie (CHATS)]
5327 *) Avoid filename truncation in cert requests.
5328 [Ben Laurie (CHATS)]
5330 *) Add assertions to check for (supposedly impossible) buffer
5332 [Ben Laurie (CHATS)]
5334 *) Don't cache truncated DNS entries in the local cache (this could
5335 potentially lead to a spoofing attack).
5336 [Ben Laurie (CHATS)]
5338 *) Fix various buffers to be large enough for hex/decimal
5339 representations in a platform independent manner.
5340 [Ben Laurie (CHATS)]
5342 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5343 resizing buffers containing secrets, and use where appropriate.
5344 [Ben Laurie (CHATS)]
5346 *) Add BIO_indent() to avoid much slightly worrying code to do
5348 [Ben Laurie (CHATS)]
5350 *) Convert sprintf()/BIO_puts() to BIO_printf().
5351 [Ben Laurie (CHATS)]
5353 *) buffer_gets() could terminate with the buffer only half
5355 [Ben Laurie (CHATS)]
5357 *) Add assertions to prevent user-supplied crypto functions from
5358 overflowing internal buffers by having large block sizes, etc.
5359 [Ben Laurie (CHATS)]
5361 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5363 [Ben Laurie (CHATS)]
5365 *) Eliminate unused copy of key in RC4.
5366 [Ben Laurie (CHATS)]
5368 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5369 [Ben Laurie (CHATS)]
5371 *) Fix off-by-one error in EGD path.
5372 [Ben Laurie (CHATS)]
5374 *) If RANDFILE path is too long, ignore instead of truncating.
5375 [Ben Laurie (CHATS)]
5377 *) Eliminate unused and incorrectly sized X.509 structure
5379 [Ben Laurie (CHATS)]
5381 *) Eliminate unused and dangerous function knumber().
5382 [Ben Laurie (CHATS)]
5384 *) Eliminate unused and dangerous structure, KSSL_ERR.
5385 [Ben Laurie (CHATS)]
5387 *) Protect against overlong session ID context length in an encoded
5388 session object. Since these are local, this does not appear to be
5390 [Ben Laurie (CHATS)]
5392 *) Change from security patch (see 0.9.6e below) that did not affect
5393 the 0.9.6 release series:
5395 Remote buffer overflow in SSL3 protocol - an attacker could
5396 supply an oversized master key in Kerberos-enabled versions.
5398 [Ben Laurie (CHATS)]
5400 *) Change the SSL kerb5 codes to match RFC 2712.
5403 *) Make -nameopt work fully for req and add -reqopt switch.
5404 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5406 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5407 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5409 *) Make sure tests can be performed even if the corresponding algorithms
5410 have been removed entirely. This was also the last step to make
5411 OpenSSL compilable with DJGPP under all reasonable conditions.
5412 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5414 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5415 to allow version independent disabling of normally unselected ciphers,
5416 which may be activated as a side-effect of selecting a single cipher.
5418 (E.g., cipher list string "RSA" enables ciphersuites that are left
5419 out of "ALL" because they do not provide symmetric encryption.
5420 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5421 [Lutz Jaenicke, Bodo Moeller]
5423 *) Add appropriate support for separate platform-dependent build
5424 directories. The recommended way to make a platform-dependent
5425 build directory is the following (tested on Linux), maybe with
5428 # Place yourself outside of the OpenSSL source tree. In
5429 # this example, the environment variable OPENSSL_SOURCE
5430 # is assumed to contain the absolute OpenSSL source directory.
5431 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5432 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5433 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5434 mkdir -p `dirname $F`
5435 ln -s $OPENSSL_SOURCE/$F $F
5438 To be absolutely sure not to disturb the source tree, a "make clean"
5439 is a good thing. If it isn't successfull, don't worry about it,
5440 it probably means the source directory is very clean.
5443 *) Make sure any ENGINE control commands make local copies of string
5444 pointers passed to them whenever necessary. Otherwise it is possible
5445 the caller may have overwritten (or deallocated) the original string
5446 data when a later ENGINE operation tries to use the stored values.
5447 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5449 *) Improve diagnostics in file reading and command-line digests.
5450 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5452 *) Add AES modes CFB and OFB to the object database. Correct an
5453 error in AES-CFB decryption.
5456 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5457 allows existing EVP_CIPHER_CTX structures to be reused after
5458 calling EVP_*Final(). This behaviour is used by encryption
5459 BIOs and some applications. This has the side effect that
5460 applications must explicitly clean up cipher contexts with
5461 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5464 *) Check the values of dna and dnb in bn_mul_recursive before calling
5465 bn_mul_comba (a non zero value means the a or b arrays do not contain
5466 n2 elements) and fallback to bn_mul_normal if either is not zero.
5469 *) Fix escaping of non-ASCII characters when using the -subj option
5470 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5473 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5474 form for "surname", serialNumber has no short form.
5475 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5476 therefore remove "mail" short name for "internet 7".
5477 The OID for unique identifiers in X509 certificates is
5478 x500UniqueIdentifier, not uniqueIdentifier.
5479 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5482 *) Add an "init" command to the ENGINE config module and auto initialize
5483 ENGINEs. Without any "init" command the ENGINE will be initialized
5484 after all ctrl commands have been executed on it. If init=1 the
5485 ENGINE is initailized at that point (ctrls before that point are run
5486 on the uninitialized ENGINE and after on the initialized one). If
5487 init=0 then the ENGINE will not be iniatialized at all.
5490 *) Fix the 'app_verify_callback' interface so that the user-defined
5491 argument is actually passed to the callback: In the
5492 SSL_CTX_set_cert_verify_callback() prototype, the callback
5493 declaration has been changed from
5496 int (*cb)(X509_STORE_CTX *,void *);
5497 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5498 i=s->ctx->app_verify_callback(&ctx)
5499 has been changed into
5500 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5502 To update applications using SSL_CTX_set_cert_verify_callback(),
5503 a dummy argument can be added to their callback functions.
5504 [D. K. Smetters <smetters@parc.xerox.com>]
5506 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5507 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5509 *) Add and OPENSSL_LOAD_CONF define which will cause
5510 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5511 This allows older applications to transparently support certain
5512 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5513 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5514 load the config file and OPENSSL_add_all_algorithms_conf() which will
5515 always load it have also been added.
5518 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5519 Adjust NIDs and EVP layer.
5520 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5522 *) Config modules support in openssl utility.
5524 Most commands now load modules from the config file,
5525 though in a few (such as version) this isn't done
5526 because it couldn't be used for anything.
5528 In the case of ca and req the config file used is
5529 the same as the utility itself: that is the -config
5530 command line option can be used to specify an
5534 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5535 use "openssl_conf" if filename is NULL use default openssl config file.
5538 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5539 config section name. Add a new flag to tolerate a missing config file
5540 and move code to CONF_modules_load_file().
5543 *) Support for crypto accelerator cards from Accelerated Encryption
5544 Processing, www.aep.ie. (Use engine 'aep')
5545 The support was copied from 0.9.6c [engine] and adapted/corrected
5546 to work with the new engine framework.
5547 [AEP Inc. and Richard Levitte]
5549 *) Support for SureWare crypto accelerator cards from Baltimore
5550 Technologies. (Use engine 'sureware')
5551 The support was copied from 0.9.6c [engine] and adapted
5552 to work with the new engine framework.
5555 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5556 make the newer ENGINE framework commands for the CHIL engine work.
5557 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5559 *) Make it possible to produce shared libraries on ReliantUNIX.
5560 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5562 *) Add the configuration target debug-linux-ppro.
5563 Make 'openssl rsa' use the general key loading routines
5564 implemented in apps.c, and make those routines able to
5565 handle the key format FORMAT_NETSCAPE and the variant
5567 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5569 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5570 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5572 *) Add -keyform to rsautl, and document -engine.
5573 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5575 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5576 BIO_R_NO_SUCH_FILE error code rather than the generic
5577 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5580 *) Add new functions
5582 ERR_peek_last_error_line
5583 ERR_peek_last_error_line_data.
5584 These are similar to
5587 ERR_peek_error_line_data,
5588 but report on the latest error recorded rather than the first one
5589 still in the error queue.
5590 [Ben Laurie, Bodo Moeller]
5592 *) default_algorithms option in ENGINE config module. This allows things
5594 default_algorithms = ALL
5595 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5598 *) Prelminary ENGINE config module.
5601 *) New experimental application configuration code.
5604 *) Change the AES code to follow the same name structure as all other
5605 symmetric ciphers, and behave the same way. Move everything to
5606 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5607 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5609 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5610 [Ben Laurie and Theo de Raadt]
5612 *) Add option to output public keys in req command.
5613 [Massimiliano Pala madwolf@openca.org]
5615 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5616 (up to about 10% better than before for P-192 and P-224).
5619 *) New functions/macros
5621 SSL_CTX_set_msg_callback(ctx, cb)
5622 SSL_CTX_set_msg_callback_arg(ctx, arg)
5623 SSL_set_msg_callback(ssl, cb)
5624 SSL_set_msg_callback_arg(ssl, arg)
5626 to request calling a callback function
5628 void cb(int write_p, int version, int content_type,
5629 const void *buf, size_t len, SSL *ssl, void *arg)
5631 whenever a protocol message has been completely received
5632 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5633 protocol version according to which the SSL library interprets
5634 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5635 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5636 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5637 specification (change_cipher_spec(20), alert(21), handshake(22)).
5638 'buf' and 'len' point to the actual message, 'ssl' to the
5639 SSL object, and 'arg' is the application-defined value set by
5640 SSL[_CTX]_set_msg_callback_arg().
5642 'openssl s_client' and 'openssl s_server' have new '-msg' options
5643 to enable a callback that displays all protocol messages.
5646 *) Change the shared library support so shared libraries are built as
5647 soon as the corresponding static library is finished, and thereby get
5648 openssl and the test programs linked against the shared library.
5649 This still only happens when the keyword "shard" has been given to
5650 the configuration scripts.
5652 NOTE: shared library support is still an experimental thing, and
5653 backward binary compatibility is still not guaranteed.
5654 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5656 *) Add support for Subject Information Access extension.
5657 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5659 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5660 additional bytes when new memory had to be allocated, not just
5661 when reusing an existing buffer.
5664 *) New command line and configuration option 'utf8' for the req command.
5665 This allows field values to be specified as UTF8 strings.
5668 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5669 runs for the former and machine-readable output for the latter.
5672 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5673 of the e-mail address in the DN (i.e., it will go into a certificate
5674 extension only). The new configuration file option 'email_in_dn = no'
5675 has the same effect.
5676 [Massimiliano Pala madwolf@openca.org]
5678 *) Change all functions with names starting with des_ to be starting
5679 with DES_ instead. Add wrappers that are compatible with libdes,
5680 but are named _ossl_old_des_*. Finally, add macros that map the
5681 des_* symbols to the corresponding _ossl_old_des_* if libdes
5682 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5683 desired, the des_* symbols will be mapped to DES_*, with one
5686 Since we provide two compatibility mappings, the user needs to
5687 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5688 compatibility is desired. The default (i.e., when that macro
5689 isn't defined) is OpenSSL 0.9.6c compatibility.
5691 There are also macros that enable and disable the support of old
5692 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5693 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5694 are defined, the default will apply: to support the old des routines.
5696 In either case, one must include openssl/des.h to get the correct
5697 definitions. Do not try to just include openssl/des_old.h, that
5700 NOTE: This is a major break of an old API into a new one. Software
5701 authors are encouraged to switch to the DES_ style functions. Some
5702 time in the future, des_old.h and the libdes compatibility functions
5703 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5704 default), and then completely removed.
5707 *) Test for certificates which contain unsupported critical extensions.
5708 If such a certificate is found during a verify operation it is
5709 rejected by default: this behaviour can be overridden by either
5710 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5711 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5712 X509_supported_extension() has also been added which returns 1 if a
5713 particular extension is supported.
5716 *) Modify the behaviour of EVP cipher functions in similar way to digests
5717 to retain compatibility with existing code.
5720 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5721 compatibility with existing code. In particular the 'ctx' parameter does
5722 not have to be to be initialized before the call to EVP_DigestInit() and
5723 it is tidied up after a call to EVP_DigestFinal(). New function
5724 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5725 EVP_MD_CTX_copy() changed to not require the destination to be
5726 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5727 requires the destination to be valid.
5729 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5730 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5733 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5734 so that complete 'Handshake' protocol structures are kept in memory
5735 instead of overwriting 'msg_type' and 'length' with 'body' data.
5738 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5739 [Massimo Santin via Richard Levitte]
5741 *) Major restructuring to the underlying ENGINE code. This includes
5742 reduction of linker bloat, separation of pure "ENGINE" manipulation
5743 (initialisation, etc) from functionality dealing with implementations
5744 of specific crypto iterfaces. This change also introduces integrated
5745 support for symmetric ciphers and digest implementations - so ENGINEs
5746 can now accelerate these by providing EVP_CIPHER and EVP_MD
5747 implementations of their own. This is detailed in crypto/engine/README
5748 as it couldn't be adequately described here. However, there are a few
5749 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5750 were changed in the original introduction of ENGINE code have now
5751 reverted back - the hooking from this code to ENGINE is now a good
5752 deal more passive and at run-time, operations deal directly with
5753 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5754 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5755 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5756 they were not being used by the framework as there is no concept of a
5757 BIGNUM_METHOD and they could not be generalised to the new
5758 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5759 ENGINE_cpy() has been removed as it cannot be consistently defined in
5763 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5766 *) Change mkdef.pl to sort symbols that get the same entry number,
5767 and make sure the automatically generated functions ERR_load_*
5768 become part of libeay.num as well.
5771 *) New function SSL_renegotiate_pending(). This returns true once
5772 renegotiation has been requested (either SSL_renegotiate() call
5773 or HelloRequest/ClientHello receveived from the peer) and becomes
5774 false once a handshake has been completed.
5775 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5776 sends a HelloRequest, but does not ensure that a handshake takes
5777 place. SSL_renegotiate_pending() is useful for checking if the
5778 client has followed the request.)
5781 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5782 By default, clients may request session resumption even during
5783 renegotiation (if session ID contexts permit); with this option,
5784 session resumption is possible only in the first handshake.
5786 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5787 more bits available for options that should not be part of
5788 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5791 *) Add some demos for certificate and certificate request creation.
5794 *) Make maximum certificate chain size accepted from the peer application
5795 settable (SSL*_get/set_max_cert_list()), as proposed by
5796 "Douglas E. Engert" <deengert@anl.gov>.
5799 *) Add support for shared libraries for Unixware-7
5800 (Boyd Lynn Gerber <gerberb@zenez.com>).
5803 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5804 be done prior to destruction. Use this to unload error strings from
5805 ENGINEs that load their own error strings. NB: This adds two new API
5806 functions to "get" and "set" this destroy handler in an ENGINE.
5809 *) Alter all existing ENGINE implementations (except "openssl" and
5810 "openbsd") to dynamically instantiate their own error strings. This
5811 makes them more flexible to be built both as statically-linked ENGINEs
5812 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5813 Also, add stub code to each that makes building them as self-contained
5814 shared-libraries easier (see README.ENGINE).
5817 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5818 implementations into applications that are completely implemented in
5819 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5820 commands that can be used to configure what shared-library to load and
5821 to control aspects of the way it is handled. Also, made an update to
5822 the README.ENGINE file that brings its information up-to-date and
5823 provides some information and instructions on the "dynamic" ENGINE
5824 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5827 *) Make it possible to unload ranges of ERR strings with a new
5828 "ERR_unload_strings" function.
5831 *) Add a copy() function to EVP_MD.
5834 *) Make EVP_MD routines take a context pointer instead of just the
5835 md_data void pointer.
5838 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5839 that the digest can only process a single chunk of data
5840 (typically because it is provided by a piece of
5841 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5842 is only going to provide a single chunk of data, and hence the
5843 framework needn't accumulate the data for oneshot drivers.
5846 *) As with "ERR", make it possible to replace the underlying "ex_data"
5847 functions. This change also alters the storage and management of global
5848 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5849 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5850 index counters. The API functions that use this state have been changed
5851 to take a "class_index" rather than pointers to the class's local STACK
5852 and counter, and there is now an API function to dynamically create new
5853 classes. This centralisation allows us to (a) plug a lot of the
5854 thread-safety problems that existed, and (b) makes it possible to clean
5855 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5856 such data would previously have always leaked in application code and
5857 workarounds were in place to make the memory debugging turn a blind eye
5858 to it. Application code that doesn't use this new function will still
5859 leak as before, but their memory debugging output will announce it now
5860 rather than letting it slide.
5862 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5863 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5864 has a return value to indicate success or failure.
5867 *) Make it possible to replace the underlying "ERR" functions such that the
5868 global state (2 LHASH tables and 2 locks) is only used by the "default"
5869 implementation. This change also adds two functions to "get" and "set"
5870 the implementation prior to it being automatically set the first time
5871 any other ERR function takes place. Ie. an application can call "get",
5872 pass the return value to a module it has just loaded, and that module
5873 can call its own "set" function using that value. This means the
5874 module's "ERR" operations will use (and modify) the error state in the
5875 application and not in its own statically linked copy of OpenSSL code.
5878 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5879 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5880 the operation, and provides a more encapsulated way for external code
5881 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5882 to use these functions rather than manually incrementing the counts.
5884 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5887 *) Add EVP test program.
5890 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5893 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5894 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5895 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5896 These allow a CRL to be built without having to access X509_CRL fields
5897 directly. Modify 'ca' application to use new functions.
5900 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5901 bug workarounds. Rollback attack detection is a security feature.
5902 The problem will only arise on OpenSSL servers when TLSv1 is not
5903 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5904 Software authors not wanting to support TLSv1 will have special reasons
5905 for their choice and can explicitly enable this option.
5906 [Bodo Moeller, Lutz Jaenicke]
5908 *) Rationalise EVP so it can be extended: don't include a union of
5909 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5910 (similar to those existing for EVP_CIPHER_CTX).
5915 EVP_MD_CTX_init(&md); /* new function call */
5916 EVP_DigestInit(&md, EVP_sha1());
5917 EVP_DigestUpdate(&md, in, len);
5918 EVP_DigestFinal(&md, out, NULL);
5919 EVP_MD_CTX_cleanup(&md); /* new function call */
5923 *) Make DES key schedule conform to the usual scheme, as well as
5924 correcting its structure. This means that calls to DES functions
5925 now have to pass a pointer to a des_key_schedule instead of a
5926 plain des_key_schedule (which was actually always a pointer
5929 des_key_schedule ks;
5931 des_set_key_checked(..., &ks);
5932 des_ncbc_encrypt(..., &ks, ...);
5934 (Note that a later change renames 'des_...' into 'DES_...'.)
5937 *) Initial reduction of linker bloat: the use of some functions, such as
5938 PEM causes large amounts of unused functions to be linked in due to
5939 poor organisation. For example pem_all.c contains every PEM function
5940 which has a knock on effect of linking in large amounts of (unused)
5941 ASN1 code. Grouping together similar functions and splitting unrelated
5942 functions prevents this.
5945 *) Cleanup of EVP macros.
5948 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5949 correct _ecb suffix.
5952 *) Add initial OCSP responder support to ocsp application. The
5953 revocation information is handled using the text based index
5954 use by the ca application. The responder can either handle
5955 requests generated internally, supplied in files (for example
5956 via a CGI script) or using an internal minimal server.
5959 *) Add configuration choices to get zlib compression for TLS.
5962 *) Changes to Kerberos SSL for RFC 2712 compliance:
5963 1. Implemented real KerberosWrapper, instead of just using
5964 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5965 2. Implemented optional authenticator field of KerberosWrapper.
5967 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5968 and authenticator structs; see crypto/krb5/.
5970 Generalized Kerberos calls to support multiple Kerberos libraries.
5971 [Vern Staats <staatsvr@asc.hpc.mil>,
5972 Jeffrey Altman <jaltman@columbia.edu>
5973 via Richard Levitte]
5975 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5976 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5977 values for each of the key sizes rather than having just
5978 parameters (and 'speed' generating keys each time).
5981 *) Speed up EVP routines.
5984 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5985 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5986 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5987 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5989 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5990 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5991 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5994 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5996 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5999 *) Added the OS2-EMX target.
6000 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
6002 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
6003 to support NCONF routines in extension code. New function CONF_set_nconf()
6004 to allow functions which take an NCONF to also handle the old LHASH
6005 structure: this means that the old CONF compatible routines can be
6006 retained (in particular wrt extensions) without having to duplicate the
6007 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
6010 *) Enhance the general user interface with mechanisms for inner control
6011 and with possibilities to have yes/no kind of prompts.
6014 *) Change all calls to low level digest routines in the library and
6015 applications to use EVP. Add missing calls to HMAC_cleanup() and
6016 don't assume HMAC_CTX can be copied using memcpy().
6017 [Verdon Walker <VWalker@novell.com>, Steve Henson]
6019 *) Add the possibility to control engines through control names but with
6020 arbitrary arguments instead of just a string.
6021 Change the key loaders to take a UI_METHOD instead of a callback
6022 function pointer. NOTE: this breaks binary compatibility with earlier
6023 versions of OpenSSL [engine].
6024 Adapt the nCipher code for these new conditions and add a card insertion
6028 *) Enhance the general user interface with mechanisms to better support
6029 dialog box interfaces, application-defined prompts, the possibility
6030 to use defaults (for example default passwords from somewhere else)
6031 and interrupts/cancellations.
6034 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6035 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6038 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6039 tidy up some unnecessarily weird code in 'sk_new()').
6040 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6042 *) Change the key loading routines for ENGINEs to use the same kind
6043 callback (pem_password_cb) as all other routines that need this
6047 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6048 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6049 than this minimum value is recommended.
6052 *) New random seeder for OpenVMS, using the system process statistics
6053 that are easily reachable.
6056 *) Windows apparently can't transparently handle global
6057 variables defined in DLLs. Initialisations such as:
6059 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6061 wont compile. This is used by the any applications that need to
6062 declare their own ASN1 modules. This was fixed by adding the option
6063 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6064 needed for static libraries under Win32.
6067 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6068 setting of purpose and trust fields. New X509_STORE trust and
6069 purpose functions and tidy up setting in other SSL functions.
6072 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6073 structure. These are inherited by X509_STORE_CTX when it is
6074 initialised. This allows various defaults to be set in the
6075 X509_STORE structure (such as flags for CRL checking and custom
6076 purpose or trust settings) for functions which only use X509_STORE_CTX
6077 internally such as S/MIME.
6079 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6080 trust settings if they are not set in X509_STORE. This allows X509_STORE
6081 purposes and trust (in S/MIME for example) to override any set by default.
6083 Add command line options for CRL checking to smime, s_client and s_server
6087 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6088 are set then the CRL is looked up in the X509_STORE structure and
6089 its validity and signature checked, then if the certificate is found
6090 in the CRL the verify fails with a revoked error.
6092 Various new CRL related callbacks added to X509_STORE_CTX structure.
6094 Command line options added to 'verify' application to support this.
6096 This needs some additional work, such as being able to handle multiple
6097 CRLs with different times, extension based lookup (rather than just
6098 by subject name) and ultimately more complete V2 CRL extension
6102 *) Add a general user interface API (crypto/ui/). This is designed
6103 to replace things like des_read_password and friends (backward
6104 compatibility functions using this new API are provided).
6105 The purpose is to remove prompting functions from the DES code
6106 section as well as provide for prompting through dialog boxes in
6107 a window system and the like.
6110 *) Add "ex_data" support to ENGINE so implementations can add state at a
6111 per-structure level rather than having to store it globally.
6114 *) Make it possible for ENGINE structures to be copied when retrieved by
6115 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6116 This causes the "original" ENGINE structure to act like a template,
6117 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6118 operational state can be localised to each ENGINE structure, despite the
6119 fact they all share the same "methods". New ENGINE structures returned in
6120 this case have no functional references and the return value is the single
6121 structural reference. This matches the single structural reference returned
6122 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6126 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6127 needs to match any other type at all we need to manually clear the
6131 *) Changes to the "openssl engine" utility to include;
6132 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6133 about an ENGINE's available control commands.
6134 - executing control commands from command line arguments using the
6135 '-pre' and '-post' switches. '-post' is only used if '-t' is
6136 specified and the ENGINE is successfully initialised. The syntax for
6137 the individual commands are colon-separated, for example;
6138 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6141 *) New dynamic control command support for ENGINEs. ENGINEs can now
6142 declare their own commands (numbers), names (strings), descriptions,
6143 and input types for run-time discovery by calling applications. A
6144 subset of these commands are implicitly classed as "executable"
6145 depending on their input type, and only these can be invoked through
6146 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6147 can be based on user input, config files, etc). The distinction is
6148 that "executable" commands cannot return anything other than a boolean
6149 result and can only support numeric or string input, whereas some
6150 discoverable commands may only be for direct use through
6151 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6152 pointers, or other custom uses. The "executable" commands are to
6153 support parameterisations of ENGINE behaviour that can be
6154 unambiguously defined by ENGINEs and used consistently across any
6155 OpenSSL-based application. Commands have been added to all the
6156 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6157 control over shared-library paths without source code alterations.
6160 *) Changed all ENGINE implementations to dynamically allocate their
6161 ENGINEs rather than declaring them statically. Apart from this being
6162 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6163 this also allows the implementations to compile without using the
6164 internal engine_int.h header.
6167 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6168 'const' value. Any code that should be able to modify a RAND_METHOD
6169 should already have non-const pointers to it (ie. they should only
6170 modify their own ones).
6173 *) Made a variety of little tweaks to the ENGINE code.
6174 - "atalla" and "ubsec" string definitions were moved from header files
6175 to C code. "nuron" string definitions were placed in variables
6176 rather than hard-coded - allowing parameterisation of these values
6177 later on via ctrl() commands.
6178 - Removed unused "#if 0"'d code.
6179 - Fixed engine list iteration code so it uses ENGINE_free() to release
6180 structural references.
6181 - Constified the RAND_METHOD element of ENGINE structures.
6182 - Constified various get/set functions as appropriate and added
6183 missing functions (including a catch-all ENGINE_cpy that duplicates
6184 all ENGINE values onto a new ENGINE except reference counts/state).
6185 - Removed NULL parameter checks in get/set functions. Setting a method
6186 or function to NULL is a way of cancelling out a previously set
6187 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6188 and doesn't justify the extra error symbols and code.
6189 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6190 flags from engine_int.h to engine.h.
6191 - Changed prototypes for ENGINE handler functions (init(), finish(),
6192 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6195 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6196 to the algorithm using long division. The binary algorithm can be
6197 used only if the modulus is odd. On 32-bit systems, it is faster
6198 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6199 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6200 up to 450 bits. In 64-bit environments, the binary algorithm
6201 appears to be advantageous for much longer moduli; here we use it
6202 for moduli up to 2048 bits.
6205 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6206 could not support the combine flag in choice fields.
6209 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6210 extensions from a certificate request to the certificate.
6213 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6214 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6215 file: this allows the display of the certificate about to be
6216 signed to be customised, to allow certain fields to be included
6217 or excluded and extension details. The old system didn't display
6218 multicharacter strings properly, omitted fields not in the policy
6219 and couldn't display additional details such as extensions.
6222 *) Function EC_POINTs_mul for multiple scalar multiplication
6223 of an arbitrary number of elliptic curve points
6224 \sum scalars[i]*points[i],
6225 optionally including the generator defined for the EC_GROUP:
6226 scalar*generator + \sum scalars[i]*points[i].
6228 EC_POINT_mul is a simple wrapper function for the typical case
6229 that the point list has just one item (besides the optional
6233 *) First EC_METHODs for curves over GF(p):
6235 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6236 operations and provides various method functions that can also
6237 operate with faster implementations of modular arithmetic.
6239 EC_GFp_mont_method() reuses most functions that are part of
6240 EC_GFp_simple_method, but uses Montgomery arithmetic.
6242 [Bodo Moeller; point addition and point doubling
6243 implementation directly derived from source code provided by
6244 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6246 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6247 crypto/ec/ec_lib.c):
6249 Curves are EC_GROUP objects (with an optional group generator)
6250 based on EC_METHODs that are built into the library.
6252 Points are EC_POINT objects based on EC_GROUP objects.
6254 Most of the framework would be able to handle curves over arbitrary
6255 finite fields, but as there are no obvious types for fields other
6256 than GF(p), some functions are limited to that for now.
6259 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6260 that the file contains a complete HTTP response.
6263 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6264 change the def and num file printf format specifier from "%-40sXXX"
6265 to "%-39s XXX". The latter will always guarantee a space after the
6266 field while the former will cause them to run together if the field
6267 is 40 of more characters long.
6270 *) Constify the cipher and digest 'method' functions and structures
6271 and modify related functions to take constant EVP_MD and EVP_CIPHER
6275 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6276 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6279 *) Modify EVP_Digest*() routines so they now return values. Although the
6280 internal software routines can never fail additional hardware versions
6284 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6286 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6287 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6290 ERR_R_NESTED_ASN1_ERROR
6292 ERR_R_MISSING_ASN1_EOS
6293 were 4 .. 9, conflicting with
6294 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6296 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6297 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6299 Add new error code 'ERR_R_INTERNAL_ERROR'.
6302 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6306 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6307 sets the subject name for a new request or supersedes the
6308 subject name in a given request. Formats that can be parsed are
6309 'CN=Some Name, OU=myOU, C=IT'
6311 'CN=Some Name/OU=myOU/C=IT'.
6313 Add options '-batch' and '-verbose' to 'openssl req'.
6314 [Massimiliano Pala <madwolf@hackmasters.net>]
6316 *) Introduce the possibility to access global variables through
6317 functions on platform were that's the best way to handle exporting
6318 global variables in shared libraries. To enable this functionality,
6319 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6320 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6321 is normally done by Configure or something similar).
6323 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6324 in the source file (foo.c) like this:
6326 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6327 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6329 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6330 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6332 OPENSSL_DECLARE_GLOBAL(int,foo);
6333 #define foo OPENSSL_GLOBAL_REF(foo)
6334 OPENSSL_DECLARE_GLOBAL(double,bar);
6335 #define bar OPENSSL_GLOBAL_REF(bar)
6337 The #defines are very important, and therefore so is including the
6338 header file everywhere where the defined globals are used.
6340 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6341 of ASN.1 items, but that structure is a bit different.
6343 The largest change is in util/mkdef.pl which has been enhanced with
6344 better and easier to understand logic to choose which symbols should
6345 go into the Windows .def files as well as a number of fixes and code
6346 cleanup (among others, algorithm keywords are now sorted
6347 lexicographically to avoid constant rewrites).
6350 *) In BN_div() keep a copy of the sign of 'num' before writing the
6351 result to 'rm' because if rm==num the value will be overwritten
6352 and produce the wrong result if 'num' is negative: this caused
6353 problems with BN_mod() and BN_nnmod().
6356 *) Function OCSP_request_verify(). This checks the signature on an
6357 OCSP request and verifies the signer certificate. The signer
6358 certificate is just checked for a generic purpose and OCSP request
6362 *) Add OCSP_check_validity() function to check the validity of OCSP
6363 responses. OCSP responses are prepared in real time and may only
6364 be a few seconds old. Simply checking that the current time lies
6365 between thisUpdate and nextUpdate max reject otherwise valid responses
6366 caused by either OCSP responder or client clock inaccuracy. Instead
6367 we allow thisUpdate and nextUpdate to fall within a certain period of
6368 the current time. The age of the response can also optionally be
6369 checked. Two new options -validity_period and -status_age added to
6373 *) If signature or public key algorithm is unrecognized print out its
6374 OID rather that just UNKNOWN.
6377 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6378 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6379 ID to be generated from the issuer certificate alone which can then be
6380 passed to OCSP_id_issuer_cmp().
6383 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6384 ASN1 modules to export functions returning ASN1_ITEM pointers
6385 instead of the ASN1_ITEM structures themselves. This adds several
6386 new macros which allow the underlying ASN1 function/structure to
6387 be accessed transparently. As a result code should not use ASN1_ITEM
6388 references directly (such as &X509_it) but instead use the relevant
6389 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6390 use of the new ASN1 code on platforms where exporting structures
6391 is problematical (for example in shared libraries) but exporting
6392 functions returning pointers to structures is not.
6395 *) Add support for overriding the generation of SSL/TLS session IDs.
6396 These callbacks can be registered either in an SSL_CTX or per SSL.
6397 The purpose of this is to allow applications to control, if they wish,
6398 the arbitrary values chosen for use as session IDs, particularly as it
6399 can be useful for session caching in multiple-server environments. A
6400 command-line switch for testing this (and any client code that wishes
6401 to use such a feature) has been added to "s_server".
6402 [Geoff Thorpe, Lutz Jaenicke]
6404 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6405 of the form '#if defined(...) || defined(...) || ...' and
6406 '#if !defined(...) && !defined(...) && ...'. This also avoids
6407 the growing number of special cases it was previously handling.
6410 *) Make all configuration macros available for application by making
6411 sure they are available in opensslconf.h, by giving them names starting
6412 with "OPENSSL_" to avoid conflicts with other packages and by making
6413 sure e_os2.h will cover all platform-specific cases together with
6415 Additionally, it is now possible to define configuration/platform-
6416 specific names (called "system identities"). In the C code, these
6417 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6418 macro with the name beginning with "OPENSSL_SYS_", which is determined
6419 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6423 *) New option -set_serial to 'req' and 'x509' this allows the serial
6424 number to use to be specified on the command line. Previously self
6425 signed certificates were hard coded with serial number 0 and the
6426 CA options of 'x509' had to use a serial number in a file which was
6430 *) New options to 'ca' utility to support V2 CRL entry extensions.
6431 Currently CRL reason, invalidity date and hold instruction are
6432 supported. Add new CRL extensions to V3 code and some new objects.
6435 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6436 disable standard block padding (aka PKCS#5 padding) in the EVP
6437 API, which was previously mandatory. This means that the data is
6438 not padded in any way and so the total length much be a multiple
6439 of the block size, otherwise an error occurs.
6442 *) Initial (incomplete) OCSP SSL support.
6445 *) New function OCSP_parse_url(). This splits up a URL into its host,
6446 port and path components: primarily to parse OCSP URLs. New -url
6447 option to ocsp utility.
6450 *) New nonce behavior. The return value of OCSP_check_nonce() now
6451 reflects the various checks performed. Applications can decide
6452 whether to tolerate certain situations such as an absent nonce
6453 in a response when one was present in a request: the ocsp application
6454 just prints out a warning. New function OCSP_add1_basic_nonce()
6455 this is to allow responders to include a nonce in a response even if
6456 the request is nonce-less.
6459 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6460 skipped when using openssl x509 multiple times on a single input file,
6461 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6464 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6465 set string type: to handle setting ASN1_TIME structures. Fix ca
6466 utility to correctly initialize revocation date of CRLs.
6469 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6470 the clients preferred ciphersuites and rather use its own preferences.
6471 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6472 Internet Explorer by ensuring unchanged hash method during stepup.
6473 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6476 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6477 to aes and add a new 'exist' option to print out symbols that don't
6481 *) Additional options to ocsp utility to allow flags to be set and
6482 additional certificates supplied.
6485 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6486 OCSP client a number of certificate to only verify the response
6490 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6491 handle the new API. Currently only ECB, CBC modes supported. Add new
6494 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6495 Encryption Standard (AES) Ciphersuites for Transport Layer
6496 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6497 not enabled by default and were not part of the "ALL" ciphersuite
6498 alias because they were not yet official; they could be
6499 explicitly requested by specifying the "AESdraft" ciphersuite
6500 group alias. In the final release of OpenSSL 0.9.7, the group
6501 alias is called "AES" and is part of "ALL".)
6502 [Ben Laurie, Steve Henson, Bodo Moeller]
6504 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6505 request to response.
6508 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6509 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6510 extract information from a certificate request. OCSP_response_create()
6511 creates a response and optionally adds a basic response structure.
6512 OCSP_basic_add1_status() adds a complete single response to a basic
6513 response and returns the OCSP_SINGLERESP structure just added (to allow
6514 extensions to be included for example). OCSP_basic_add1_cert() adds a
6515 certificate to a basic response and OCSP_basic_sign() signs a basic
6516 response with various flags. New helper functions ASN1_TIME_check()
6517 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6518 (converts ASN1_TIME to GeneralizedTime).
6521 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6522 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6523 structure from a certificate. X509_pubkey_digest() digests the public_key
6524 contents: this is used in various key identifiers.
6527 *) Make sk_sort() tolerate a NULL argument.
6528 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6530 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6531 passed by the function are trusted implicitly. If any of them signed the
6532 response then it is assumed to be valid and is not verified.
6535 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6536 to data. This was previously part of the PKCS7 ASN1 code. This
6537 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6538 [Steve Henson, reported by Kenneth R. Robinette
6539 <support@securenetterm.com>]
6541 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6542 routines: without these tracing memory leaks is very painful.
6543 Fix leaks in PKCS12 and PKCS7 routines.
6546 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6547 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6548 effectively meant GeneralizedTime would never be used. Now it
6549 is initialised to -1 but X509_time_adj() now has to check the value
6550 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6551 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6552 [Steve Henson, reported by Kenneth R. Robinette
6553 <support@securenetterm.com>]
6555 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6556 result in a zero length in the ASN1_INTEGER structure which was
6557 not consistent with the structure when d2i_ASN1_INTEGER() was used
6558 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6559 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6560 where it did not print out a minus for negative ASN1_INTEGER.
6563 *) Add summary printout to ocsp utility. The various functions which
6564 convert status values to strings have been renamed to:
6565 OCSP_response_status_str(), OCSP_cert_status_str() and
6566 OCSP_crl_reason_str() and are no longer static. New options
6567 to verify nonce values and to disable verification. OCSP response
6568 printout format cleaned up.
6571 *) Add additional OCSP certificate checks. These are those specified
6572 in RFC2560. This consists of two separate checks: the CA of the
6573 certificate being checked must either be the OCSP signer certificate
6574 or the issuer of the OCSP signer certificate. In the latter case the
6575 OCSP signer certificate must contain the OCSP signing extended key
6576 usage. This check is performed by attempting to match the OCSP
6577 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6578 in the OCSP_CERTID structures of the response.
6581 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6582 and related routines. This uses the standard OpenSSL certificate
6583 verify routines to perform initial checks (just CA validity) and
6584 to obtain the certificate chain. Then additional checks will be
6585 performed on the chain. Currently the root CA is checked to see
6586 if it is explicitly trusted for OCSP signing. This is used to set
6587 a root CA as a global signing root: that is any certificate that
6588 chains to that CA is an acceptable OCSP signing certificate.
6591 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6592 extensions from a separate configuration file.
6593 As when reading extensions from the main configuration file,
6594 the '-extensions ...' option may be used for specifying the
6596 [Massimiliano Pala <madwolf@comune.modena.it>]
6598 *) New OCSP utility. Allows OCSP requests to be generated or
6599 read. The request can be sent to a responder and the output
6600 parsed, outputed or printed in text form. Not complete yet:
6601 still needs to check the OCSP response validity.
6604 *) New subcommands for 'openssl ca':
6605 'openssl ca -status <serial>' prints the status of the cert with
6606 the given serial number (according to the index file).
6607 'openssl ca -updatedb' updates the expiry status of certificates
6609 [Massimiliano Pala <madwolf@comune.modena.it>]
6611 *) New '-newreq-nodes' command option to CA.pl. This is like
6612 '-newreq', but calls 'openssl req' with the '-nodes' option
6613 so that the resulting key is not encrypted.
6614 [Damien Miller <djm@mindrot.org>]
6616 *) New configuration for the GNU Hurd.
6617 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6619 *) Initial code to implement OCSP basic response verify. This
6620 is currently incomplete. Currently just finds the signer's
6621 certificate and verifies the signature on the response.
6624 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6625 value of OPENSSLDIR. This is available via the new '-d' option
6626 to 'openssl version', and is also included in 'openssl version -a'.
6629 *) Allowing defining memory allocation callbacks that will be given
6630 file name and line number information in additional arguments
6631 (a const char* and an int). The basic functionality remains, as
6632 well as the original possibility to just replace malloc(),
6633 realloc() and free() by functions that do not know about these
6634 additional arguments. To register and find out the current
6635 settings for extended allocation functions, the following
6636 functions are provided:
6638 CRYPTO_set_mem_ex_functions
6639 CRYPTO_set_locked_mem_ex_functions
6640 CRYPTO_get_mem_ex_functions
6641 CRYPTO_get_locked_mem_ex_functions
6643 These work the same way as CRYPTO_set_mem_functions and friends.
6644 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6645 extended allocation function is enabled.
6646 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6647 a conventional allocation function is enabled.
6648 [Richard Levitte, Bodo Moeller]
6650 *) Finish off removing the remaining LHASH function pointer casts.
6651 There should no longer be any prototype-casting required when using
6652 the LHASH abstraction, and any casts that remain are "bugs". See
6653 the callback types and macros at the head of lhash.h for details
6654 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6657 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6658 If /dev/[u]random devices are not available or do not return enough
6659 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6661 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6662 /etc/entropy will be queried once each in this sequence, quering stops
6663 when enough entropy was collected without querying more sockets.
6666 *) Change the Unix RAND_poll() variant to be able to poll several
6667 random devices, as specified by DEVRANDOM, until a sufficient amount
6668 of data has been collected. We spend at most 10 ms on each file
6669 (select timeout) and read in non-blocking mode. DEVRANDOM now
6670 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6671 (previously it was just the string "/dev/urandom"), so on typical
6672 platforms the 10 ms delay will never occur.
6673 Also separate out the Unix variant to its own file, rand_unix.c.
6674 For VMS, there's a currently-empty rand_vms.c.
6677 *) Move OCSP client related routines to ocsp_cl.c. These
6678 provide utility functions which an application needing
6679 to issue a request to an OCSP responder and analyse the
6680 response will typically need: as opposed to those which an
6681 OCSP responder itself would need which will be added later.
6683 OCSP_request_sign() signs an OCSP request with an API similar
6684 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6685 response. OCSP_response_get1_basic() extracts basic response
6686 from response. OCSP_resp_find_status(): finds and extracts status
6687 information from an OCSP_CERTID structure (which will be created
6688 when the request structure is built). These are built from lower
6689 level functions which work on OCSP_SINGLERESP structures but
6690 wont normally be used unless the application wishes to examine
6691 extensions in the OCSP response for example.
6693 Replace nonce routines with a pair of functions.
6694 OCSP_request_add1_nonce() adds a nonce value and optionally
6695 generates a random value. OCSP_check_nonce() checks the
6696 validity of the nonce in an OCSP response.
6699 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6700 This doesn't copy the supplied OCSP_CERTID and avoids the
6701 need to free up the newly created id. Change return type
6702 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6703 This can then be used to add extensions to the request.
6704 Deleted OCSP_request_new(), since most of its functionality
6705 is now in OCSP_REQUEST_new() (and the case insensitive name
6706 clash) apart from the ability to set the request name which
6707 will be added elsewhere.
6710 *) Update OCSP API. Remove obsolete extensions argument from
6711 various functions. Extensions are now handled using the new
6712 OCSP extension code. New simple OCSP HTTP function which
6713 can be used to send requests and parse the response.
6716 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6717 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6718 uses the special reorder version of SET OF to sort the attributes
6719 and reorder them to match the encoded order. This resolves a long
6720 standing problem: a verify on a PKCS7 structure just after signing
6721 it used to fail because the attribute order did not match the
6722 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6723 it uses the received order. This is necessary to tolerate some broken
6724 software that does not order SET OF. This is handled by encoding
6725 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6726 to produce the required SET OF.
6729 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6730 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6731 files to get correct declarations of the ASN.1 item variables.
6734 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6735 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6736 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6737 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6738 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6739 ASN1_ITEM and no wrapper functions.
6742 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6743 replace the old function pointer based I/O routines. Change most of
6744 the *_d2i_bio() and *_d2i_fp() functions to use these.
6747 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6748 lines, recognice more "algorithms" that can be deselected, and make
6749 it complain about algorithm deselection that isn't recognised.
6752 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6753 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6754 to use new functions. Add NO_ASN1_OLD which can be set to remove
6755 some old style ASN1 functions: this can be used to determine if old
6756 code will still work when these eventually go away.
6759 *) New extension functions for OCSP structures, these follow the
6760 same conventions as certificates and CRLs.
6763 *) New function X509V3_add1_i2d(). This automatically encodes and
6764 adds an extension. Its behaviour can be customised with various
6765 flags to append, replace or delete. Various wrappers added for
6766 certifcates and CRLs.
6769 *) Fix to avoid calling the underlying ASN1 print routine when
6770 an extension cannot be parsed. Correct a typo in the
6771 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6774 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
6775 entries for variables.
6778 *) Add functionality to apps/openssl.c for detecting locking
6779 problems: As the program is single-threaded, all we have
6780 to do is register a locking callback using an array for
6781 storing which locks are currently held by the program.
6784 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6785 SSL_get_ex_data_X509_STORE_idx(), which is used in
6786 ssl_verify_cert_chain() and thus can be called at any time
6787 during TLS/SSL handshakes so that thread-safety is essential.
6788 Unfortunately, the ex_data design is not at all suited
6789 for multi-threaded use, so it probably should be abolished.
6792 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6793 [Broadcom, tweaked and integrated by Geoff Thorpe]
6795 *) Move common extension printing code to new function
6796 X509V3_print_extensions(). Reorganise OCSP print routines and
6797 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6800 *) New function X509_signature_print() to remove duplication in some
6804 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6805 set (this was treated exactly the same as SET OF previously). This
6806 is used to reorder the STACK representing the structure to match the
6807 encoding. This will be used to get round a problem where a PKCS7
6808 structure which was signed could not be verified because the STACK
6809 order did not reflect the encoded order.
6812 *) Reimplement the OCSP ASN1 module using the new code.
6815 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6816 for its ASN1 operations. The old style function pointers still exist
6817 for now but they will eventually go away.
6820 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6821 completely replaces the old ASN1 functionality with a table driven
6822 encoder and decoder which interprets an ASN1_ITEM structure describing
6823 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6824 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6825 has also been converted to the new form.
6828 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6829 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6830 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6831 for negative moduli.
6834 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6835 of not touching the result's sign bit.
6838 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6842 *) Changed the LHASH code to use prototypes for callbacks, and created
6843 macros to declare and implement thin (optionally static) functions
6844 that provide type-safety and avoid function pointer casting for the
6845 type-specific callbacks.
6848 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6850 [Veers Staats <staatsvr@asc.hpc.mil>,
6851 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6853 *) Reformat the FAQ so the different questions and answers can be divided
6854 in sections depending on the subject.
6857 *) Have the zlib compression code load ZLIB.DLL dynamically under
6861 *) New function BN_mod_sqrt for computing square roots modulo a prime
6862 (using the probabilistic Tonelli-Shanks algorithm unless
6863 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6864 be handled deterministically).
6865 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6867 *) Make BN_mod_inverse faster by explicitly handling small quotients
6868 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6869 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6872 *) New function BN_kronecker.
6875 *) Fix BN_gcd so that it works on negative inputs; the result is
6876 positive unless both parameters are zero.
6877 Previously something reasonably close to an infinite loop was
6878 possible because numbers could be growing instead of shrinking
6879 in the implementation of Euclid's algorithm.
6882 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6883 sign of the number in question.
6885 Fix BN_is_word(a,w) to work correctly for w == 0.
6887 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6888 because its test if the absolute value of 'a' equals 'w'.
6889 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6890 it exists mostly for use in the implementations of BN_is_zero(),
6891 BN_is_one(), and BN_is_word().
6894 *) New function BN_swap.
6897 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6898 the exponentiation functions are more likely to produce reasonable
6899 results on negative inputs.
6902 *) Change BN_mod_mul so that the result is always non-negative.
6903 Previously, it could be negative if one of the factors was negative;
6904 I don't think anyone really wanted that behaviour.
6907 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6908 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6909 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6910 and add new functions:
6919 BN_mod_lshift1_quick
6923 These functions always generate non-negative results.
6925 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6926 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6928 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6929 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6930 be reduced modulo m.
6931 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6934 The following entry accidentily appeared in the CHANGES file
6935 distributed with OpenSSL 0.9.7. The modifications described in
6936 it do *not* apply to OpenSSL 0.9.7.
6938 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6939 was actually never needed) and in BN_mul(). The removal in BN_mul()
6940 required a small change in bn_mul_part_recursive() and the addition
6941 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6942 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6943 bn_sub_words() and bn_add_words() except they take arrays with
6948 *) In 'openssl passwd', verify passwords read from the terminal
6949 unless the '-salt' option is used (which usually means that
6950 verification would just waste user's time since the resulting
6951 hash is going to be compared with some given password hash)
6952 or the new '-noverify' option is used.
6954 This is an incompatible change, but it does not affect
6955 non-interactive use of 'openssl passwd' (passwords on the command
6956 line, '-stdin' option, '-in ...' option) and thus should not
6960 *) Remove all references to RSAref, since there's no more need for it.
6963 *) Make DSO load along a path given through an environment variable
6964 (SHLIB_PATH) with shl_load().
6967 *) Constify the ENGINE code as a result of BIGNUM constification.
6968 Also constify the RSA code and most things related to it. In a
6969 few places, most notable in the depth of the ASN.1 code, ugly
6970 casts back to non-const were required (to be solved at a later
6974 *) Make it so the openssl application has all engines loaded by default.
6977 *) Constify the BIGNUM routines a little more.
6980 *) Add the following functions:
6982 ENGINE_load_cswift()
6984 ENGINE_load_atalla()
6986 ENGINE_load_builtin_engines()
6988 That way, an application can itself choose if external engines that
6989 are built-in in OpenSSL shall ever be used or not. The benefit is
6990 that applications won't have to be linked with libdl or other dso
6991 libraries unless it's really needed.
6993 Changed 'openssl engine' to load all engines on demand.
6994 Changed the engine header files to avoid the duplication of some
6995 declarations (they differed!).
6998 *) 'openssl engine' can now list capabilities.
7001 *) Better error reporting in 'openssl engine'.
7004 *) Never call load_dh_param(NULL) in s_server.
7007 *) Add engine application. It can currently list engines by name and
7008 identity, and test if they are actually available.
7011 *) Improve RPM specification file by forcing symbolic linking and making
7012 sure the installed documentation is also owned by root.root.
7013 [Damien Miller <djm@mindrot.org>]
7015 *) Give the OpenSSL applications more possibilities to make use of
7016 keys (public as well as private) handled by engines.
7019 *) Add OCSP code that comes from CertCo.
7022 *) Add VMS support for the Rijndael code.
7025 *) Added untested support for Nuron crypto accelerator.
7028 *) Add support for external cryptographic devices. This code was
7029 previously distributed separately as the "engine" branch.
7030 [Geoff Thorpe, Richard Levitte]
7032 *) Rework the filename-translation in the DSO code. It is now possible to
7033 have far greater control over how a "name" is turned into a filename
7034 depending on the operating environment and any oddities about the
7035 different shared library filenames on each system.
7038 *) Support threads on FreeBSD-elf in Configure.
7041 *) Fix for SHA1 assembly problem with MASM: it produces
7042 warnings about corrupt line number information when assembling
7043 with debugging information. This is caused by the overlapping
7045 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7048 NCONF_get_number() has no error checking at all. As a replacement,
7049 NCONF_get_number_e() is defined (_e for "error checking") and is
7050 promoted strongly. The old NCONF_get_number is kept around for
7051 binary backward compatibility.
7052 Make it possible for methods to load from something other than a BIO,
7053 by providing a function pointer that is given a name instead of a BIO.
7054 For example, this could be used to load configuration data from an
7058 *) Fix for non blocking accept BIOs. Added new I/O special reason
7059 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7060 with non blocking I/O was not possible because no retry code was
7061 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7065 *) Added the beginnings of Rijndael support.
7068 *) Fix for bug in DirectoryString mask setting. Add support for
7069 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7070 to allow certificate printing to more controllable, additional
7071 'certopt' option to 'x509' to allow new printing options to be
7075 *) Clean old EAY MD5 hack from e_os.h.
7078 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7080 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7081 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7082 [Joe Orton, Steve Henson]
7084 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7086 *) Fix additional bug revealed by the NISCC test suite:
7088 Stop bug triggering large recursion when presented with
7089 certain ASN.1 tags (CVE-2003-0851)
7092 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7094 *) Fix various bugs revealed by running the NISCC test suite:
7096 Stop out of bounds reads in the ASN1 code when presented with
7097 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7099 If verify callback ignores invalid public key errors don't try to check
7100 certificate signature with the NULL public key.
7104 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7105 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7109 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7110 extra data after the compression methods not only for TLS 1.0
7111 but also for SSL 3.0 (as required by the specification).
7112 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7114 *) Change X509_certificate_type() to mark the key as exported/exportable
7115 when it's 512 *bits* long, not 512 bytes.
7118 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7120 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7121 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7122 a protocol version number mismatch like a decryption error
7123 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7126 *) Turn on RSA blinding by default in the default implementation
7127 to avoid a timing attack. Applications that don't want it can call
7128 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7129 They would be ill-advised to do so in most cases.
7130 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7132 *) Change RSA blinding code so that it works when the PRNG is not
7133 seeded (in this case, the secret RSA exponent is abused as
7134 an unpredictable seed -- if it is not unpredictable, there
7135 is no point in blinding anyway). Make RSA blinding thread-safe
7136 by remembering the creator's thread ID in rsa->blinding and
7137 having all other threads use local one-time blinding factors
7138 (this requires more computation than sharing rsa->blinding, but
7139 avoids excessive locking; and if an RSA object is not shared
7140 between threads, blinding will still be very fast).
7143 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7145 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7146 via timing by performing a MAC computation even if incorrrect
7147 block cipher padding has been found. This is a countermeasure
7148 against active attacks where the attacker has to distinguish
7149 between bad padding and a MAC verification error. (CVE-2003-0078)
7151 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7152 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7153 Martin Vuagnoux (EPFL, Ilion)]
7155 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7157 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7158 memory from it's contents. This is done with a counter that will
7159 place alternating values in each byte. This can be used to solve
7160 two issues: 1) the removal of calls to memset() by highly optimizing
7161 compilers, and 2) cleansing with other values than 0, since those can
7162 be read through on certain media, for example a swap space on disk.
7165 *) Bugfix: client side session caching did not work with external caching,
7166 because the session->cipher setting was not restored when reloading
7167 from the external cache. This problem was masked, when
7168 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7169 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7172 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7173 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7174 [Zeev Lieber <zeev-l@yahoo.com>]
7176 *) Undo an undocumented change introduced in 0.9.6e which caused
7177 repeated calls to OpenSSL_add_all_ciphers() and
7178 OpenSSL_add_all_digests() to be ignored, even after calling
7182 *) Change the default configuration reader to deal with last line not
7183 being properly terminated.
7186 *) Change X509_NAME_cmp() so it applies the special rules on handling
7187 DN values that are of type PrintableString, as well as RDNs of type
7188 emailAddress where the value has the type ia5String.
7189 [stefank@valicert.com via Richard Levitte]
7191 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7192 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7193 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7194 the bitwise-OR of the two for use by the majority of applications
7195 wanting this behaviour, and update the docs. The documented
7196 behaviour and actual behaviour were inconsistent and had been
7197 changing anyway, so this is more a bug-fix than a behavioural
7199 [Geoff Thorpe, diagnosed by Nadav Har'El]
7201 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7202 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7205 *) Fix initialization code race conditions in
7206 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7207 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7208 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7209 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7210 ssl2_get_cipher_by_char(),
7211 ssl3_get_cipher_by_char().
7212 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7214 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7215 the cached sessions are flushed, as the remove_cb() might use ex_data
7216 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7217 (see [openssl.org #212]).
7218 [Geoff Thorpe, Lutz Jaenicke]
7220 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7221 length, instead of the encoding length to d2i_ASN1_OBJECT.
7224 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7226 *) [In 0.9.6g-engine release:]
7227 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7228 [Lynn Gazis <lgazis@rainbow.com>]
7230 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7232 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7233 and get fix the header length calculation.
7234 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7235 Alon Kantor <alonk@checkpoint.com> (and others),
7238 *) Use proper error handling instead of 'assertions' in buffer
7239 overflow checks added in 0.9.6e. This prevents DoS (the
7240 assertions could call abort()).
7241 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7243 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7245 *) Add various sanity checks to asn1_get_length() to reject
7246 the ASN1 length bytes if they exceed sizeof(long), will appear
7247 negative or the content length exceeds the length of the
7249 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7251 *) Fix cipher selection routines: ciphers without encryption had no flags
7252 for the cipher strength set and where therefore not handled correctly
7253 by the selection routines (PR #130).
7256 *) Fix EVP_dsa_sha macro.
7260 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7261 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7262 that was added in OpenSSL 0.9.6d.
7264 As the countermeasure turned out to be incompatible with some
7265 broken SSL implementations, the new option is part of SSL_OP_ALL.
7266 SSL_OP_ALL is usually employed when compatibility with weird SSL
7267 implementations is desired (e.g. '-bugs' option to 's_client' and
7268 's_server'), so the new option is automatically set in many
7272 *) Changes in security patch:
7274 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7275 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7276 Air Force Materiel Command, USAF, under agreement number
7279 *) Add various sanity checks to asn1_get_length() to reject
7280 the ASN1 length bytes if they exceed sizeof(long), will appear
7281 negative or the content length exceeds the length of the
7282 supplied buffer. (CVE-2002-0659)
7283 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7285 *) Assertions for various potential buffer overflows, not known to
7287 [Ben Laurie (CHATS)]
7289 *) Various temporary buffers to hold ASCII versions of integers were
7290 too small for 64 bit platforms. (CVE-2002-0655)
7291 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7293 *) Remote buffer overflow in SSL3 protocol - an attacker could
7294 supply an oversized session ID to a client. (CVE-2002-0656)
7295 [Ben Laurie (CHATS)]
7297 *) Remote buffer overflow in SSL2 protocol - an attacker could
7298 supply an oversized client master key. (CVE-2002-0656)
7299 [Ben Laurie (CHATS)]
7301 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7303 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7304 encoded as NULL) with id-dsa-with-sha1.
7305 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7307 *) Check various X509_...() return values in apps/req.c.
7308 [Nils Larsch <nla@trustcenter.de>]
7310 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7311 an end-of-file condition would erronously be flagged, when the CRLF
7312 was just at the end of a processed block. The bug was discovered when
7313 processing data through a buffering memory BIO handing the data to a
7314 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7315 <ptsekov@syntrex.com> and Nedelcho Stanev.
7318 *) Implement a countermeasure against a vulnerability recently found
7319 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7320 before application data chunks to avoid the use of known IVs
7321 with data potentially chosen by the attacker.
7324 *) Fix length checks in ssl3_get_client_hello().
7327 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7328 to prevent ssl3_read_internal() from incorrectly assuming that
7329 ssl3_read_bytes() found application data while handshake
7330 processing was enabled when in fact s->s3->in_read_app_data was
7331 merely automatically cleared during the initial handshake.
7332 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7334 *) Fix object definitions for Private and Enterprise: they were not
7335 recognized in their shortname (=lowercase) representation. Extend
7336 obj_dat.pl to issue an error when using undefined keywords instead
7337 of silently ignoring the problem (Svenning Sorensen
7338 <sss@sss.dnsalias.net>).
7341 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7342 generators, i.e. generators other than 2 and 5. (Previously, the
7343 code did not properly initialise the 'add' and 'rem' values to
7344 BN_generate_prime().)
7346 In the new general case, we do not insist that 'generator' is
7347 actually a primitive root: This requirement is rather pointless;
7348 a generator of the order-q subgroup is just as good, if not
7352 *) Map new X509 verification errors to alerts. Discovered and submitted by
7353 Tom Wu <tom@arcot.com>.
7356 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7357 returning non-zero before the data has been completely received
7358 when using non-blocking I/O.
7359 [Bodo Moeller; problem pointed out by John Hughes]
7361 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7362 [Ben Laurie, Lutz Jaenicke]
7364 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7365 Yoram Zahavi <YoramZ@gilian.com>).
7368 *) Add information about CygWin 1.3 and on, and preserve proper
7369 configuration for the versions before that.
7370 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7372 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7373 check whether we deal with a copy of a session and do not delete from
7374 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7375 <izhar@checkpoint.com>.
7378 *) Do not store session data into the internal session cache, if it
7379 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7380 flag is set). Proposed by Aslam <aslam@funk.com>.
7383 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7387 *) [In 0.9.6d-engine release:]
7388 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7389 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7391 *) Add the configuration target linux-s390x.
7392 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7394 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7395 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7396 variable as an indication that a ClientHello message has been
7397 received. As the flag value will be lost between multiple
7398 invocations of ssl3_accept when using non-blocking I/O, the
7399 function may not be aware that a handshake has actually taken
7400 place, thus preventing a new session from being added to the
7403 To avoid this problem, we now set s->new_session to 2 instead of
7404 using a local variable.
7405 [Lutz Jaenicke, Bodo Moeller]
7407 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7408 if the SSL_R_LENGTH_MISMATCH error is detected.
7409 [Geoff Thorpe, Bodo Moeller]
7411 *) New 'shared_ldflag' column in Configure platform table.
7414 *) Fix EVP_CIPHER_mode macro.
7415 ["Dan S. Camper" <dan@bti.net>]
7417 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7418 type, we must throw them away by setting rr->length to 0.
7419 [D P Chang <dpc@qualys.com>]
7421 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7423 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7424 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7425 worked incorrectly for those cases where range = 10..._2 and
7426 3*range is two bits longer than range.)
7429 *) Only add signing time to PKCS7 structures if it is not already
7433 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7434 OBJ_ld_ce should be OBJ_id_ce.
7435 Also some ip-pda OIDs in crypto/objects/objects.txt were
7436 incorrect (cf. RFC 3039).
7437 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7439 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7440 returns early because it has nothing to do.
7441 [Andy Schneider <andy.schneider@bjss.co.uk>]
7443 *) [In 0.9.6c-engine release:]
7444 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7445 [Andy Schneider <andy.schneider@bjss.co.uk>]
7447 *) [In 0.9.6c-engine release:]
7448 Add support for Cryptographic Appliance's keyserver technology.
7449 (Use engine 'keyclient')
7450 [Cryptographic Appliances and Geoff Thorpe]
7452 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7453 is called via tools/c89.sh because arguments have to be
7454 rearranged (all '-L' options must appear before the first object
7456 [Richard Shapiro <rshapiro@abinitio.com>]
7458 *) [In 0.9.6c-engine release:]
7459 Add support for Broadcom crypto accelerator cards, backported
7461 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7463 *) [In 0.9.6c-engine release:]
7464 Add support for SureWare crypto accelerator cards from
7465 Baltimore Technologies. (Use engine 'sureware')
7466 [Baltimore Technologies and Mark Cox]
7468 *) [In 0.9.6c-engine release:]
7469 Add support for crypto accelerator cards from Accelerated
7470 Encryption Processing, www.aep.ie. (Use engine 'aep')
7471 [AEP Inc. and Mark Cox]
7473 *) Add a configuration entry for gcc on UnixWare.
7474 [Gary Benson <gbenson@redhat.com>]
7476 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7477 messages are stored in a single piece (fixed-length part and
7478 variable-length part combined) and fix various bugs found on the way.
7481 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7482 instead. BIO_gethostbyname() does not know what timeouts are
7483 appropriate, so entries would stay in cache even when they have
7485 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7487 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7488 faced with a pathologically small ClientHello fragment that does
7489 not contain client_version: Instead of aborting with an error,
7490 simply choose the highest available protocol version (i.e.,
7491 TLS 1.0 unless it is disabled). In practice, ClientHello
7492 messages are never sent like this, but this change gives us
7493 strictly correct behaviour at least for TLS.
7496 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7497 never resets s->method to s->ctx->method when called from within
7498 one of the SSL handshake functions.
7499 [Bodo Moeller; problem pointed out by Niko Baric]
7501 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7502 (sent using the client's version number) if client_version is
7503 smaller than the protocol version in use. Also change
7504 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7505 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7506 the client will at least see that alert.
7509 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7513 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7514 client receives HelloRequest while in a handshake.
7515 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7517 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7518 should end in 'break', not 'goto end' which circuments various
7519 cleanups done in state SSL_ST_OK. But session related stuff
7520 must be disabled for SSL_ST_OK in the case that we just sent a
7523 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7524 before just sending a HelloRequest.
7525 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7527 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7528 reveal whether illegal block cipher padding was found or a MAC
7529 verification error occured. (Neither SSLerr() codes nor alerts
7530 are directly visible to potential attackers, but the information
7531 may leak via logfiles.)
7533 Similar changes are not required for the SSL 2.0 implementation
7534 because the number of padding bytes is sent in clear for SSL 2.0,
7535 and the extra bytes are just ignored. However ssl/s2_pkt.c
7536 failed to verify that the purported number of padding bytes is in
7540 *) Add OpenUNIX-8 support including shared libraries
7541 (Boyd Lynn Gerber <gerberb@zenez.com>).
7544 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7545 'wristwatch attack' using huge encoding parameters (cf.
7546 James H. Manger's CRYPTO 2001 paper). Note that the
7547 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7548 encoding parameters and hence was not vulnerable.
7551 *) BN_sqr() bug fix.
7552 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7554 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7555 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7556 followed by modular reduction.
7557 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7559 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7560 equivalent based on BN_pseudo_rand() instead of BN_rand().
7563 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7564 This function was broken, as the check for a new client hello message
7565 to handle SGC did not allow these large messages.
7566 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7569 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7572 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7573 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7576 *) Rework the configuration and shared library support for Tru64 Unix.
7577 The configuration part makes use of modern compiler features and
7578 still retains old compiler behavior for those that run older versions
7579 of the OS. The shared library support part includes a variant that
7580 uses the RPATH feature, and is available through the special
7581 configuration target "alpha-cc-rpath", which will never be selected
7583 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7585 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7586 with the same message size as in ssl3_get_certificate_request().
7587 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7588 messages might inadvertently be reject as too long.
7589 [Petr Lampa <lampa@fee.vutbr.cz>]
7591 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7594 *) Modified SSL library such that the verify_callback that has been set
7595 specificly for an SSL object with SSL_set_verify() is actually being
7596 used. Before the change, a verify_callback set with this function was
7597 ignored and the verify_callback() set in the SSL_CTX at the time of
7598 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7599 to allow the necessary settings.
7602 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7603 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7604 done automatically (in contradiction to the requirements of the C
7605 standard). This made problems when used from OpenSSH.
7608 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7609 dh->length and always used
7611 BN_rand_range(priv_key, dh->p).
7613 BN_rand_range() is not necessary for Diffie-Hellman, and this
7614 specific range makes Diffie-Hellman unnecessarily inefficient if
7615 dh->length (recommended exponent length) is much smaller than the
7616 length of dh->p. We could use BN_rand_range() if the order of
7617 the subgroup was stored in the DH structure, but we only have
7622 BN_rand(priv_key, l, ...)
7624 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7630 RSA_eay_public_encrypt
7631 RSA_eay_private_decrypt
7632 RSA_eay_private_encrypt (signing)
7633 RSA_eay_public_decrypt (signature verification)
7635 (default implementations for RSA_public_encrypt,
7636 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7637 always reject numbers >= n.
7640 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7641 to synchronize access to 'locking_thread'. This is necessary on
7642 systems where access to 'locking_thread' (an 'unsigned long'
7643 variable) is not atomic.
7646 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7647 *before* setting the 'crypto_lock_rand' flag. The previous code had
7648 a race condition if 0 is a valid thread ID.
7649 [Travis Vitek <vitek@roguewave.com>]
7651 *) Add support for shared libraries under Irix.
7652 [Albert Chin-A-Young <china@thewrittenword.com>]
7654 *) Add configuration option to build on Linux on both big-endian and
7656 [Ralf Baechle <ralf@uni-koblenz.de>]
7658 *) Add the possibility to create shared libraries on HP-UX.
7661 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7663 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7664 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7665 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7666 PRNG state recovery was possible based on the output of
7667 one PRNG request appropriately sized to gain knowledge on
7668 'md' followed by enough consecutive 1-byte PRNG requests
7669 to traverse all of 'state'.
7671 1. When updating 'md_local' (the current thread's copy of 'md')
7672 during PRNG output generation, hash all of the previous
7673 'md_local' value, not just the half used for PRNG output.
7675 2. Make the number of bytes from 'state' included into the hash
7676 independent from the number of PRNG bytes requested.
7678 The first measure alone would be sufficient to avoid
7679 Markku-Juhani's attack. (Actually it had never occurred
7680 to me that the half of 'md_local' used for chaining was the
7681 half from which PRNG output bytes were taken -- I had always
7682 assumed that the secret half would be used.) The second
7683 measure makes sure that additional data from 'state' is never
7684 mixed into 'md_local' in small portions; this heuristically
7685 further strengthens the PRNG.
7688 *) Fix crypto/bn/asm/mips3.s.
7691 *) When only the key is given to "enc", the IV is undefined. Print out
7692 an error message in this case.
7695 *) Handle special case when X509_NAME is empty in X509 printing routines.
7698 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7699 positive and less than q.
7702 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7703 used: it isn't thread safe and the add_lock_callback should handle
7705 [Paul Rose <Paul.Rose@bridge.com>]
7707 *) Verify that incoming data obeys the block size in
7708 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7712 [Ulf Möller, Bodo Möller]
7714 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7715 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7716 when fixing the server behaviour for backwards-compatible 'client
7717 hello' messages. (Note that the attack is impractical against
7718 SSL 3.0 and TLS 1.0 anyway because length and version checking
7719 means that the probability of guessing a valid ciphertext is
7720 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7723 Before 0.9.5, the countermeasure (hide the error by generating a
7724 random 'decryption result') did not work properly because
7725 ERR_clear_error() was missing, meaning that SSL_get_error() would
7726 detect the supposedly ignored error.
7728 Both problems are now fixed.
7731 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7732 (previously it was 1024).
7735 *) Fix for compatibility mode trust settings: ignore trust settings
7736 unless some valid trust or reject settings are present.
7739 *) Fix for blowfish EVP: its a variable length cipher.
7742 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7743 parameters in DSA public key structures and return an error in the
7744 DSA routines if parameters are absent.
7747 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7748 in the current directory if neither $RANDFILE nor $HOME was set.
7749 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7750 caused some confusion to Windows users who haven't defined $HOME.
7751 Thus RAND_file_name() is changed again: e_os.h can define a
7752 DEFAULT_HOME, which will be used if $HOME is not set.
7753 For Windows, we use "C:"; on other platforms, we still require
7754 environment variables.
7756 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7757 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7758 having multiple threads call RAND_poll() concurrently.
7761 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7762 combination of a flag and a thread ID variable.
7763 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7764 flag), *other* threads can enter ssleay_add_bytes without obeying
7765 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7766 that they do not hold after the first thread unsets add_do_not_lock).
7769 *) Change bctest again: '-x' expressions are not available in all
7773 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7775 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7776 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7778 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7779 the default extension for executables, if any. Also, make the perl
7780 scripts that use symlink() to test if it really exists and use "cp"
7781 if it doesn't. All this made OpenSSL compilable and installable in
7785 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7786 If SEQUENCE is length is indefinite just set c->slen to the total
7787 amount of data available.
7788 [Steve Henson, reported by shige@FreeBSD.org]
7789 [This change does not apply to 0.9.7.]
7791 *) Change bctest to avoid here-documents inside command substitution
7792 (workaround for FreeBSD /bin/sh bug).
7793 For compatibility with Ultrix, avoid shell functions (introduced
7794 in the bctest version that searches along $PATH).
7797 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7798 with des_encrypt() defined on some operating systems, like Solaris
7802 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7803 On the Importance of Eliminating Errors in Cryptographic
7804 Computations, J. Cryptology 14 (2001) 2, 101-119,
7805 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7808 *) MIPS assembler BIGNUM division bug fix.
7811 *) Disabled incorrect Alpha assembler code.
7814 *) Fix PKCS#7 decode routines so they correctly update the length
7815 after reading an EOC for the EXPLICIT tag.
7817 [This change does not apply to 0.9.7.]
7819 *) Fix bug in PKCS#12 key generation routines. This was triggered
7820 if a 3DES key was generated with a 0 initial byte. Include
7821 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7822 (but broken) behaviour.
7825 *) Enhance bctest to search for a working bc along $PATH and print
7827 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7829 *) Fix memory leaks in err.c: free err_data string if necessary;
7830 don't write to the wrong index in ERR_set_error_data.
7833 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7837 *) Replace rdtsc with _emit statements for VC++ version 5.
7838 [Jeremy Cooper <jeremy@baymoo.org>]
7840 *) Make it possible to reuse SSLv2 sessions.
7843 *) In copy_email() check for >= 0 as a return value for
7844 X509_NAME_get_index_by_NID() since 0 is a valid index.
7845 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7847 *) Avoid coredump with unsupported or invalid public keys by checking if
7848 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7849 PKCS7_verify() fails with non detached data.
7852 *) Don't use getenv in library functions when run as setuid/setgid.
7853 New function OPENSSL_issetugid().
7856 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7857 due to incorrect handling of multi-threading:
7859 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7861 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7863 3. Count how many times MemCheck_off() has been called so that
7864 nested use can be treated correctly. This also avoids
7865 inband-signalling in the previous code (which relied on the
7866 assumption that thread ID 0 is impossible).
7869 *) Add "-rand" option also to s_client and s_server.
7872 *) Fix CPU detection on Irix 6.x.
7873 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7874 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7876 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7879 [This change does not apply to 0.9.7.]
7881 *) Use the cached encoding of an X509_NAME structure rather than
7882 copying it. This is apparently the reason for the libsafe "errors"
7883 but the code is actually correct.
7886 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7887 Bleichenbacher's DSA attack.
7888 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7889 to be set and top=0 forces the highest bit to be set; top=-1 is new
7890 and leaves the highest bit random.
7891 [Ulf Moeller, Bodo Moeller]
7893 *) In the NCONF_...-based implementations for CONF_... queries
7894 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7895 a temporary CONF structure with the data component set to NULL
7896 (which gives segmentation faults in lh_retrieve).
7897 Instead, use NULL for the CONF pointer in CONF_get_string and
7898 CONF_get_number (which may use environment variables) and directly
7899 return NULL from CONF_get_section.
7902 *) Fix potential buffer overrun for EBCDIC.
7905 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7906 keyUsage if basicConstraints absent for a CA.
7909 *) Make SMIME_write_PKCS7() write mail header values with a format that
7910 is more generally accepted (no spaces before the semicolon), since
7911 some programs can't parse those values properly otherwise. Also make
7912 sure BIO's that break lines after each write do not create invalid
7916 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7917 macros previously used would not encode an empty SEQUENCE OF
7918 and break the signature.
7920 [This change does not apply to 0.9.7.]
7922 *) Zero the premaster secret after deriving the master secret in
7926 *) Add some EVP_add_digest_alias registrations (as found in
7927 OpenSSL_add_all_digests()) to SSL_library_init()
7928 aka OpenSSL_add_ssl_algorithms(). This provides improved
7929 compatibility with peers using X.509 certificates
7930 with unconventional AlgorithmIdentifier OIDs.
7933 *) Fix for Irix with NO_ASM.
7934 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7936 *) ./config script fixes.
7937 [Ulf Moeller, Richard Levitte]
7939 *) Fix 'openssl passwd -1'.
7942 *) Change PKCS12_key_gen_asc() so it can cope with non null
7943 terminated strings whose length is passed in the passlen
7944 parameter, for example from PEM callbacks. This was done
7945 by adding an extra length parameter to asc2uni().
7946 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7948 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7949 call failed, free the DSA structure.
7952 *) Fix to uni2asc() to cope with zero length Unicode strings.
7953 These are present in some PKCS#12 files.
7956 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7957 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7958 when writing a 32767 byte record.
7959 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7961 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7962 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7964 (RSA objects have a reference count access to which is protected
7965 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7966 so they are meant to be shared between threads.)
7967 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7968 "Reddie, Steven" <Steven.Reddie@ca.com>]
7970 *) Fix a deadlock in CRYPTO_mem_leaks().
7973 *) Use better test patterns in bntest.
7976 *) rand_win.c fix for Borland C.
7979 *) BN_rshift bugfix for n == 0.
7982 *) Add a 'bctest' script that checks for some known 'bc' bugs
7983 so that 'make test' does not abort just because 'bc' is broken.
7986 *) Store verify_result within SSL_SESSION also for client side to
7987 avoid potential security hole. (Re-used sessions on the client side
7988 always resulted in verify_result==X509_V_OK, not using the original
7989 result of the server certificate verification.)
7992 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7993 SSL3_RT_APPLICATION_DATA, return 0.
7994 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7998 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7999 releases, have been re-implemented by renaming the previous
8000 implementations of ssl2_read and ssl3_read to ssl2_read_internal
8001 and ssl3_read_internal, respectively, and adding 'peek' parameters
8002 to them. The new ssl[23]_{read,peek} functions are calls to
8003 ssl[23]_read_internal with the 'peek' flag set appropriately.
8004 A 'peek' parameter has also been added to ssl3_read_bytes, which
8005 does the actual work for ssl3_read_internal.
8008 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
8009 the method-specific "init()" handler. Also clean up ex_data after
8010 calling the method-specific "finish()" handler. Previously, this was
8011 happening the other way round.
8014 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
8015 The previous value, 12, was not always sufficient for BN_mod_exp().
8018 *) Make sure that shared libraries get the internal name engine with
8019 the full version number and not just 0. This should mark the
8020 shared libraries as not backward compatible. Of course, this should
8021 be changed again when we can guarantee backward binary compatibility.
8024 *) Fix typo in get_cert_by_subject() in by_dir.c
8025 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8027 *) Rework the system to generate shared libraries:
8029 - Make note of the expected extension for the shared libraries and
8030 if there is a need for symbolic links from for example libcrypto.so.0
8031 to libcrypto.so.0.9.7. There is extended info in Configure for
8034 - Make as few rebuilds of the shared libraries as possible.
8036 - Still avoid linking the OpenSSL programs with the shared libraries.
8038 - When installing, install the shared libraries separately from the
8042 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8044 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8045 and not in SSL_clear because the latter is also used by the
8046 accept/connect functions; previously, the settings made by
8047 SSL_set_read_ahead would be lost during the handshake.
8048 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8050 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8051 Previously, it would create entries for disableed algorithms no
8055 *) Added several new manual pages for SSL_* function.
8058 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8060 *) In ssl23_get_client_hello, generate an error message when faced
8061 with an initial SSL 3.0/TLS record that is too small to contain the
8062 first two bytes of the ClientHello message, i.e. client_version.
8063 (Note that this is a pathologic case that probably has never happened
8064 in real life.) The previous approach was to use the version number
8065 from the record header as a substitute; but our protocol choice
8066 should not depend on that one because it is not authenticated
8067 by the Finished messages.
8070 *) More robust randomness gathering functions for Windows.
8071 [Jeffrey Altman <jaltman@columbia.edu>]
8073 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8074 not set then we don't setup the error code for issuer check errors
8075 to avoid possibly overwriting other errors which the callback does
8076 handle. If an application does set the flag then we assume it knows
8077 what it is doing and can handle the new informational codes
8081 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8082 a general "ANY" type, as such it should be able to decode anything
8083 including tagged types. However it didn't check the class so it would
8084 wrongly interpret tagged types in the same way as their universal
8085 counterpart and unknown types were just rejected. Changed so that the
8086 tagged and unknown types are handled in the same way as a SEQUENCE:
8087 that is the encoding is stored intact. There is also a new type
8088 "V_ASN1_OTHER" which is used when the class is not universal, in this
8089 case we have no idea what the actual type is so we just lump them all
8093 *) On VMS, stdout may very well lead to a file that is written to
8094 in a record-oriented fashion. That means that every write() will
8095 write a separate record, which will be read separately by the
8096 programs trying to read from it. This can be very confusing.
8098 The solution is to put a BIO filter in the way that will buffer
8099 text until a linefeed is reached, and then write everything a
8100 line at a time, so every record written will be an actual line,
8101 not chunks of lines and not (usually doesn't happen, but I've
8102 seen it once) several lines in one record. BIO_f_linebuffer() is
8105 Currently, it's a VMS-only method, because that's where it has
8106 been tested well enough.
8109 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8110 it can return incorrect results.
8111 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8112 but it was in 0.9.6-beta[12].)
8115 *) Disable the check for content being present when verifying detached
8116 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8117 include zero length content when signing messages.
8120 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8121 BIO_ctrl (for BIO pairs).
8124 *) Add DSO method for VMS.
8127 *) Bug fix: Montgomery multiplication could produce results with the
8131 *) Add RPM specification openssl.spec and modify it to build three
8132 packages. The default package contains applications, application
8133 documentation and run-time libraries. The devel package contains
8134 include files, static libraries and function documentation. The
8135 doc package contains the contents of the doc directory. The original
8136 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8139 *) Add a large number of documentation files for many SSL routines.
8140 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8142 *) Add a configuration entry for Sony News 4.
8143 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8145 *) Don't set the two most significant bits to one when generating a
8146 random number < q in the DSA library.
8149 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8150 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8151 the underlying transport is blocking) if a handshake took place.
8152 (The default behaviour is needed by applications such as s_client
8153 and s_server that use select() to determine when to use SSL_read;
8154 but for applications that know in advance when to expect data, it
8155 just makes things more complicated.)
8158 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8162 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8163 work better on such systems.
8164 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8166 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8167 Update PKCS12_parse() so it copies the friendlyName and the
8168 keyid to the certificates aux info.
8171 *) Fix bug in PKCS7_verify() which caused an infinite loop
8172 if there was more than one signature.
8173 [Sven Uszpelkat <su@celocom.de>]
8175 *) Major change in util/mkdef.pl to include extra information
8176 about each symbol, as well as presentig variables as well
8177 as functions. This change means that there's n more need
8178 to rebuild the .num files when some algorithms are excluded.
8181 *) Allow the verify time to be set by an application,
8182 rather than always using the current time.
8185 *) Phase 2 verify code reorganisation. The certificate
8186 verify code now looks up an issuer certificate by a
8187 number of criteria: subject name, authority key id
8188 and key usage. It also verifies self signed certificates
8189 by the same criteria. The main comparison function is
8190 X509_check_issued() which performs these checks.
8192 Lot of changes were necessary in order to support this
8193 without completely rewriting the lookup code.
8195 Authority and subject key identifier are now cached.
8197 The LHASH 'certs' is X509_STORE has now been replaced
8198 by a STACK_OF(X509_OBJECT). This is mainly because an
8199 LHASH can't store or retrieve multiple objects with
8200 the same hash value.
8202 As a result various functions (which were all internal
8203 use only) have changed to handle the new X509_STORE
8204 structure. This will break anything that messed round
8205 with X509_STORE internally.
8207 The functions X509_STORE_add_cert() now checks for an
8208 exact match, rather than just subject name.
8210 The X509_STORE API doesn't directly support the retrieval
8211 of multiple certificates matching a given criteria, however
8212 this can be worked round by performing a lookup first
8213 (which will fill the cache with candidate certificates)
8214 and then examining the cache for matches. This is probably
8215 the best we can do without throwing out X509_LOOKUP
8216 entirely (maybe later...).
8218 The X509_VERIFY_CTX structure has been enhanced considerably.
8220 All certificate lookup operations now go via a get_issuer()
8221 callback. Although this currently uses an X509_STORE it
8222 can be replaced by custom lookups. This is a simple way
8223 to bypass the X509_STORE hackery necessary to make this
8224 work and makes it possible to use more efficient techniques
8225 in future. A very simple version which uses a simple
8226 STACK for its trusted certificate store is also provided
8227 using X509_STORE_CTX_trusted_stack().
8229 The verify_cb() and verify() callbacks now have equivalents
8230 in the X509_STORE_CTX structure.
8232 X509_STORE_CTX also has a 'flags' field which can be used
8233 to customise the verify behaviour.
8236 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8237 excludes S/MIME capabilities.
8240 *) When a certificate request is read in keep a copy of the
8241 original encoding of the signed data and use it when outputing
8242 again. Signatures then use the original encoding rather than
8243 a decoded, encoded version which may cause problems if the
8244 request is improperly encoded.
8247 *) For consistency with other BIO_puts implementations, call
8248 buffer_write(b, ...) directly in buffer_puts instead of calling
8251 In BIO_puts, increment b->num_write as in BIO_write.
8252 [Peter.Sylvester@EdelWeb.fr]
8254 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8255 BN_zero, we may not return a BIGNUM with an array consisting of
8259 *) Avoid calling abort() from within the library when problems are
8260 detected, except if preprocessor symbols have been defined
8261 (such as REF_CHECK, BN_DEBUG etc.).
8264 *) New openssl application 'rsautl'. This utility can be
8265 used for low level RSA operations. DER public key
8266 BIO/fp routines also added.
8269 *) New Configure entry and patches for compiling on QNX 4.
8270 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8272 *) A demo state-machine implementation was sponsored by
8273 Nuron (http://www.nuron.com/) and is now available in
8274 demos/state_machine.
8277 *) New options added to the 'dgst' utility for signature
8278 generation and verification.
8281 *) Unrecognized PKCS#7 content types are now handled via a
8282 catch all ASN1_TYPE structure. This allows unsupported
8283 types to be stored as a "blob" and an application can
8284 encode and decode it manually.
8287 *) Fix various signed/unsigned issues to make a_strex.c
8289 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8291 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8292 length if passed a buffer. ASN1_INTEGER_to_BN failed
8293 if passed a NULL BN and its argument was negative.
8294 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8296 *) Modification to PKCS#7 encoding routines to output definite
8297 length encoding. Since currently the whole structures are in
8298 memory there's not real point in using indefinite length
8299 constructed encoding. However if OpenSSL is compiled with
8300 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8303 *) Added BIO_vprintf() and BIO_vsnprintf().
8306 *) Added more prefixes to parse for in the the strings written
8307 through a logging bio, to cover all the levels that are available
8308 through syslog. The prefixes are now:
8310 PANIC, EMERG, EMR => LOG_EMERG
8311 ALERT, ALR => LOG_ALERT
8312 CRIT, CRI => LOG_CRIT
8313 ERROR, ERR => LOG_ERR
8314 WARNING, WARN, WAR => LOG_WARNING
8315 NOTICE, NOTE, NOT => LOG_NOTICE
8316 INFO, INF => LOG_INFO
8317 DEBUG, DBG => LOG_DEBUG
8319 and as before, if none of those prefixes are present at the
8320 beginning of the string, LOG_ERR is chosen.
8322 On Win32, the LOG_* levels are mapped according to this:
8324 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8325 LOG_WARNING => EVENTLOG_WARNING_TYPE
8326 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8330 *) Made it possible to reconfigure with just the configuration
8331 argument "reconf" or "reconfigure". The command line arguments
8332 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8333 and are retrieved from there when reconfiguring.
8337 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8339 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8342 *) The obj_dat.pl script was messing up the sorting of object
8343 names. The reason was that it compared the quoted version
8344 of strings as a result "OCSP" > "OCSP Signing" because
8345 " > SPACE. Changed script to store unquoted versions of
8346 names and add quotes on output. It was also omitting some
8347 names from the lookup table if they were given a default
8348 value (that is if SN is missing it is given the same
8349 value as LN and vice versa), these are now added on the
8350 grounds that if an object has a name we should be able to
8351 look it up. Finally added warning output when duplicate
8352 short or long names are found.
8355 *) Changes needed for Tandem NSK.
8356 [Scott Uroff <scott@xypro.com>]
8358 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8359 RSA_padding_check_SSLv23(), special padding was never detected
8360 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8361 version rollback attacks was not effective.
8363 In s23_clnt.c, don't use special rollback-attack detection padding
8364 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8365 client; similarly, in s23_srvr.c, don't do the rollback check if
8366 SSL 2.0 is the only protocol enabled in the server.
8369 *) Make it possible to get hexdumps of unprintable data with 'openssl
8370 asn1parse'. By implication, the functions ASN1_parse_dump() and
8371 BIO_dump_indent() are added.
8374 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8375 these print out strings and name structures based on various
8376 flags including RFC2253 support and proper handling of
8377 multibyte characters. Added options to the 'x509' utility
8378 to allow the various flags to be set.
8381 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8382 Also change the functions X509_cmp_current_time() and
8383 X509_gmtime_adj() work with an ASN1_TIME structure,
8384 this will enable certificates using GeneralizedTime in validity
8385 dates to be checked.
8388 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8389 negative public key encodings) on by default,
8390 NO_NEG_PUBKEY_BUG can be set to disable it.
8393 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8394 content octets. An i2c_ASN1_OBJECT is unnecessary because
8395 the encoding can be trivially obtained from the structure.
8398 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8399 not read locks (CRYPTO_r_[un]lock).
8402 *) A first attempt at creating official support for shared
8403 libraries through configuration. I've kept it so the
8404 default is static libraries only, and the OpenSSL programs
8405 are always statically linked for now, but there are
8406 preparations for dynamic linking in place.
8407 This has been tested on Linux and Tru64.
8410 *) Randomness polling function for Win9x, as described in:
8411 Peter Gutmann, Software Generation of Practically Strong
8415 *) Fix so PRNG is seeded in req if using an already existing
8419 *) New options to smime application. -inform and -outform
8420 allow alternative formats for the S/MIME message including
8421 PEM and DER. The -content option allows the content to be
8422 specified separately. This should allow things like Netscape
8423 form signing output easier to verify.
8426 *) Fix the ASN1 encoding of tags using the 'long form'.
8429 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8430 STRING types. These convert content octets to and from the
8431 underlying type. The actual tag and length octets are
8432 already assumed to have been read in and checked. These
8433 are needed because all other string types have virtually
8434 identical handling apart from the tag. By having versions
8435 of the ASN1 functions that just operate on content octets
8436 IMPLICIT tagging can be handled properly. It also allows
8437 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8438 and ASN1_INTEGER are identical apart from the tag.
8441 *) Change the handling of OID objects as follows:
8443 - New object identifiers are inserted in objects.txt, following
8444 the syntax given in objects.README.
8445 - objects.pl is used to process obj_mac.num and create a new
8447 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8450 This is currently kind of a hack, and the perl code in objects.pl
8451 isn't very elegant, but it works as I intended. The simplest way
8452 to check that it worked correctly is to look in obj_dat.h and
8453 check the array nid_objs and make sure the objects haven't moved
8454 around (this is important!). Additions are OK, as well as
8455 consistent name changes.
8458 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8461 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8462 The given file adds to whatever has already been seeded into the
8463 random pool through the RANDFILE configuration file option or
8464 environment variable, or the default random state file.
8467 *) mkstack.pl now sorts each macro group into lexical order.
8468 Previously the output order depended on the order the files
8469 appeared in the directory, resulting in needless rewriting
8473 *) Patches to make OpenSSL compile under Win32 again. Mostly
8474 work arounds for the VC++ problem that it treats func() as
8475 func(void). Also stripped out the parts of mkdef.pl that
8476 added extra typesafe functions: these no longer exist.
8479 *) Reorganisation of the stack code. The macros are now all
8480 collected in safestack.h . Each macro is defined in terms of
8481 a "stack macro" of the form SKM_<name>(type, a, b). The
8482 DEBUG_SAFESTACK is now handled in terms of function casts,
8483 this has the advantage of retaining type safety without the
8484 use of additional functions. If DEBUG_SAFESTACK is not defined
8485 then the non typesafe macros are used instead. Also modified the
8486 mkstack.pl script to handle the new form. Needs testing to see
8487 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8488 the default if no major problems. Similar behaviour for ASN1_SET_OF
8489 and PKCS12_STACK_OF.
8492 *) When some versions of IIS use the 'NET' form of private key the
8493 key derivation algorithm is different. Normally MD5(password) is
8494 used as a 128 bit RC4 key. In the modified case
8495 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
8496 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8497 as the old Netscape_RSA functions except they have an additional
8498 'sgckey' parameter which uses the modified algorithm. Also added
8499 an -sgckey command line option to the rsa utility. Thanks to
8500 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8501 algorithm to openssl-dev.
8504 *) The evp_local.h macros were using 'c.##kname' which resulted in
8505 invalid expansion on some systems (SCO 5.0.5 for example).
8506 Corrected to 'c.kname'.
8507 [Phillip Porch <root@theporch.com>]
8509 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8510 a STACK of email addresses from a certificate or request, these look
8511 in the subject name and the subject alternative name extensions and
8512 omit any duplicate addresses.
8515 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8516 This makes DSA verification about 2 % faster.
8519 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8520 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8521 plus overhead for 1024 bit moduli).
8522 This makes exponentiations about 0.5 % faster for 1024 bit
8523 exponents (as measured by "openssl speed rsa2048").
8526 *) Rename memory handling macros to avoid conflicts with other
8528 Malloc => OPENSSL_malloc
8529 Malloc_locked => OPENSSL_malloc_locked
8530 Realloc => OPENSSL_realloc
8531 Free => OPENSSL_free
8534 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8535 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8538 *) CygWin32 support.
8539 [John Jarvie <jjarvie@newsguy.com>]
8541 *) The type-safe stack code has been rejigged. It is now only compiled
8542 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8543 by default all type-specific stack functions are "#define"d back to
8544 standard stack functions. This results in more streamlined output
8545 but retains the type-safety checking possibilities of the original
8549 *) The STACK code has been cleaned up, and certain type declarations
8550 that didn't make a lot of sense have been brought in line. This has
8551 also involved a cleanup of sorts in safestack.h to more correctly
8552 map type-safe stack functions onto their plain stack counterparts.
8553 This work has also resulted in a variety of "const"ifications of
8554 lots of the code, especially "_cmp" operations which should normally
8555 be prototyped with "const" parameters anyway.
8558 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8559 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8560 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8561 where all of 'md' is used each time the PRNG is used, but 'state'
8562 is used only indexed by a cyclic counter. As entropy may not be
8563 well distributed from the beginning, 'md' is important as a
8564 chaining variable. However, the output function chains only half
8565 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8566 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8567 in all of 'state' being rewritten, with the new values depending
8568 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8571 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8572 the handshake is continued after ssl_verify_cert_chain();
8573 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8574 can lead to 'unexplainable' connection aborts later.
8575 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8577 *) Major EVP API cipher revision.
8578 Add hooks for extra EVP features. This allows various cipher
8579 parameters to be set in the EVP interface. Support added for variable
8580 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8581 setting of RC2 and RC5 parameters.
8583 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8586 Remove lots of duplicated code from the EVP library. For example *every*
8587 cipher init() function handles the 'iv' in the same way according to the
8588 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8589 for CFB and OFB modes they zero ctx->num.
8591 New functionality allows removal of S/MIME code RC2 hack.
8593 Most of the routines have the same form and so can be declared in terms
8596 By shifting this to the top level EVP_CipherInit() it can be removed from
8597 all individual ciphers. If the cipher wants to handle IVs or keys
8598 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8601 Change lots of functions like EVP_EncryptUpdate() to now return a
8602 value: although software versions of the algorithms cannot fail
8603 any installed hardware versions can.
8606 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8607 this option is set, tolerate broken clients that send the negotiated
8608 protocol version number instead of the requested protocol version
8612 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8613 i.e. non-zero for export ciphersuites, zero otherwise.
8614 Previous versions had this flag inverted, inconsistent with
8615 rsa_tmp_cb (..._TMP_RSA_CB).
8616 [Bodo Moeller; problem reported by Amit Chopra]
8618 *) Add missing DSA library text string. Work around for some IIS
8619 key files with invalid SEQUENCE encoding.
8622 *) Add a document (doc/standards.txt) that list all kinds of standards
8623 and so on that are implemented in OpenSSL.
8626 *) Enhance c_rehash script. Old version would mishandle certificates
8627 with the same subject name hash and wouldn't handle CRLs at all.
8628 Added -fingerprint option to crl utility, to support new c_rehash
8632 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8635 *) Fix for SSL server purpose checking. Server checking was
8636 rejecting certificates which had extended key usage present
8637 but no ssl client purpose.
8638 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8640 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8641 is a little unclear about how a blank password is handled.
8642 Since the password in encoded as a BMPString with terminating
8643 double NULL a zero length password would end up as just the
8644 double NULL. However no password at all is different and is
8645 handled differently in the PKCS#12 key generation code. NS
8646 treats a blank password as zero length. MSIE treats it as no
8647 password on export: but it will try both on import. We now do
8648 the same: PKCS12_parse() tries zero length and no password if
8649 the password is set to "" or NULL (NULL is now a valid password:
8650 it wasn't before) as does the pkcs12 application.
8653 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8654 perror when PEM_read_bio_X509_REQ fails, the error message must
8655 be obtained from the error queue.
8658 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8659 it in ERR_remove_state if appropriate, and change ERR_get_state
8660 accordingly to avoid race conditions (this is necessary because
8661 thread_hash is no longer constant once set).
8664 *) Bugfix for linux-elf makefile.one.
8667 *) RSA_get_default_method() will now cause a default
8668 RSA_METHOD to be chosen if one doesn't exist already.
8669 Previously this was only set during a call to RSA_new()
8670 or RSA_new_method(NULL) meaning it was possible for
8671 RSA_get_default_method() to return NULL.
8674 *) Added native name translation to the existing DSO code
8675 that will convert (if the flag to do so is set) filenames
8676 that are sufficiently small and have no path information
8677 into a canonical native form. Eg. "blah" converted to
8678 "libblah.so" or "blah.dll" etc.
8681 *) New function ERR_error_string_n(e, buf, len) which is like
8682 ERR_error_string(e, buf), but writes at most 'len' bytes
8683 including the 0 terminator. For ERR_error_string_n, 'buf'
8685 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8687 *) CONF library reworked to become more general. A new CONF
8688 configuration file reader "class" is implemented as well as a
8689 new functions (NCONF_*, for "New CONF") to handle it. The now
8690 old CONF_* functions are still there, but are reimplemented to
8691 work in terms of the new functions. Also, a set of functions
8692 to handle the internal storage of the configuration data is
8693 provided to make it easier to write new configuration file
8694 reader "classes" (I can definitely see something reading a
8695 configuration file in XML format, for example), called _CONF_*,
8696 or "the configuration storage API"...
8698 The new configuration file reading functions are:
8700 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8701 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8703 NCONF_default, NCONF_WIN32
8705 NCONF_dump_fp, NCONF_dump_bio
8707 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8708 NCONF_new creates a new CONF object. This works in the same way
8709 as other interfaces in OpenSSL, like the BIO interface.
8710 NCONF_dump_* dump the internal storage of the configuration file,
8711 which is useful for debugging. All other functions take the same
8712 arguments as the old CONF_* functions wth the exception of the
8713 first that must be a `CONF *' instead of a `LHASH *'.
8715 To make it easer to use the new classes with the old CONF_* functions,
8716 the function CONF_set_default_method is provided.
8719 *) Add '-tls1' option to 'openssl ciphers', which was already
8720 mentioned in the documentation but had not been implemented.
8721 (This option is not yet really useful because even the additional
8722 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8725 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8726 OpenSSL-based applications) load shared libraries and bind to
8727 them in a portable way.
8728 [Geoff Thorpe, with contributions from Richard Levitte]
8730 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8732 *) Make sure _lrotl and _lrotr are only used with MSVC.
8734 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8735 (the default implementation of RAND_status).
8737 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8738 to '-clrext' (= clear extensions), as intended and documented.
8739 [Bodo Moeller; inconsistency pointed out by Michael Attili
8742 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8743 was larger than the MD block size.
8744 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8746 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8747 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8748 using the passed key: if the passed key was a private key the result
8749 of X509_print(), for example, would be to print out all the private key
8753 *) des_quad_cksum() byte order bug fix.
8754 [Ulf Möller, using the problem description in krb4-0.9.7, where
8755 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8757 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8759 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8761 *) For easily testing in shell scripts whether some command
8762 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8763 returns with exit code 0 iff no command of the given name is available.
8764 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8765 the output goes to stdout and nothing is printed to stderr.
8766 Additional arguments are always ignored.
8768 Since for each cipher there is a command of the same name,
8769 the 'no-cipher' compilation switches can be tested this way.
8771 ('openssl no-XXX' is not able to detect pseudo-commands such
8772 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8775 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8778 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8779 is set; it will be thrown away anyway because each handshake creates
8781 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8782 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8783 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
8784 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8787 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8788 'Q' and 'R' lose their special meanings (quit/renegotiate).
8789 This is part of what -quiet does; unlike -quiet, -ign_eof
8790 does not suppress any output.
8793 *) Add compatibility options to the purpose and trust code. The
8794 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8795 accepts a certificate or CA, this was the previous behaviour,
8796 with all the associated security issues.
8798 X509_TRUST_COMPAT is the old trust behaviour: only and
8799 automatically trust self signed roots in certificate store. A
8800 new trust setting X509_TRUST_DEFAULT is used to specify that
8801 a purpose has no associated trust setting and it should instead
8802 use the value in the default purpose.
8805 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8806 and fix a memory leak.
8809 *) In util/mkerr.pl (which implements 'make errors'), preserve
8810 reason strings from the previous version of the .c file, as
8811 the default to have only downcase letters (and digits) in
8812 automatically generated reasons codes is not always appropriate.
8815 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8816 using strerror. Previously, ERR_reason_error_string() returned
8817 library names as reason strings for SYSerr; but SYSerr is a special
8818 case where small numbers are errno values, not library numbers.
8821 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8822 converts DSA parameters into DH parameters. (When creating parameters,
8823 DSA_generate_parameters is used.)
8826 *) Include 'length' (recommended exponent length) in C code generated
8827 by 'openssl dhparam -C'.
8830 *) The second argument to set_label in perlasm was already being used
8831 so couldn't be used as a "file scope" flag. Moved to third argument
8835 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8836 instead of RAND_bytes for encryption IVs and salts.
8839 *) Include RAND_status() into RAND_METHOD instead of implementing
8840 it only for md_rand.c Otherwise replacing the PRNG by calling
8841 RAND_set_rand_method would be impossible.
8844 *) Don't let DSA_generate_key() enter an infinite loop if the random
8845 number generation fails.
8848 *) New 'rand' application for creating pseudo-random output.
8851 *) Added configuration support for Linux/IA64
8852 [Rolf Haberrecker <rolf@suse.de>]
8854 *) Assembler module support for Mingw32.
8857 *) Shared library support for HPUX (in shlib/).
8858 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8860 *) Shared library support for Solaris gcc.
8861 [Lutz Behnke <behnke@trustcenter.de>]
8863 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8865 *) PKCS7_encrypt() was adding text MIME headers twice because they
8866 were added manually and by SMIME_crlf_copy().
8869 *) In bntest.c don't call BN_rand with zero bits argument.
8870 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8872 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8873 case was implemented. This caused BN_div_recp() to fail occasionally.
8876 *) Add an optional second argument to the set_label() in the perl
8877 assembly language builder. If this argument exists and is set
8878 to 1 it signals that the assembler should use a symbol whose
8879 scope is the entire file, not just the current function. This
8880 is needed with MASM which uses the format label:: for this scope.
8881 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8883 *) Change the ASN1 types so they are typedefs by default. Before
8884 almost all types were #define'd to ASN1_STRING which was causing
8885 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8889 *) Change names of new functions to the new get1/get0 naming
8890 convention: After 'get1', the caller owns a reference count
8891 and has to call ..._free; 'get0' returns a pointer to some
8892 data structure without incrementing reference counters.
8893 (Some of the existing 'get' functions increment a reference
8894 counter, some don't.)
8895 Similarly, 'set1' and 'add1' functions increase reference
8896 counters or duplicate objects.
8899 *) Allow for the possibility of temp RSA key generation failure:
8900 the code used to assume it always worked and crashed on failure.
8903 *) Fix potential buffer overrun problem in BIO_printf().
8904 [Ulf Möller, using public domain code by Patrick Powell; problem
8905 pointed out by David Sacerdote <das33@cornell.edu>]
8907 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8908 RAND_egd() and RAND_status(). In the command line application,
8909 the EGD socket can be specified like a seed file using RANDFILE
8913 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8914 Some CAs (e.g. Verisign) distribute certificates in this form.
8917 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8918 list to exclude them. This means that no special compilation option
8919 is needed to use anonymous DH: it just needs to be included in the
8923 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8924 EVP_MD_type. The old functionality is available in a new macro called
8925 EVP_MD_md(). Change code that uses it and update docs.
8928 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8929 where the 'void *' argument is replaced by a function pointer argument.
8930 Previously 'void *' was abused to point to functions, which works on
8931 many platforms, but is not correct. As these functions are usually
8932 called by macros defined in OpenSSL header files, most source code
8933 should work without changes.
8936 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8937 sections with information on -D... compiler switches used for
8938 compiling the library so that applications can see them. To enable
8939 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8940 must be defined. E.g.,
8941 #define OPENSSL_ALGORITHM_DEFINES
8942 #include <openssl/opensslconf.h>
8943 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8944 [Richard Levitte, Ulf and Bodo Möller]
8946 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8950 *) Change the 'other' type in certificate aux info to a STACK_OF
8951 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8952 the required ASN1 format: arbitrary types determined by an OID.
8955 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8956 argument to 'req'. This is not because the function is newer or
8957 better than others it just uses the work 'NEW' in the certificate
8958 request header lines. Some software needs this.
8961 *) Reorganise password command line arguments: now passwords can be
8962 obtained from various sources. Delete the PEM_cb function and make
8963 it the default behaviour: i.e. if the callback is NULL and the
8964 usrdata argument is not NULL interpret it as a null terminated pass
8965 phrase. If usrdata and the callback are NULL then the pass phrase
8966 is prompted for as usual.
8969 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8970 the support is automatically enabled. The resulting binaries will
8971 autodetect the card and use it if present.
8972 [Ben Laurie and Compaq Inc.]
8974 *) Work around for Netscape hang bug. This sends certificate request
8975 and server done in one record. Since this is perfectly legal in the
8976 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8977 the bugs/SSLv3 entry for more info.
8980 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8983 *) Add -rand argument to smime and pkcs12 applications and read/write
8987 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8990 *) Add command line password options to the remaining applications.
8993 *) Bug fix for BN_div_recp() for numerators with an even number of
8997 *) More tests in bntest.c, and changed test_bn output.
9000 *) ./config recognizes MacOS X now.
9003 *) Bug fix for BN_div() when the first words of num and divsor are
9004 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
9007 *) Add support for various broken PKCS#8 formats, and command line
9008 options to produce them.
9011 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
9012 get temporary BIGNUMs from a BN_CTX.
9015 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
9019 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
9020 include a #define from the old name to the new. The original intent
9021 was that statically linked binaries could for example just call
9022 SSLeay_add_all_ciphers() to just add ciphers to the table and not
9023 link with digests. This never worked becayse SSLeay_add_all_digests()
9024 and SSLeay_add_all_ciphers() were in the same source file so calling
9025 one would link with the other. They are now in separate source files.
9028 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9031 *) Use a less unusual form of the Miller-Rabin primality test (it used
9032 a binary algorithm for exponentiation integrated into the Miller-Rabin
9033 loop, our standard modexp algorithms are faster).
9036 *) Support for the EBCDIC character set completed.
9037 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9039 *) Source code cleanups: use const where appropriate, eliminate casts,
9040 use void * instead of char * in lhash.
9043 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9044 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9045 this the server could overwrite ephemeral keys that the client
9049 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9050 using 50 iterations of the Rabin-Miller test.
9052 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9053 iterations of the Rabin-Miller test as required by the appendix
9054 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9055 As BN_is_prime_fasttest includes trial division, DSA parameter
9056 generation becomes much faster.
9058 This implies a change for the callback functions in DSA_is_prime
9059 and DSA_generate_parameters: The callback function is called once
9060 for each positive witness in the Rabin-Miller test, not just
9061 occasionally in the inner loop; and the parameters to the
9062 callback function now provide an iteration count for the outer
9063 loop rather than for the current invocation of the inner loop.
9064 DSA_generate_parameters additionally can call the callback
9065 function with an 'iteration count' of -1, meaning that a
9066 candidate has passed the trial division test (when q is generated
9067 from an application-provided seed, trial division is skipped).
9070 *) New function BN_is_prime_fasttest that optionally does trial
9071 division before starting the Rabin-Miller test and has
9072 an additional BN_CTX * argument (whereas BN_is_prime always
9073 has to allocate at least one BN_CTX).
9074 'callback(1, -1, cb_arg)' is called when a number has passed the
9075 trial division stage.
9078 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9082 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9085 *) New function BN_pseudo_rand().
9088 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9089 bignum version of BN_from_montgomery() with the working code from
9090 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9094 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9095 made it impossible to use the same SSL_SESSION data structure in
9096 SSL2 clients in multiple threads.
9099 *) The return value of RAND_load_file() no longer counts bytes obtained
9100 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9101 to seed the PRNG (previously an explicit byte count was required).
9102 [Ulf Möller, Bodo Möller]
9104 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9105 used (char *) instead of (void *) and had casts all over the place.
9108 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9111 *) Retain source code compatibility for BN_prime_checks macro:
9112 BN_is_prime(..., BN_prime_checks, ...) now uses
9113 BN_prime_checks_for_size to determine the appropriate number of
9114 Rabin-Miller iterations.
9117 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9118 DH_CHECK_P_NOT_SAFE_PRIME.
9119 (Check if this is true? OpenPGP calls them "strong".)
9122 *) Merge the functionality of "dh" and "gendh" programs into a new program
9123 "dhparam". The old programs are retained for now but will handle DH keys
9124 (instead of parameters) in future.
9127 *) Make the ciphers, s_server and s_client programs check the return values
9128 when a new cipher list is set.
9131 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9132 ciphers. Before when the 56bit ciphers were enabled the sorting was
9135 The syntax for the cipher sorting has been extended to support sorting by
9136 cipher-strength (using the strength_bits hard coded in the tables).
9137 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9139 Fix a bug in the cipher-command parser: when supplying a cipher command
9140 string with an "undefined" symbol (neither command nor alphanumeric
9141 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9142 an error is flagged.
9144 Due to the strength-sorting extension, the code of the
9145 ssl_create_cipher_list() function was completely rearranged. I hope that
9146 the readability was also increased :-)
9147 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9149 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9150 for the first serial number and places 2 in the serial number file. This
9151 avoids problems when the root CA is created with serial number zero and
9152 the first user certificate has the same issuer name and serial number
9156 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9157 the new code. Add documentation for this stuff.
9160 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9161 X509_*() to X509at_*() on the grounds that they don't handle X509
9162 structures and behave in an analagous way to the X509v3 functions:
9163 they shouldn't be called directly but wrapper functions should be used
9166 So we also now have some wrapper functions that call the X509at functions
9167 when passed certificate requests. (TO DO: similar things can be done with
9168 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9169 things. Some of these need some d2i or i2d and print functionality
9170 because they handle more complex structures.)
9173 *) Add missing #ifndefs that caused missing symbols when building libssl
9174 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9175 NO_RSA in ssl/s2*.c.
9176 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9178 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9179 has a return value which indicates the quality of the random data
9180 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9181 error queue. New function RAND_pseudo_bytes() generates output that is
9182 guaranteed to be unique but not unpredictable. RAND_add is like
9183 RAND_seed, but takes an extra argument for an entropy estimate
9184 (RAND_seed always assumes full entropy).
9187 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9188 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9189 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9190 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9191 false-positive rate of at most 2^-80 for random input.
9194 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9197 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9198 in the 0.9.5 release), this returns the chain
9199 from an X509_CTX structure with a dup of the stack and all
9200 the X509 reference counts upped: so the stack will exist
9201 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9204 Also make SSL_SESSION_print() print out the verify return
9208 *) Add manpage for the pkcs12 command. Also change the default
9209 behaviour so MAC iteration counts are used unless the new
9210 -nomaciter option is used. This improves file security and
9211 only older versions of MSIE (4.0 for example) need it.
9214 *) Honor the no-xxx Configure options when creating .DEF files.
9217 *) Add PKCS#10 attributes to field table: challengePassword,
9218 unstructuredName and unstructuredAddress. These are taken from
9219 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9220 international characters are used.
9222 More changes to X509_ATTRIBUTE code: allow the setting of types
9223 based on strings. Remove the 'loc' parameter when adding
9224 attributes because these will be a SET OF encoding which is sorted
9228 *) Initial changes to the 'req' utility to allow request generation
9229 automation. This will allow an application to just generate a template
9230 file containing all the field values and have req construct the
9233 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9234 used all over the place including certificate requests and PKCS#7
9235 structures. They are currently handled manually where necessary with
9236 some primitive wrappers for PKCS#7. The new functions behave in a
9237 manner analogous to the X509 extension functions: they allow
9238 attributes to be looked up by NID and added.
9240 Later something similar to the X509V3 code would be desirable to
9241 automatically handle the encoding, decoding and printing of the
9242 more complex types. The string types like challengePassword can
9243 be handled by the string table functions.
9245 Also modified the multi byte string table handling. Now there is
9246 a 'global mask' which masks out certain types. The table itself
9247 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9248 is useful when for example there is only one permissible type
9249 (as in countryName) and using the mask might result in no valid
9253 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9254 SSL_get_peer_finished to allow applications to obtain the latest
9255 Finished messages sent to the peer or expected from the peer,
9256 respectively. (SSL_get_peer_finished is usually the Finished message
9257 actually received from the peer, otherwise the protocol will be aborted.)
9259 As the Finished message are message digests of the complete handshake
9260 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9261 be used for external authentication procedures when the authentication
9262 provided by SSL/TLS is not desired or is not enough.
9265 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9266 the host supports BWX extension and if Compaq C is present on the
9267 $PATH. Just exploiting of the BWX extension results in 20-30%
9268 performance kick for some algorithms, e.g. DES and RC4 to mention
9269 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9273 *) Add support for MS "fast SGC". This is arguably a violation of the
9274 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9275 weak crypto and after checking the certificate is SGC a second one
9276 with strong crypto. MS SGC stops the first handshake after receiving
9277 the server certificate message and sends a second client hello. Since
9278 a server will typically do all the time consuming operations before
9279 expecting any further messages from the client (server key exchange
9280 is the most expensive) there is little difference between the two.
9282 To get OpenSSL to support MS SGC we have to permit a second client
9283 hello message after we have sent server done. In addition we have to
9284 reset the MAC if we do get this second client hello.
9287 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9288 if a DER encoded private key is RSA or DSA traditional format. Changed
9289 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9290 format DER encoded private key. Newer code should use PKCS#8 format which
9291 has the key type encoded in the ASN1 structure. Added DER private key
9292 support to pkcs8 application.
9295 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9296 ciphersuites has been selected (as required by the SSL 3/TLS 1
9297 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9298 is set, we interpret this as a request to violate the specification
9299 (the worst that can happen is a handshake failure, and 'correct'
9300 behaviour would result in a handshake failure anyway).
9303 *) In SSL_CTX_add_session, take into account that there might be multiple
9304 SSL_SESSION structures with the same session ID (e.g. when two threads
9305 concurrently obtain them from an external cache).
9306 The internal cache can handle only one SSL_SESSION with a given ID,
9307 so if there's a conflict, we now throw out the old one to achieve
9311 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9312 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9313 some routines that use cipher OIDs: some ciphers do not have OIDs
9314 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9318 *) Simplify the trust setting structure and code. Now we just have
9319 two sequences of OIDs for trusted and rejected settings. These will
9320 typically have values the same as the extended key usage extension
9321 and any application specific purposes.
9323 The trust checking code now has a default behaviour: it will just
9324 check for an object with the same NID as the passed id. Functions can
9325 be provided to override either the default behaviour or the behaviour
9326 for a given id. SSL client, server and email already have functions
9327 in place for compatibility: they check the NID and also return "trusted"
9328 if the certificate is self signed.
9331 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9332 traditional format into an EVP_PKEY structure.
9335 *) Add a password callback function PEM_cb() which either prompts for
9336 a password if usr_data is NULL or otherwise assumes it is a null
9337 terminated password. Allow passwords to be passed on command line
9338 environment or config files in a few more utilities.
9341 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9342 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9343 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9344 Update documentation.
9347 *) Support for ASN1 "NULL" type. This could be handled before by using
9348 ASN1_TYPE but there wasn't any function that would try to read a NULL
9349 and produce an error if it couldn't. For compatibility we also have
9350 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9351 don't allocate anything because they don't need to.
9354 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9356 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9358 *) Rebuild of the memory allocation routines used by OpenSSL code and
9359 possibly others as well. The purpose is to make an interface that
9360 provide hooks so anyone can build a separate set of allocation and
9361 deallocation routines to be used by OpenSSL, for example memory
9362 pool implementations, or something else, which was previously hard
9363 since Malloc(), Realloc() and Free() were defined as macros having
9364 the values malloc, realloc and free, respectively (except for Win32
9365 compilations). The same is provided for memory debugging code.
9366 OpenSSL already comes with functionality to find memory leaks, but
9367 this gives people a chance to debug other memory problems.
9369 With these changes, a new set of functions and macros have appeared:
9371 CRYPTO_set_mem_debug_functions() [F]
9372 CRYPTO_get_mem_debug_functions() [F]
9373 CRYPTO_dbg_set_options() [F]
9374 CRYPTO_dbg_get_options() [F]
9375 CRYPTO_malloc_debug_init() [M]
9377 The memory debug functions are NULL by default, unless the library
9378 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9379 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9380 gives the standard debugging functions that come with OpenSSL) or
9381 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9382 provided by the library user) must be used. When the standard
9383 debugging functions are used, CRYPTO_dbg_set_options can be used to
9384 request additional information:
9385 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9386 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9388 Also, things like CRYPTO_set_mem_functions will always give the
9389 expected result (the new set of functions is used for allocation
9390 and deallocation) at all times, regardless of platform and compiler
9393 To finish it up, some functions that were never use in any other
9394 way than through macros have a new API and new semantic:
9397 CRYPTO_dbg_realloc()
9400 All macros of value have retained their old syntax.
9401 [Richard Levitte and Bodo Moeller]
9403 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9404 ordering of SMIMECapabilities wasn't in "strength order" and there
9405 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9409 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9410 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9411 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9413 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9414 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9415 functionality to handle multipart/signed properly) and a utility
9416 called 'smime' to call all this stuff. This is based on code I
9417 originally wrote for Celo who have kindly allowed it to be
9418 included in OpenSSL.
9421 *) Add variants des_set_key_checked and des_set_key_unchecked of
9422 des_set_key (aka des_key_sched). Global variable des_check_key
9423 decides which of these is called by des_set_key; this way
9424 des_check_key behaves as it always did, but applications and
9425 the library itself, which was buggy for des_check_key == 1,
9426 have a cleaner way to pick the version they need.
9429 *) New function PKCS12_newpass() which changes the password of a
9433 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9434 dynamic mix. In both cases the ids can be used as an index into the
9435 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9436 functions so they accept a list of the field values and the
9437 application doesn't need to directly manipulate the X509_TRUST
9441 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9445 *) Modify the way the V3 extension code looks up extensions. This now
9446 works in a similar way to the object code: we have some "standard"
9447 extensions in a static table which is searched with OBJ_bsearch()
9448 and the application can add dynamic ones if needed. The file
9449 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9450 updated whenever a new extension is added to the core code and kept
9451 in ext_nid order. There is a simple program 'tabtest.c' which checks
9452 this. New extensions are not added too often so this file can readily
9453 be maintained manually.
9455 There are two big advantages in doing things this way. The extensions
9456 can be looked up immediately and no longer need to be "added" using
9457 X509V3_add_standard_extensions(): this function now does nothing.
9458 [Side note: I get *lots* of email saying the extension code doesn't
9459 work because people forget to call this function]
9460 Also no dynamic allocation is done unless new extensions are added:
9461 so if we don't add custom extensions there is no need to call
9462 X509V3_EXT_cleanup().
9465 *) Modify enc utility's salting as follows: make salting the default. Add a
9466 magic header, so unsalted files fail gracefully instead of just decrypting
9467 to garbage. This is because not salting is a big security hole, so people
9468 should be discouraged from doing it.
9471 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9472 digest to be passed on the command line but it only used this
9473 parameter when signing a certificate. Modified so all relevant
9474 operations are affected by the digest parameter including the
9475 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9476 DSA key was used because it didn't fix the digest.
9479 *) Initial certificate chain verify code. Currently tests the untrusted
9480 certificates for consistency with the verify purpose (which is set
9481 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9483 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9484 this is because it will reject chains with invalid extensions whereas
9485 every previous version of OpenSSL and SSLeay made no checks at all.
9487 Trust code: checks the root CA for the relevant trust settings. Trust
9488 settings have an initial value consistent with the verify purpose: e.g.
9489 if the verify purpose is for SSL client use it expects the CA to be
9490 trusted for SSL client use. However the default value can be changed to
9491 permit custom trust settings: one example of this would be to only trust
9492 certificates from a specific "secure" set of CAs.
9494 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9495 which should be used for version portability: especially since the
9496 verify structure is likely to change more often now.
9498 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9499 to set them. If not set then assume SSL clients will verify SSL servers
9502 Two new options to the verify program: -untrusted allows a set of
9503 untrusted certificates to be passed in and -purpose which sets the
9504 intended purpose of the certificate. If a purpose is set then the
9505 new chain verify code is used to check extension consistency.
9508 *) Support for the authority information access extension.
9511 *) Modify RSA and DSA PEM read routines to transparently handle
9512 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9513 public keys in a format compatible with certificate
9514 SubjectPublicKeyInfo structures. Unfortunately there were already
9515 functions called *_PublicKey_* which used various odd formats so
9516 these are retained for compatibility: however the DSA variants were
9517 never in a public release so they have been deleted. Changed dsa/rsa
9518 utilities to handle the new format: note no releases ever handled public
9519 keys so we should be OK.
9521 The primary motivation for this change is to avoid the same fiasco
9522 that dogs private keys: there are several incompatible private key
9523 formats some of which are standard and some OpenSSL specific and
9524 require various evil hacks to allow partial transparent handling and
9525 even then it doesn't work with DER formats. Given the option anything
9526 other than PKCS#8 should be dumped: but the other formats have to
9527 stay in the name of compatibility.
9529 With public keys and the benefit of hindsight one standard format
9530 is used which works with EVP_PKEY, RSA or DSA structures: though
9531 it clearly returns an error if you try to read the wrong kind of key.
9533 Added a -pubkey option to the 'x509' utility to output the public key.
9534 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9535 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9536 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9537 that do the same as the EVP_PKEY_assign_*() except they up the
9538 reference count of the added key (they don't "swallow" the
9542 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9543 CRLs would fail if the file contained no certificates or no CRLs:
9544 added a new function to read in both types and return the number
9545 read: this means that if none are read it will be an error. The
9546 DER versions of the certificate and CRL reader would always fail
9547 because it isn't possible to mix certificates and CRLs in DER format
9548 without choking one or the other routine. Changed this to just read
9549 a certificate: this is the best we can do. Also modified the code
9550 in apps/verify.c to take notice of return codes: it was previously
9551 attempting to read in certificates from NULL pointers and ignoring
9552 any errors: this is one reason why the cert and CRL reader seemed
9553 to work. It doesn't check return codes from the default certificate
9554 routines: these may well fail if the certificates aren't installed.
9557 *) Code to support otherName option in GeneralName.
9560 *) First update to verify code. Change the verify utility
9561 so it warns if it is passed a self signed certificate:
9562 for consistency with the normal behaviour. X509_verify
9563 has been modified to it will now verify a self signed
9564 certificate if *exactly* the same certificate appears
9565 in the store: it was previously impossible to trust a
9566 single self signed certificate. This means that:
9567 openssl verify ss.pem
9568 now gives a warning about a self signed certificate but
9569 openssl verify -CAfile ss.pem ss.pem
9573 *) For servers, store verify_result in SSL_SESSION data structure
9574 (and add it to external session representation).
9575 This is needed when client certificate verifications fails,
9576 but an application-provided verification callback (set by
9577 SSL_CTX_set_cert_verify_callback) allows accepting the session
9578 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9579 but returns 1): When the session is reused, we have to set
9580 ssl->verify_result to the appropriate error code to avoid
9582 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9584 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9585 case in PKCS7_dataInit() where the signed PKCS7 structure
9586 didn't contain any existing data because it was being created.
9587 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9589 *) Add a salt to the key derivation routines in enc.c. This
9590 forms the first 8 bytes of the encrypted file. Also add a
9591 -S option to allow a salt to be input on the command line.
9594 *) New function X509_cmp(). Oddly enough there wasn't a function
9595 to compare two certificates. We do this by working out the SHA1
9596 hash and comparing that. X509_cmp() will be needed by the trust
9600 *) SSL_get1_session() is like SSL_get_session(), but increments
9601 the reference count in the SSL_SESSION returned.
9602 [Geoff Thorpe <geoff@eu.c2.net>]
9604 *) Fix for 'req': it was adding a null to request attributes.
9605 Also change the X509_LOOKUP and X509_INFO code to handle
9606 certificate auxiliary information.
9609 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9613 *) Add the possibility to add extra information to the memory leak
9614 detecting output, to form tracebacks, showing from where each
9615 allocation was originated: CRYPTO_push_info("constant string") adds
9616 the string plus current file name and line number to a per-thread
9617 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9618 is like calling CYRPTO_pop_info() until the stack is empty.
9619 Also updated memory leak detection code to be multi-thread-safe.
9622 *) Add options -text and -noout to pkcs7 utility and delete the
9623 encryption options which never did anything. Update docs.
9626 *) Add options to some of the utilities to allow the pass phrase
9627 to be included on either the command line (not recommended on
9628 OSes like Unix) or read from the environment. Update the
9629 manpages and fix a few bugs.
9632 *) Add a few manpages for some of the openssl commands.
9635 *) Fix the -revoke option in ca. It was freeing up memory twice,
9636 leaking and not finding already revoked certificates.
9639 *) Extensive changes to support certificate auxiliary information.
9640 This involves the use of X509_CERT_AUX structure and X509_AUX
9641 functions. An X509_AUX function such as PEM_read_X509_AUX()
9642 can still read in a certificate file in the usual way but it
9643 will also read in any additional "auxiliary information". By
9644 doing things this way a fair degree of compatibility can be
9645 retained: existing certificates can have this information added
9646 using the new 'x509' options.
9648 Current auxiliary information includes an "alias" and some trust
9649 settings. The trust settings will ultimately be used in enhanced
9650 certificate chain verification routines: currently a certificate
9651 can only be trusted if it is self signed and then it is trusted
9655 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9656 The problem was that one of the replacement routines had not been working
9657 since SSLeay releases. For now the offending routine has been replaced
9658 with non-optimised assembler. Even so, this now gives around 95%
9659 performance improvement for 1024 bit RSA signs.
9662 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9663 handling. Most clients have the effective key size in bits equal to
9664 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9665 A few however don't do this and instead use the size of the decrypted key
9666 to determine the RC2 key length and the AlgorithmIdentifier to determine
9667 the effective key length. In this case the effective key length can still
9668 be 40 bits but the key length can be 168 bits for example. This is fixed
9669 by manually forcing an RC2 key into the EVP_PKEY structure because the
9670 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9671 the key length and effective key length are equal.
9674 *) Add a bunch of functions that should simplify the creation of
9675 X509_NAME structures. Now you should be able to do:
9676 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9677 and have it automatically work out the correct field type and fill in
9678 the structures. The more adventurous can try:
9679 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9680 and it will (hopefully) work out the correct multibyte encoding.
9683 *) Change the 'req' utility to use the new field handling and multibyte
9684 copy routines. Before the DN field creation was handled in an ad hoc
9685 way in req, ca, and x509 which was rather broken and didn't support
9686 BMPStrings or UTF8Strings. Since some software doesn't implement
9687 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9688 using the dirstring_type option. See the new comment in the default
9689 openssl.cnf for more info.
9692 *) Make crypto/rand/md_rand.c more robust:
9693 - Assure unique random numbers after fork().
9694 - Make sure that concurrent threads access the global counter and
9695 md serializably so that we never lose entropy in them
9696 or use exactly the same state in multiple threads.
9697 Access to the large state is not always serializable because
9698 the additional locking could be a performance killer, and
9699 md should be large enough anyway.
9702 *) New file apps/app_rand.c with commonly needed functionality
9703 for handling the random seed file.
9705 Use the random seed file in some applications that previously did not:
9707 dsaparam -genkey (which also ignored its '-rand' option),
9710 x509 (when signing).
9711 Except on systems with /dev/urandom, it is crucial to have a random
9712 seed file at least for key creation, DSA signing, and for DH exchanges;
9713 for RSA signatures we could do without one.
9715 gendh and gendsa (unlike genrsa) used to read only the first byte
9716 of each file listed in the '-rand' option. The function as previously
9717 found in genrsa is now in app_rand.c and is used by all programs
9718 that support '-rand'.
9721 *) In RAND_write_file, use mode 0600 for creating files;
9722 don't just chmod when it may be too late.
9725 *) Report an error from X509_STORE_load_locations
9726 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9729 *) New function ASN1_mbstring_copy() this copies a string in either
9730 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9731 into an ASN1_STRING type. A mask of permissible types is passed
9732 and it chooses the "minimal" type to use or an error if not type
9736 *) Add function equivalents to the various macros in asn1.h. The old
9737 macros are retained with an M_ prefix. Code inside the library can
9738 use the M_ macros. External code (including the openssl utility)
9739 should *NOT* in order to be "shared library friendly".
9742 *) Add various functions that can check a certificate's extensions
9743 to see if it usable for various purposes such as SSL client,
9744 server or S/MIME and CAs of these types. This is currently
9745 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9746 verification. Also added a -purpose flag to x509 utility to
9747 print out all the purposes.
9750 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9754 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9755 for, obtain and decode and extension and obtain its critical flag.
9756 This allows all the necessary extension code to be handled in a
9757 single function call.
9760 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9761 platforms. See crypto/rc4/rc4_enc.c for further details.
9764 *) New -noout option to asn1parse. This causes no output to be produced
9765 its main use is when combined with -strparse and -out to extract data
9766 from a file (which may not be in ASN.1 format).
9769 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9770 when producing the local key id.
9771 [Richard Levitte <levitte@stacken.kth.se>]
9773 *) New option -dhparam in s_server. This allows a DH parameter file to be
9774 stated explicitly. If it is not stated then it tries the first server
9775 certificate file. The previous behaviour hard coded the filename
9779 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9780 a public key to be input or output. For example:
9781 openssl rsa -in key.pem -pubout -out pubkey.pem
9782 Also added necessary DSA public key functions to handle this.
9785 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9786 in the message. This was handled by allowing
9787 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9788 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9790 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9791 to the end of the strings whereas this didn't. This would cause problems
9792 if strings read with d2i_ASN1_bytes() were later modified.
9793 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9795 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9796 data and it contains EOF it will end up returning an error. This is
9797 caused by input 46 bytes long. The cause is due to the way base64
9798 BIOs find the start of base64 encoded data. They do this by trying a
9799 trial decode on each line until they find one that works. When they
9800 do a flag is set and it starts again knowing it can pass all the
9801 data directly through the decoder. Unfortunately it doesn't reset
9802 the context it uses. This means that if EOF is reached an attempt
9803 is made to pass two EOFs through the context and this causes the
9804 resulting error. This can also cause other problems as well. As is
9805 usual with these problems it takes *ages* to find and the fix is
9806 trivial: move one line.
9807 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9809 *) Ugly workaround to get s_client and s_server working under Windows. The
9810 old code wouldn't work because it needed to select() on sockets and the
9811 tty (for keypresses and to see if data could be written). Win32 only
9812 supports select() on sockets so we select() with a 1s timeout on the
9813 sockets and then see if any characters are waiting to be read, if none
9814 are present then we retry, we also assume we can always write data to
9815 the tty. This isn't nice because the code then blocks until we've
9816 received a complete line of data and it is effectively polling the
9817 keyboard at 1s intervals: however it's quite a bit better than not
9818 working at all :-) A dedicated Windows application might handle this
9819 with an event loop for example.
9822 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9823 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9824 will be called when RSA_sign() and RSA_verify() are used. This is useful
9825 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9826 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9827 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9828 This necessitated the support of an extra signature type NID_md5_sha1
9829 for SSL signatures and modifications to the SSL library to use it instead
9830 of calling RSA_public_decrypt() and RSA_private_encrypt().
9833 *) Add new -verify -CAfile and -CApath options to the crl program, these
9834 will lookup a CRL issuers certificate and verify the signature in a
9835 similar way to the verify program. Tidy up the crl program so it
9836 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9837 less strict. It will now permit CRL extensions even if it is not
9838 a V2 CRL: this will allow it to tolerate some broken CRLs.
9841 *) Initialize all non-automatic variables each time one of the openssl
9842 sub-programs is started (this is necessary as they may be started
9843 multiple times from the "OpenSSL>" prompt).
9844 [Lennart Bang, Bodo Moeller]
9846 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9847 removing all other RSA functionality (this is what NO_RSA does). This
9848 is so (for example) those in the US can disable those operations covered
9849 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9853 *) Non-copying interface to BIO pairs.
9854 (still largely untested)
9857 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9858 ASCII string. This was handled independently in various places before.
9861 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9862 UTF8 strings a character at a time.
9865 *) Use client_version from client hello to select the protocol
9866 (s23_srvr.c) and for RSA client key exchange verification
9867 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9870 *) Add various utility functions to handle SPKACs, these were previously
9871 handled by poking round in the structure internals. Added new function
9872 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9873 print, verify and generate SPKACs. Based on an original idea from
9874 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9877 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9880 *) Allow the config file extension section to be overwritten on the
9881 command line. Based on an original idea from Massimiliano Pala
9882 <madwolf@comune.modena.it>. The new option is called -extensions
9883 and can be applied to ca, req and x509. Also -reqexts to override
9884 the request extensions in req and -crlexts to override the crl extensions
9888 *) Add new feature to the SPKAC handling in ca. Now you can include
9889 the same field multiple times by preceding it by "XXXX." for example:
9892 this is the same syntax as used in the req config file.
9895 *) Allow certificate extensions to be added to certificate requests. These
9896 are specified in a 'req_extensions' option of the req section of the
9897 config file. They can be printed out with the -text option to req but
9898 are otherwise ignored at present.
9901 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9902 data read consists of only the final block it would not decrypted because
9903 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9904 A misplaced 'break' also meant the decrypted final block might not be
9905 copied until the next read.
9908 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9909 a few extra parameters to the DH structure: these will be useful if
9910 for example we want the value of 'q' or implement X9.42 DH.
9913 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9914 provides hooks that allow the default DSA functions or functions on a
9915 "per key" basis to be replaced. This allows hardware acceleration and
9916 hardware key storage to be handled without major modification to the
9917 library. Also added low level modexp hooks and CRYPTO_EX structure and
9918 associated functions.
9921 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9922 as "read only": it can't be written to and the buffer it points to will
9923 not be freed. Reading from a read only BIO is much more efficient than
9924 a normal memory BIO. This was added because there are several times when
9925 an area of memory needs to be read from a BIO. The previous method was
9926 to create a memory BIO and write the data to it, this results in two
9927 copies of the data and an O(n^2) reading algorithm. There is a new
9928 function BIO_new_mem_buf() which creates a read only memory BIO from
9929 an area of memory. Also modified the PKCS#7 routines to use read only
9933 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9934 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9935 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9936 but a retry condition occured while trying to read the rest.
9939 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9940 NID_pkcs7_encrypted by default: this was wrong since this should almost
9941 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9942 the encrypted data type: this is a more sensible place to put it and it
9943 allows the PKCS#12 code to be tidied up that duplicated this
9947 *) Changed obj_dat.pl script so it takes its input and output files on
9948 the command line. This should avoid shell escape redirection problems
9952 *) Initial support for certificate extension requests, these are included
9953 in things like Xenroll certificate requests. Included functions to allow
9954 extensions to be obtained and added.
9957 *) -crlf option to s_client and s_server for sending newlines as
9958 CRLF (as required by many protocols).
9961 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9963 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9964 [Ralf S. Engelschall]
9966 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9967 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9969 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9973 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9974 DH parameters/keys (q is lost during that conversion, but the resulting
9975 DH parameters contain its length).
9977 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9978 much faster than DH_generate_parameters (which creates parameters
9979 where p = 2*q + 1), and also the smaller q makes DH computations
9980 much more efficient (160-bit exponentiation instead of 1024-bit
9981 exponentiation); so this provides a convenient way to support DHE
9982 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9983 utter importance to use
9984 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9986 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9987 when such DH parameters are used, because otherwise small subgroup
9988 attacks may become possible!
9991 *) Avoid memory leak in i2d_DHparams.
9994 *) Allow the -k option to be used more than once in the enc program:
9995 this allows the same encrypted message to be read by multiple recipients.
9998 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9999 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
10000 it will always use the numerical form of the OID, even if it has a short
10004 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
10005 method only got called if p,q,dmp1,dmq1,iqmp components were present,
10006 otherwise bn_mod_exp was called. In the case of hardware keys for example
10007 no private key components need be present and it might store extra data
10008 in the RSA structure, which cannot be accessed from bn_mod_exp.
10009 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
10010 private key operations.
10013 *) Added support for SPARC Linux.
10016 *) pem_password_cb function type incompatibly changed from
10017 typedef int pem_password_cb(char *buf, int size, int rwflag);
10019 ....(char *buf, int size, int rwflag, void *userdata);
10020 so that applications can pass data to their callbacks:
10021 The PEM[_ASN1]_{read,write}... functions and macros now take an
10022 additional void * argument, which is just handed through whenever
10023 the password callback is called.
10024 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10026 New function SSL_CTX_set_default_passwd_cb_userdata.
10028 Compatibility note: As many C implementations push function arguments
10029 onto the stack in reverse order, the new library version is likely to
10030 interoperate with programs that have been compiled with the old
10031 pem_password_cb definition (PEM_whatever takes some data that
10032 happens to be on the stack as its last argument, and the callback
10033 just ignores this garbage); but there is no guarantee whatsoever that
10036 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10037 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10038 problems not only on Windows, but also on some Unix platforms.
10039 To avoid problematic command lines, these definitions are now in an
10040 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10041 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10044 *) MIPS III/IV assembler module is reimplemented.
10047 *) More DES library cleanups: remove references to srand/rand and
10048 delete an unused file.
10051 *) Add support for the the free Netwide assembler (NASM) under Win32,
10052 since not many people have MASM (ml) and it can be hard to obtain.
10053 This is currently experimental but it seems to work OK and pass all
10054 the tests. Check out INSTALL.W32 for info.
10057 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10058 without temporary keys kept an extra copy of the server key,
10059 and connections with temporary keys did not free everything in case
10063 *) New function RSA_check_key and new openssl rsa option -check
10064 for verifying the consistency of RSA keys.
10065 [Ulf Moeller, Bodo Moeller]
10067 *) Various changes to make Win32 compile work:
10068 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10069 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10070 comparison" warnings.
10071 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10074 *) Add a debugging option to PKCS#5 v2 key generation function: when
10075 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10076 derived keys are printed to stderr.
10079 *) Copy the flags in ASN1_STRING_dup().
10080 [Roman E. Pavlov <pre@mo.msk.ru>]
10082 *) The x509 application mishandled signing requests containing DSA
10083 keys when the signing key was also DSA and the parameters didn't match.
10085 It was supposed to omit the parameters when they matched the signing key:
10086 the verifying software was then supposed to automatically use the CA's
10087 parameters if they were absent from the end user certificate.
10089 Omitting parameters is no longer recommended. The test was also
10090 the wrong way round! This was probably due to unusual behaviour in
10091 EVP_cmp_parameters() which returns 1 if the parameters match.
10092 This meant that parameters were omitted when they *didn't* match and
10093 the certificate was useless. Certificates signed with 'ca' didn't have
10095 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10097 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10098 The interface is as follows:
10099 Applications can use
10100 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10101 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10102 "off" is now the default.
10103 The library internally uses
10104 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10105 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10106 to disable memory-checking temporarily.
10108 Some inconsistent states that previously were possible (and were
10109 even the default) are now avoided.
10111 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10112 with each memory chunk allocated; this is occasionally more helpful
10113 than just having a counter.
10115 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10117 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10121 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10122 which largely parallels "options", but is for changing API behaviour,
10123 whereas "options" are about protocol behaviour.
10124 Initial "mode" flags are:
10126 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10127 a single record has been written.
10128 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10129 retries use the same buffer location.
10130 (But all of the contents must be
10134 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10137 *) Fix problems with no-hmac etc.
10138 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10140 *) New functions RSA_get_default_method(), RSA_set_method() and
10141 RSA_get_method(). These allows replacement of RSA_METHODs without having
10142 to mess around with the internals of an RSA structure.
10145 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10146 Also really enable memory leak checks in openssl.c and in some
10148 [Chad C. Mulligan, Bodo Moeller]
10150 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10151 up the length of negative integers. This has now been simplified to just
10152 store the length when it is first determined and use it later, rather
10153 than trying to keep track of where data is copied and updating it to
10155 [Steve Henson, reported by Brien Wheeler
10156 <bwheeler@authentica-security.com>]
10158 *) Add a new function PKCS7_signatureVerify. This allows the verification
10159 of a PKCS#7 signature but with the signing certificate passed to the
10160 function itself. This contrasts with PKCS7_dataVerify which assumes the
10161 certificate is present in the PKCS#7 structure. This isn't always the
10162 case: certificates can be omitted from a PKCS#7 structure and be
10163 distributed by "out of band" means (such as a certificate database).
10166 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10167 function prototypes in pem.h, also change util/mkdef.pl to add the
10168 necessary function names.
10171 *) mk1mf.pl (used by Windows builds) did not properly read the
10172 options set by Configure in the top level Makefile, and Configure
10173 was not even able to write more than one option correctly.
10174 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10177 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10178 file to be loaded from a BIO or FILE pointer. The BIO version will
10179 for example allow memory BIOs to contain config info.
10182 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10183 Whoever hopes to achieve shared-library compatibility across versions
10184 must use this, not the compile-time macro.
10185 (Exercise 0.9.4: Which is the minimum library version required by
10187 Note: All this applies only to multi-threaded programs, others don't
10191 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10192 through a BIO pair triggered the default case, i.e.
10193 SSLerr(...,SSL_R_UNKNOWN_STATE).
10196 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10197 can use the SSL library even if none of the specific BIOs is
10201 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10202 for the encoded length.
10203 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10205 *) Add initial documentation of the X509V3 functions.
10208 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10209 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10210 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10211 secure PKCS#8 private key format with a high iteration count.
10214 *) Fix determination of Perl interpreter: A perl or perl5
10215 _directory_ in $PATH was also accepted as the interpreter.
10216 [Ralf S. Engelschall]
10218 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10219 wrong with it but it was very old and did things like calling
10220 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10221 unusual formatting.
10224 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10225 to use the new extension code.
10228 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10229 with macros. This should make it easier to change their form, add extra
10230 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10234 *) Add to configuration table a new entry that can specify an alternative
10235 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10236 according to Mark Crispin <MRC@Panda.COM>.
10240 *) DES CBC did not update the IV. Weird.
10243 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10244 Changing the behaviour of the former might break existing programs --
10245 where IV updating is needed, des_ncbc_encrypt can be used.
10248 *) When bntest is run from "make test" it drives bc to check its
10249 calculations, as well as internally checking them. If an internal check
10250 fails, it needs to cause bc to give a non-zero result or make test carries
10251 on without noticing the failure. Fixed.
10254 *) DES library cleanups.
10257 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10258 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10259 ciphers. NOTE: although the key derivation function has been verified
10260 against some published test vectors it has not been extensively tested
10261 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10265 *) Instead of "mkdir -p", which is not fully portable, use new
10266 Perl script "util/mkdir-p.pl".
10269 *) Rewrite the way password based encryption (PBE) is handled. It used to
10270 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10271 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10272 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10273 the 'parameter' field of the AlgorithmIdentifier is passed to the
10274 underlying key generation function so it must do its own ASN1 parsing.
10275 This has also changed the EVP_PBE_CipherInit() function which now has a
10276 'parameter' argument instead of literal salt and iteration count values
10277 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10280 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10281 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10282 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10283 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10284 value was just used as a "magic string" and not used directly its
10285 value doesn't matter.
10288 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10292 *) "linux-sparc64" configuration (ultrapenguin).
10293 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10294 "linux-sparc" configuration.
10295 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10297 *) config now generates no-xxx options for missing ciphers.
10300 *) Support the EBCDIC character set (work in progress).
10301 File ebcdic.c not yet included because it has a different license.
10302 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10304 *) Support BS2000/OSD-POSIX.
10305 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10307 *) Make callbacks for key generation use void * instead of char *.
10310 *) Make S/MIME samples compile (not yet tested).
10313 *) Additional typesafe stacks.
10316 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10320 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10322 *) New configuration variant "sco5-gcc".
10324 *) Updated some demos.
10325 [Sean O Riordain, Wade Scholine]
10327 *) Add missing BIO_free at exit of pkcs12 application.
10330 *) Fix memory leak in conf.c.
10333 *) Updates for Win32 to assembler version of MD5.
10336 *) Set #! path to perl in apps/der_chop to where we found it
10337 instead of using a fixed path.
10340 *) SHA library changes for irix64-mips4-cc.
10343 *) Improvements for VMS support.
10347 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10349 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10350 This also avoids the problems with SC4.2 and unpatched SC5.
10351 [Andy Polyakov <appro@fy.chalmers.se>]
10353 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10354 These are required because of the typesafe stack would otherwise break
10355 existing code. If old code used a structure member which used to be STACK
10356 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10357 sk_num or sk_value it would produce an error because the num, data members
10358 are not present in STACK_OF. Now it just produces a warning. sk_set
10359 replaces the old method of assigning a value to sk_value
10360 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10361 that does this will no longer work (and should use sk_set instead) but
10362 this could be regarded as a "questionable" behaviour anyway.
10365 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10366 correctly handle encrypted S/MIME data.
10369 *) Change type of various DES function arguments from des_cblock
10370 (which means, in function argument declarations, pointer to char)
10371 to des_cblock * (meaning pointer to array with 8 char elements),
10372 which allows the compiler to do more typechecking; it was like
10373 that back in SSLeay, but with lots of ugly casts.
10375 Introduce new type const_des_cblock.
10378 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10379 problems: find RecipientInfo structure that matches recipient certificate
10380 and initialise the ASN1 structures properly based on passed cipher.
10383 *) Belatedly make the BN tests actually check the results.
10386 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10387 to and from BNs: it was completely broken. New compilation option
10388 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10389 key elements as negative integers.
10392 *) Reorganize and speed up MD5.
10393 [Andy Polyakov <appro@fy.chalmers.se>]
10396 [Richard Levitte <richard@levitte.org>]
10398 *) New option -out to asn1parse to allow the parsed structure to be
10399 output to a file. This is most useful when combined with the -strparse
10400 option to examine the output of things like OCTET STRINGS.
10403 *) Make SSL library a little more fool-proof by not requiring any longer
10404 that SSL_set_{accept,connect}_state be called before
10405 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10406 in many applications because usually everything *appeared* to work as
10407 intended anyway -- now it really works as intended).
10410 *) Move openssl.cnf out of lib/.
10413 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10414 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10415 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10416 [Ralf S. Engelschall]
10418 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10419 handle PKCS#7 enveloped data properly.
10420 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10422 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10423 copying pointers. The cert_st handling is changed by this in
10424 various ways (and thus what used to be known as ctx->default_cert
10425 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10426 any longer when s->cert does not give us what we need).
10427 ssl_cert_instantiate becomes obsolete by this change.
10428 As soon as we've got the new code right (possibly it already is?),
10429 we have solved a couple of bugs of the earlier code where s->cert
10430 was used as if it could not have been shared with other SSL structures.
10432 Note that using the SSL API in certain dirty ways now will result
10433 in different behaviour than observed with earlier library versions:
10434 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10435 does not influence s as it used to.
10437 In order to clean up things more thoroughly, inside SSL_SESSION
10438 we don't use CERT any longer, but a new structure SESS_CERT
10439 that holds per-session data (if available); currently, this is
10440 the peer's certificate chain and, for clients, the server's certificate
10441 and temporary key. CERT holds only those values that can have
10442 meaningful defaults in an SSL_CTX.
10445 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10446 from the internal representation. Various PKCS#7 fixes: remove some
10447 evil casts and set the enc_dig_alg field properly based on the signing
10451 *) Allow PKCS#12 password to be set from the command line or the
10452 environment. Let 'ca' get its config file name from the environment
10453 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10457 *) Allow certificate policies extension to use an IA5STRING for the
10458 organization field. This is contrary to the PKIX definition but
10459 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10463 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10464 without disallowing inline assembler and the like for non-pedantic builds.
10467 *) Support Borland C++ builder.
10468 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10470 *) Support Mingw32.
10473 *) SHA-1 cleanups and performance enhancements.
10474 [Andy Polyakov <appro@fy.chalmers.se>]
10476 *) Sparc v8plus assembler for the bignum library.
10477 [Andy Polyakov <appro@fy.chalmers.se>]
10479 *) Accept any -xxx and +xxx compiler options in Configure.
10482 *) Update HPUX configuration.
10485 *) Add missing sk_<type>_unshift() function to safestack.h
10486 [Ralf S. Engelschall]
10488 *) New function SSL_CTX_use_certificate_chain_file that sets the
10489 "extra_cert"s in addition to the certificate. (This makes sense
10490 only for "PEM" format files, as chains as a whole are not
10494 *) Support verify_depth from the SSL API.
10495 x509_vfy.c had what can be considered an off-by-one-error:
10496 Its depth (which was not part of the external interface)
10497 was actually counting the number of certificates in a chain;
10498 now it really counts the depth.
10501 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10502 instead of X509err, which often resulted in confusing error
10503 messages since the error codes are not globally unique
10504 (e.g. an alleged error in ssl3_accept when a certificate
10505 didn't match the private key).
10507 *) New function SSL_CTX_set_session_id_context that allows to set a default
10508 value (so that you don't need SSL_set_session_id_context for each
10509 connection using the SSL_CTX).
10512 *) OAEP decoding bug fix.
10515 *) Support INSTALL_PREFIX for package builders, as proposed by
10519 *) New Configure options "threads" and "no-threads". For systems
10520 where the proper compiler options are known (currently Solaris
10521 and Linux), "threads" is the default.
10524 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10527 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10528 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10529 such as /usr/local/bin.
10532 *) "make linux-shared" to build shared libraries.
10533 [Niels Poppe <niels@netbox.org>]
10535 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10538 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10539 extension adding in x509 utility.
10542 *) Remove NOPROTO sections and error code comments.
10545 *) Partial rewrite of the DEF file generator to now parse the ANSI
10549 *) New Configure options --prefix=DIR and --openssldir=DIR.
10552 *) Complete rewrite of the error code script(s). It is all now handled
10553 by one script at the top level which handles error code gathering,
10554 header rewriting and C source file generation. It should be much better
10555 than the old method: it now uses a modified version of Ulf's parser to
10556 read the ANSI prototypes in all header files (thus the old K&R definitions
10557 aren't needed for error creation any more) and do a better job of
10558 translating function codes into names. The old 'ASN1 error code imbedded
10559 in a comment' is no longer necessary and it doesn't use .err files which
10560 have now been deleted. Also the error code call doesn't have to appear all
10561 on one line (which resulted in some large lines...).
10564 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10567 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10568 0 (which usually indicates a closed connection), but continue reading.
10571 *) Fix some race conditions.
10574 *) Add support for CRL distribution points extension. Add Certificate
10575 Policies and CRL distribution points documentation.
10578 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10581 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10582 8 of keying material. Merlin has also confirmed interop with this fix
10583 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10584 [Merlin Hughes <merlin@baltimore.ie>]
10586 *) Fix lots of warnings.
10587 [Richard Levitte <levitte@stacken.kth.se>]
10589 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10590 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10591 [Richard Levitte <levitte@stacken.kth.se>]
10593 *) Fix problems with sizeof(long) == 8.
10594 [Andy Polyakov <appro@fy.chalmers.se>]
10596 *) Change functions to ANSI C.
10599 *) Fix typos in error codes.
10600 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10602 *) Remove defunct assembler files from Configure.
10605 *) SPARC v8 assembler BIGNUM implementation.
10606 [Andy Polyakov <appro@fy.chalmers.se>]
10608 *) Support for Certificate Policies extension: both print and set.
10609 Various additions to support the r2i method this uses.
10612 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10613 return a const string when you are expecting an allocated buffer.
10616 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10617 types DirectoryString and DisplayText.
10620 *) Add code to allow r2i extensions to access the configuration database,
10621 add an LHASH database driver and add several ctx helper functions.
10624 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10625 fail when they extended the size of a BIGNUM.
10628 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10629 support typesafe stack.
10632 *) Fix typo in SSL_[gs]et_options().
10633 [Nils Frostberg <nils@medcom.se>]
10635 *) Delete various functions and files that belonged to the (now obsolete)
10636 old X509V3 handling code.
10639 *) New Configure option "rsaref".
10642 *) Don't auto-generate pem.h.
10645 *) Introduce type-safe ASN.1 SETs.
10648 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10649 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10651 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10652 that links with OpenSSL (well at least cause lots of warnings), but fear
10653 not: the conversion is trivial, and it eliminates loads of evil casts. A
10654 few STACKed things have been converted already. Feel free to convert more.
10655 In the fullness of time, I'll do away with the STACK type altogether.
10658 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10659 specified in <certfile> by updating the entry in the index.txt file.
10660 This way one no longer has to edit the index.txt file manually for
10661 revoking a certificate. The -revoke option does the gory details now.
10662 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10664 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10665 `-text' option at all and this way the `-noout -text' combination was
10666 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10667 [Ralf S. Engelschall]
10669 *) Make sure a corresponding plain text error message exists for the
10670 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10671 verify callback function determined that a certificate was revoked.
10672 [Ralf S. Engelschall]
10674 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10675 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10676 all available cipers including rc5, which was forgotten until now.
10677 In order to let the testing shell script know which algorithms
10678 are available, a new (up to now undocumented) command
10679 "openssl list-cipher-commands" is used.
10682 *) Bugfix: s_client occasionally would sleep in select() when
10683 it should have checked SSL_pending() first.
10686 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10687 the raw DSA values prior to ASN.1 encoding.
10690 *) Tweaks to Configure
10691 [Niels Poppe <niels@netbox.org>]
10693 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10697 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10700 *) New config option to avoid instructions that are illegal on the 80386.
10701 The default code is faster, but requires at least a 486.
10704 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10705 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10706 same as SSL2_VERSION anyway.
10709 *) New "-showcerts" option for s_client.
10712 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10713 application. Various cleanups and fixes.
10716 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10717 modify error routines to work internally. Add error codes and PBE init
10718 to library startup routines.
10721 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10722 packing functions to asn1 and evp. Changed function names and error
10723 codes along the way.
10726 *) PKCS12 integration: and so it begins... First of several patches to
10727 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10728 objects to objects.h
10731 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10732 and display support for Thawte strong extranet extension.
10735 *) Add LinuxPPC support.
10736 [Jeff Dubrule <igor@pobox.org>]
10738 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10739 bn_div_words in alpha.s.
10740 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10742 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10743 OAEP isn't supported when OpenSSL is built with RSAref.
10744 [Ulf Moeller <ulf@fitug.de>]
10746 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10747 so they no longer are missing under -DNOPROTO.
10748 [Soren S. Jorvang <soren@t.dk>]
10751 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10753 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10754 doesn't work when the session is reused. Coming soon!
10757 *) Fix a security hole, that allows sessions to be reused in the wrong
10758 context thus bypassing client cert protection! All software that uses
10759 client certs and session caches in multiple contexts NEEDS PATCHING to
10760 allow session reuse! A fuller solution is in the works.
10761 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10763 *) Some more source tree cleanups (removed obsolete files
10764 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10765 permission on "config" script to be executable) and a fix for the INSTALL
10767 [Ulf Moeller <ulf@fitug.de>]
10769 *) Remove some legacy and erroneous uses of malloc, free instead of
10771 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10773 *) Make rsa_oaep_test return non-zero on error.
10774 [Ulf Moeller <ulf@fitug.de>]
10776 *) Add support for native Solaris shared libraries. Configure
10777 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10778 if someone would make that last step automatic.
10779 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10781 *) ctx_size was not built with the right compiler during "make links". Fixed.
10784 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10785 except NULL ciphers". This means the default cipher list will no longer
10786 enable NULL ciphers. They need to be specifically enabled e.g. with
10787 the string "DEFAULT:eNULL".
10790 *) Fix to RSA private encryption routines: if p < q then it would
10791 occasionally produce an invalid result. This will only happen with
10792 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10795 *) Be less restrictive and allow also `perl util/perlpath.pl
10796 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10797 because this way one can also use an interpreter named `perl5' (which is
10798 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10799 installed as `perl').
10800 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10802 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10803 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10805 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10806 advapi32.lib to Win32 build and change the pem test comparision
10807 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10808 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10809 and crypto/des/ede_cbcm_enc.c.
10812 *) DES quad checksum was broken on big-endian architectures. Fixed.
10815 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10816 Win32 test batch file so it (might) work again. The Win32 test batch file
10817 is horrible: I feel ill....
10820 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10821 in e_os.h. Audit of header files to check ANSI and non ANSI
10822 sections: 10 functions were absent from non ANSI section and not exported
10823 from Windows DLLs. Fixed up libeay.num for new functions.
10826 *) Make `openssl version' output lines consistent.
10827 [Ralf S. Engelschall]
10829 *) Fix Win32 symbol export lists for BIO functions: Added
10830 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10831 to ms/libeay{16,32}.def.
10832 [Ralf S. Engelschall]
10834 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10835 fine under Unix and passes some trivial tests I've now added. But the
10836 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10837 added to make sure no one expects that this stuff really works in the
10838 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10839 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10841 [Ralf S. Engelschall]
10843 *) Fix the generation of two part addresses in perl.
10844 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10846 *) Add config entry for Linux on MIPS.
10847 [John Tobey <jtobey@channel1.com>]
10849 *) Make links whenever Configure is run, unless we are on Windoze.
10852 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10853 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10857 *) Add a useful kludge to allow package maintainers to specify compiler and
10858 other platforms details on the command line without having to patch the
10859 Configure script everytime: One now can use ``perl Configure
10860 <id>:<details>'', i.e. platform ids are allowed to have details appended
10861 to them (seperated by colons). This is treated as there would be a static
10862 pre-configured entry in Configure's %table under key <id> with value
10863 <details> and ``perl Configure <id>'' is called. So, when you want to
10864 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10865 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10866 now, which overrides the FreeBSD-elf entry on-the-fly.
10867 [Ralf S. Engelschall]
10869 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10872 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10873 on the `perl Configure ...' command line. This way one can compile
10874 OpenSSL libraries with Position Independent Code (PIC) which is needed
10875 for linking it into DSOs.
10876 [Ralf S. Engelschall]
10878 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10882 *) Cleaned up the LICENSE document: The official contact for any license
10883 questions now is the OpenSSL core team under openssl-core@openssl.org.
10884 And add a paragraph about the dual-license situation to make sure people
10885 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10886 to the OpenSSL toolkit.
10887 [Ralf S. Engelschall]
10889 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10890 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10891 Additonally cleaned up the `make links' target: Remove unnecessary
10892 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10893 to speed processing and no longer clutter the display with confusing
10894 stuff. Instead only the actually done links are displayed.
10895 [Ralf S. Engelschall]
10897 *) Permit null encryption ciphersuites, used for authentication only. It used
10898 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10899 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10903 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10904 signed attributes when verifying signatures (this would break them),
10905 the detached data encoding was wrong and public keys obtained using
10906 X509_get_pubkey() weren't freed.
10909 *) Add text documentation for the BUFFER functions. Also added a work around
10910 to a Win95 console bug. This was triggered by the password read stuff: the
10911 last character typed gets carried over to the next fread(). If you were
10912 generating a new cert request using 'req' for example then the last
10913 character of the passphrase would be CR which would then enter the first
10917 *) Added the new `Includes OpenSSL Cryptography Software' button as
10918 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10919 button and can be used by applications based on OpenSSL to show the
10920 relationship to the OpenSSL project.
10921 [Ralf S. Engelschall]
10923 *) Remove confusing variables in function signatures in files
10924 ssl/ssl_lib.c and ssl/ssl.h.
10925 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10927 *) Don't install bss_file.c under PREFIX/include/
10928 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10930 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10931 functions that return function pointers and has support for NT specific
10932 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10933 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10934 unsigned to signed types: this was killing the Win32 compile.
10937 *) Add new certificate file to stack functions,
10938 SSL_add_dir_cert_subjects_to_stack() and
10939 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10940 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10941 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10942 This means that Apache-SSL and similar packages don't have to mess around
10943 to add as many CAs as they want to the preferred list.
10946 *) Experiment with doxygen documentation. Currently only partially applied to
10948 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10949 openssl.doxy as the configuration file.
10952 *) Get rid of remaining C++-style comments which strict C compilers hate.
10953 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10955 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10956 compiled in by default: it has problems with large keys.
10959 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10960 DH private keys and/or callback functions which directly correspond to
10961 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10962 is needed for applications which have to configure certificates on a
10963 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10965 For the RSA certificate situation is makes no difference, but
10966 for the DSA certificate situation this fixes the "no shared cipher"
10967 problem where the OpenSSL cipher selection procedure failed because the
10968 temporary keys were not overtaken from the context and the API provided
10969 no way to reconfigure them.
10970 The new functions now let applications reconfigure the stuff and they
10971 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10972 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10973 non-public-API function ssl_cert_instantiate() is used as a helper
10974 function and also to reduce code redundancy inside ssl_rsa.c.
10975 [Ralf S. Engelschall]
10977 *) Move s_server -dcert and -dkey options out of the undocumented feature
10978 area because they are useful for the DSA situation and should be
10979 recognized by the users.
10980 [Ralf S. Engelschall]
10982 *) Fix the cipher decision scheme for export ciphers: the export bits are
10983 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10984 SSL_EXP_MASK. So, the original variable has to be used instead of the
10985 already masked variable.
10986 [Richard Levitte <levitte@stacken.kth.se>]
10988 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10989 [Richard Levitte <levitte@stacken.kth.se>]
10991 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10992 from `int' to `unsigned int' because it's a length and initialized by
10993 EVP_DigestFinal() which expects an `unsigned int *'.
10994 [Richard Levitte <levitte@stacken.kth.se>]
10996 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10997 script. Instead use the usual Shell->Perl transition trick.
10998 [Ralf S. Engelschall]
11000 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
11001 (in addition to RSA certificates) to match the behaviour of `openssl dsa
11002 -noout -modulus' as it's already the case for `openssl rsa -noout
11003 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
11004 currently the public key is printed (a decision which was already done by
11005 `openssl dsa -modulus' in the past) which serves a similar purpose.
11006 Additionally the NO_RSA no longer completely removes the whole -modulus
11007 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
11009 [Ralf S. Engelschall]
11011 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
11012 BIO. See the source (crypto/evp/bio_ok.c) for more info.
11013 [Arne Ansper <arne@ats.cyber.ee>]
11015 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
11016 to be added. Now both 'req' and 'ca' can use new objects defined in the
11020 *) Add cool BIO that does syslog (or event log on NT).
11021 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
11023 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
11024 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11025 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11026 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11029 *) Add preliminary config info for new extension code.
11032 *) Make RSA_NO_PADDING really use no padding.
11033 [Ulf Moeller <ulf@fitug.de>]
11035 *) Generate errors when private/public key check is done.
11038 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11039 for some CRL extensions and new objects added.
11042 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11043 key usage extension and fuller support for authority key id.
11046 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11047 padding method for RSA, which is recommended for new applications in PKCS
11048 #1 v2.0 (RFC 2437, October 1998).
11049 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11050 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11051 against Bleichbacher's attack on RSA.
11052 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11055 *) Updates to the new SSL compression code
11056 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11058 *) Fix so that the version number in the master secret, when passed
11059 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11060 (because the server will not accept higher), that the version number
11061 is 0x03,0x01, not 0x03,0x00
11062 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11064 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11065 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11066 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11069 *) Support for RAW extensions where an arbitrary extension can be
11070 created by including its DER encoding. See apps/openssl.cnf for
11074 *) Make sure latest Perl versions don't interpret some generated C array
11075 code as Perl array code in the crypto/err/err_genc.pl script.
11076 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11078 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11079 not many people have the assembler. Various Win32 compilation fixes and
11080 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11081 build instructions.
11084 *) Modify configure script 'Configure' to automatically create crypto/date.h
11085 file under Win32 and also build pem.h from pem.org. New script
11086 util/mkfiles.pl to create the MINFO file on environments that can't do a
11087 'make files': perl util/mkfiles.pl >MINFO should work.
11090 *) Major rework of DES function declarations, in the pursuit of correctness
11091 and purity. As a result, many evil casts evaporated, and some weirdness,
11092 too. You may find this causes warnings in your code. Zapping your evil
11093 casts will probably fix them. Mostly.
11096 *) Fix for a typo in asn1.h. Bug fix to object creation script
11097 obj_dat.pl. It considered a zero in an object definition to mean
11098 "end of object": none of the objects in objects.h have any zeros
11099 so it wasn't spotted.
11100 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11102 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11103 Masking (CBCM). In the absence of test vectors, the best I have been able
11104 to do is check that the decrypt undoes the encrypt, so far. Send me test
11105 vectors if you have them.
11108 *) Correct calculation of key length for export ciphers (too much space was
11109 allocated for null ciphers). This has not been tested!
11112 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11113 message is now correct (it understands "crypto" and "ssl" on its
11114 command line). There is also now an "update" option. This will update
11115 the util/ssleay.num and util/libeay.num files with any new functions.
11117 perl util/mkdef.pl crypto ssl update
11118 it will update them.
11121 *) Overhauled the Perl interface (perl/*):
11122 - ported BN stuff to OpenSSL's different BN library
11123 - made the perl/ source tree CVS-aware
11124 - renamed the package from SSLeay to OpenSSL (the files still contain
11125 their history because I've copied them in the repository)
11126 - removed obsolete files (the test scripts will be replaced
11127 by better Test::Harness variants in the future)
11128 [Ralf S. Engelschall]
11130 *) First cut for a very conservative source tree cleanup:
11131 1. merge various obsolete readme texts into doc/ssleay.txt
11132 where we collect the old documents and readme texts.
11133 2. remove the first part of files where I'm already sure that we no
11134 longer need them because of three reasons: either they are just temporary
11135 files which were left by Eric or they are preserved original files where
11136 I've verified that the diff is also available in the CVS via "cvs diff
11137 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11138 the crypto/md/ stuff).
11139 [Ralf S. Engelschall]
11141 *) More extension code. Incomplete support for subject and issuer alt
11142 name, issuer and authority key id. Change the i2v function parameters
11143 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11144 what that's for :-) Fix to ASN1 macro which messed up
11145 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11148 *) Preliminary support for ENUMERATED type. This is largely copied from the
11152 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11153 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11155 *) Make sure `make rehash' target really finds the `openssl' program.
11156 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11158 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11159 like to hear about it if this slows down other processors.
11162 *) Add CygWin32 platform information to Configure script.
11163 [Alan Batie <batie@aahz.jf.intel.com>]
11165 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11166 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11168 *) New program nseq to manipulate netscape certificate sequences
11171 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11175 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11176 but the BN code had some problems that would cause failures when
11177 doing certificate verification and some other functions.
11178 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11180 *) Add ASN1 and PEM code to support netscape certificate sequences.
11183 *) Add ASN1 and PEM code to support netscape certificate sequences.
11186 *) Add several PKIX and private extended key usage OIDs.
11189 *) Modify the 'ca' program to handle the new extension code. Modify
11190 openssl.cnf for new extension format, add comments.
11193 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11194 and add a sample to openssl.cnf so req -x509 now adds appropriate
11198 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11199 error code, add initial support to X509_print() and x509 application.
11202 *) Takes a deep breath and start addding X509 V3 extension support code. Add
11203 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11204 stuff is currently isolated and isn't even compiled yet.
11207 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11208 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11209 Removed the versions check from X509 routines when loading extensions:
11210 this allows certain broken certificates that don't set the version
11211 properly to be processed.
11214 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11215 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11216 can still be regenerated with "make depend".
11219 *) Spelling mistake in C version of CAST-128.
11220 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11222 *) Changes to the error generation code. The perl script err-code.pl
11223 now reads in the old error codes and retains the old numbers, only
11224 adding new ones if necessary. It also only changes the .err files if new
11225 codes are added. The makefiles have been modified to only insert errors
11226 when needed (to avoid needlessly modifying header files). This is done
11227 by only inserting errors if the .err file is newer than the auto generated
11228 C file. To rebuild all the error codes from scratch (the old behaviour)
11229 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11230 or delete all the .err files.
11233 *) CAST-128 was incorrectly implemented for short keys. The C version has
11234 been fixed, but is untested. The assembler versions are also fixed, but
11235 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11236 to regenerate it if needed.
11237 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11238 Hagino <itojun@kame.net>]
11240 *) File was opened incorrectly in randfile.c.
11241 [Ulf Möller <ulf@fitug.de>]
11243 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11244 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11245 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11246 al: it's just almost always a UTCTime. Note this patch adds new error
11247 codes so do a "make errors" if there are problems.
11250 *) Correct Linux 1 recognition in config.
11251 [Ulf Möller <ulf@fitug.de>]
11253 *) Remove pointless MD5 hash when using DSA keys in ca.
11254 [Anonymous <nobody@replay.com>]
11256 *) Generate an error if given an empty string as a cert directory. Also
11257 generate an error if handed NULL (previously returned 0 to indicate an
11258 error, but didn't set one).
11259 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11261 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11264 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11265 parameters. This was causing a warning which killed off the Win32 compile.
11268 *) Remove C++ style comments from crypto/bn/bn_local.h.
11269 [Neil Costigan <neil.costigan@celocom.com>]
11271 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11272 based on a text string, looking up short and long names and finally
11273 "dot" format. The "dot" format stuff didn't work. Added new function
11274 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11275 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11276 OID is not part of the table.
11279 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11280 X509_LOOKUP_by_alias().
11283 *) Sort openssl functions by name.
11286 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11287 encryption from sample DSA keys (in case anyone is interested the password
11291 *) Make _all_ *_free functions accept a NULL pointer.
11292 [Frans Heymans <fheymans@isaserver.be>]
11294 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11296 [Anonymous <nobody@replay.com>]
11298 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11299 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11301 *) Don't blow it for numeric -newkey arguments to apps/req.
11302 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11304 *) Temp key "for export" tests were wrong in s3_srvr.c.
11305 [Anonymous <nobody@replay.com>]
11307 *) Add prototype for temp key callback functions
11308 SSL_CTX_set_tmp_{rsa,dh}_callback().
11311 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11312 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11315 *) X509_name_add_entry() freed the wrong thing after an error.
11316 [Arne Ansper <arne@ats.cyber.ee>]
11318 *) rsa_eay.c would attempt to free a NULL context.
11319 [Arne Ansper <arne@ats.cyber.ee>]
11321 *) BIO_s_socket() had a broken should_retry() on Windoze.
11322 [Arne Ansper <arne@ats.cyber.ee>]
11324 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11325 [Arne Ansper <arne@ats.cyber.ee>]
11327 *) Make sure the already existing X509_STORE->depth variable is initialized
11328 in X509_STORE_new(), but document the fact that this variable is still
11329 unused in the certificate verification process.
11330 [Ralf S. Engelschall]
11332 *) Fix the various library and apps files to free up pkeys obtained from
11333 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11336 *) Fix reference counting in X509_PUBKEY_get(). This makes
11337 demos/maurice/example2.c work, amongst others, probably.
11338 [Steve Henson and Ben Laurie]
11340 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11341 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11342 are no longer created. This way we have a single and consistent command
11343 line interface `openssl <command>', similar to `cvs <command>'.
11344 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11346 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11347 BIT STRING wrapper always have zero unused bits.
11350 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11353 *) Make the top-level INSTALL documentation easier to understand.
11356 *) Makefiles updated to exit if an error occurs in a sub-directory
11357 make (including if user presses ^C) [Paul Sutton]
11359 *) Make Montgomery context stuff explicit in RSA data structure.
11362 *) Fix build order of pem and err to allow for generated pem.h.
11365 *) Fix renumbering bug in X509_NAME_delete_entry().
11368 *) Enhanced the err-ins.pl script so it makes the error library number
11369 global and can add a library name. This is needed for external ASN1 and
11370 other error libraries.
11373 *) Fixed sk_insert which never worked properly.
11376 *) Fix ASN1 macros so they can handle indefinite length construted
11377 EXPLICIT tags. Some non standard certificates use these: they can now
11381 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11382 into a single doc/ssleay.txt bundle. This way the information is still
11383 preserved but no longer messes up this directory. Now it's new room for
11384 the new set of documenation files.
11385 [Ralf S. Engelschall]
11387 *) SETs were incorrectly DER encoded. This was a major pain, because they
11388 shared code with SEQUENCEs, which aren't coded the same. This means that
11389 almost everything to do with SETs or SEQUENCEs has either changed name or
11390 number of arguments.
11391 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11393 *) Fix test data to work with the above.
11396 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11397 was already fixed by Eric for 0.9.1 it seems.
11398 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11400 *) Autodetect FreeBSD3.
11403 *) Fix various bugs in Configure. This affects the following platforms:
11407 unixware-2.0-pentium
11411 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11412 before they are needed.
11415 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11419 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11421 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11422 changed SSLeay to OpenSSL in version strings.
11423 [Ralf S. Engelschall]
11425 *) Some fixups to the top-level documents.
11428 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11429 because the symlink to include/ was missing.
11430 [Ralf S. Engelschall]
11432 *) Incorporated the popular no-RSA/DSA-only patches
11433 which allow to compile a RSA-free SSLeay.
11434 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11436 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11437 when "ssleay" is still not found.
11438 [Ralf S. Engelschall]
11440 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11441 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11443 *) Updated the README file.
11444 [Ralf S. Engelschall]
11446 *) Added various .cvsignore files in the CVS repository subdirs
11447 to make a "cvs update" really silent.
11448 [Ralf S. Engelschall]
11450 *) Recompiled the error-definition header files and added
11451 missing symbols to the Win32 linker tables.
11452 [Ralf S. Engelschall]
11454 *) Cleaned up the top-level documents;
11455 o new files: CHANGES and LICENSE
11456 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11457 o merged COPYRIGHT into LICENSE
11458 o removed obsolete TODO file
11459 o renamed MICROSOFT to INSTALL.W32
11460 [Ralf S. Engelschall]
11462 *) Removed dummy files from the 0.9.1b source tree:
11463 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11464 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11465 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11466 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11467 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11468 [Ralf S. Engelschall]
11470 *) Added various platform portability fixes.
11473 *) The Genesis of the OpenSSL rpject:
11474 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11475 Young and Tim J. Hudson created while they were working for C2Net until
11477 [The OpenSSL Project]
11480 Changes between 0.9.0b and 0.9.1b [not released]
11482 *) Updated a few CA certificates under certs/
11485 *) Changed some BIGNUM api stuff.
11488 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11489 DGUX x86, Linux Alpha, etc.
11492 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11493 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11497 *) Add -strparse option to asn1pars program which parses nested
11499 [Dr Stephen Henson <shenson@bigfoot.com>]
11501 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11504 *) DSA fix for "ca" program.
11507 *) Added "-genkey" option to "dsaparam" program.
11510 *) Added RIPE MD160 (rmd160) message digest.
11513 *) Added -a (all) option to "ssleay version" command.
11516 *) Added PLATFORM define which is the id given to Configure.
11519 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11522 *) Extended the ASN.1 parser routines.
11525 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11528 *) Added a BN_CTX to the BN library.
11531 *) Fixed the weak key values in DES library
11534 *) Changed API in EVP library for cipher aliases.
11537 *) Added support for RC2/64bit cipher.
11540 *) Converted the lhash library to the crypto/mem.c functions.
11543 *) Added more recognized ASN.1 object ids.
11546 *) Added more RSA padding checks for SSL/TLS.
11549 *) Added BIO proxy/filter functionality.
11552 *) Added extra_certs to SSL_CTX which can be used
11553 send extra CA certificates to the client in the CA cert chain sending
11554 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11557 *) Now Fortezza is denied in the authentication phase because
11558 this is key exchange mechanism is not supported by SSLeay at all.
11561 *) Additional PKCS1 checks.
11564 *) Support the string "TLSv1" for all TLS v1 ciphers.
11567 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11568 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11571 *) Fixed a few memory leaks.
11574 *) Fixed various code and comment typos.
11577 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11578 bytes sent in the client random.
11579 [Edward Bishop <ebishop@spyglass.com>]