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 *) Client DoS due to large DH parameter
14 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
15 malicious server can send a very large prime value to the client. This will
16 cause the client to spend an unreasonably long period of time generating a
17 key for this prime resulting in a hang until the client has finished. This
18 could be exploited in a Denial Of Service attack.
20 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
24 *) Cache timing vulnerability in RSA Key Generation
26 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
27 a cache timing side channel attack. An attacker with sufficient access to
28 mount cache timing attacks during the RSA key generation process could
29 recover the private key.
31 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
32 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
36 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
37 parameter is no longer accepted, as it leads to a corrupt table. NULL
38 pem_str is reserved for alias entries only.
41 *) Revert blinding in ECDSA sign and instead make problematic addition
42 length-invariant. Switch even to fixed-length Montgomery multiplication.
45 *) Change generating and checking of primes so that the error rate of not
46 being prime depends on the intended use based on the size of the input.
47 For larger primes this will result in more rounds of Miller-Rabin.
48 The maximal error rate for primes with more than 1080 bits is lowered
50 [Kurt Roeckx, Annie Yousar]
52 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
55 *) Add blinding to ECDSA and DSA signatures to protect against side channel
56 attacks discovered by Keegan Ryan (NCC Group).
59 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
60 now allow empty (zero character) pass phrases.
63 *) Certificate time validation (X509_cmp_time) enforces stricter
64 compliance with RFC 5280. Fractional seconds and timezone offsets
65 are no longer allowed.
68 Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
70 *) Constructed ASN.1 types with a recursive definition could exceed the stack
72 Constructed ASN.1 types with a recursive definition (such as can be found
73 in PKCS7) could eventually exceed the stack given malicious input with
74 excessive recursion. This could result in a Denial Of Service attack. There
75 are no such structures used within SSL/TLS that come from untrusted sources
76 so this is considered safe.
78 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
83 Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
85 *) Read/write after SSL object in error state
87 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
88 mechanism. The intent was that if a fatal error occurred during a handshake
89 then OpenSSL would move into the error state and would immediately fail if
90 you attempted to continue the handshake. This works as designed for the
91 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
92 SSL_connect()), however due to a bug it does not work correctly if
93 SSL_read() or SSL_write() is called directly. In that scenario, if the
94 handshake fails then a fatal error will be returned in the initial function
95 call. If SSL_read()/SSL_write() is subsequently called by the application
96 for the same SSL object then it will succeed and the data is passed without
97 being decrypted/encrypted directly from the SSL/TLS record layer.
99 In order to exploit this issue an application bug would have to be present
100 that resulted in a call to SSL_read()/SSL_write() being issued after having
101 already received a fatal error.
103 This issue was reported to OpenSSL by David Benjamin (Google).
107 *) rsaz_1024_mul_avx2 overflow bug on x86_64
109 There is an overflow bug in the AVX2 Montgomery multiplication procedure
110 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
111 Analysis suggests that attacks against RSA and DSA as a result of this
112 defect would be very difficult to perform and are not believed likely.
113 Attacks against DH1024 are considered just feasible, because most of the
114 work necessary to deduce information about a private key may be performed
115 offline. The amount of resources required for such an attack would be
116 significant. However, for an attack on TLS to be meaningful, the server
117 would have to share the DH1024 private key among multiple clients, which is
118 no longer an option since CVE-2016-0701.
120 This only affects processors that support the AVX2 but not ADX extensions
121 like Intel Haswell (4th generation).
123 This issue was reported to OpenSSL by David Benjamin (Google). The issue
124 was originally found via the OSS-Fuzz project.
128 Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
130 *) bn_sqrx8x_internal carry bug on x86_64
132 There is a carry propagating bug in the x86_64 Montgomery squaring
133 procedure. No EC algorithms are affected. Analysis suggests that attacks
134 against RSA and DSA as a result of this defect would be very difficult to
135 perform and are not believed likely. Attacks against DH are considered just
136 feasible (although very difficult) because most of the work necessary to
137 deduce information about a private key may be performed offline. The amount
138 of resources required for such an attack would be very significant and
139 likely only accessible to a limited number of attackers. An attacker would
140 additionally need online access to an unpatched system using the target
141 private key in a scenario with persistent DH parameters and a private
142 key that is shared between multiple clients.
144 This only affects processors that support the BMI1, BMI2 and ADX extensions
145 like Intel Broadwell (5th generation) and later or AMD Ryzen.
147 This issue was reported to OpenSSL by the OSS-Fuzz project.
151 *) Malformed X.509 IPAddressFamily could cause OOB read
153 If an X.509 certificate has a malformed IPAddressFamily extension,
154 OpenSSL could do a one-byte buffer overread. The most likely result
155 would be an erroneous display of the certificate in text format.
157 This issue was reported to OpenSSL by the OSS-Fuzz project.
161 Changes between 1.0.2k and 1.0.2l [25 May 2017]
163 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
164 platform rather than 'mingw'.
167 Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
169 *) Truncated packet could crash via OOB read
171 If one side of an SSL/TLS path is running on a 32-bit host and a specific
172 cipher is being used, then a truncated packet can cause that host to
173 perform an out-of-bounds read, usually resulting in a crash.
175 This issue was reported to OpenSSL by Robert Święcki of Google.
179 *) BN_mod_exp may produce incorrect results on x86_64
181 There is a carry propagating bug in the x86_64 Montgomery squaring
182 procedure. No EC algorithms are affected. Analysis suggests that attacks
183 against RSA and DSA as a result of this defect would be very difficult to
184 perform and are not believed likely. Attacks against DH are considered just
185 feasible (although very difficult) because most of the work necessary to
186 deduce information about a private key may be performed offline. The amount
187 of resources required for such an attack would be very significant and
188 likely only accessible to a limited number of attackers. An attacker would
189 additionally need online access to an unpatched system using the target
190 private key in a scenario with persistent DH parameters and a private
191 key that is shared between multiple clients. For example this can occur by
192 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
193 similar to CVE-2015-3193 but must be treated as a separate problem.
195 This issue was reported to OpenSSL by the OSS-Fuzz project.
199 *) Montgomery multiplication may produce incorrect results
201 There is a carry propagating bug in the Broadwell-specific Montgomery
202 multiplication procedure that handles input lengths divisible by, but
203 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
204 and DH private keys are impossible. This is because the subroutine in
205 question is not used in operations with the private key itself and an input
206 of the attacker's direct choice. Otherwise the bug can manifest itself as
207 transient authentication and key negotiation failures or reproducible
208 erroneous outcome of public-key operations with specially crafted input.
209 Among EC algorithms only Brainpool P-512 curves are affected and one
210 presumably can attack ECDH key negotiation. Impact was not analyzed in
211 detail, because pre-requisites for attack are considered unlikely. Namely
212 multiple clients have to choose the curve in question and the server has to
213 share the private key among them, neither of which is default behaviour.
214 Even then only clients that chose the curve will be affected.
216 This issue was publicly reported as transient failures and was not
217 initially recognized as a security issue. Thanks to Richard Morgan for
218 providing reproducible case.
222 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
223 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
224 prevent issues where no progress is being made and the peer continually
225 sends unrecognised record types, using up resources processing them.
228 Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
230 *) Missing CRL sanity check
232 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
233 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
234 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
236 This issue only affects the OpenSSL 1.0.2i
240 Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
242 *) OCSP Status Request extension unbounded memory growth
244 A malicious client can send an excessively large OCSP Status Request
245 extension. If that client continually requests renegotiation, sending a
246 large OCSP Status Request extension each time, then there will be unbounded
247 memory growth on the server. This will eventually lead to a Denial Of
248 Service attack through memory exhaustion. Servers with a default
249 configuration are vulnerable even if they do not support OCSP. Builds using
250 the "no-ocsp" build time option are not affected.
252 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
256 *) In order to mitigate the SWEET32 attack, the DES ciphers were moved from
259 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
264 *) OOB write in MDC2_Update()
266 An overflow can occur in MDC2_Update() either if called directly or
267 through the EVP_DigestUpdate() function using MDC2. If an attacker
268 is able to supply very large amounts of input data after a previous
269 call to EVP_EncryptUpdate() with a partial block then a length check
270 can overflow resulting in a heap corruption.
272 The amount of data needed is comparable to SIZE_MAX which is impractical
275 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
279 *) Malformed SHA512 ticket DoS
281 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
282 DoS attack where a malformed ticket will result in an OOB read which will
285 The use of SHA512 in TLS session tickets is comparatively rare as it requires
286 a custom server callback and ticket lookup mechanism.
288 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
292 *) OOB write in BN_bn2dec()
294 The function BN_bn2dec() does not check the return value of BN_div_word().
295 This can cause an OOB write if an application uses this function with an
296 overly large BIGNUM. This could be a problem if an overly large certificate
297 or CRL is printed out from an untrusted source. TLS is not affected because
298 record limits will reject an oversized certificate before it is parsed.
300 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
304 *) OOB read in TS_OBJ_print_bio()
306 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
307 the total length the OID text representation would use and not the amount
308 of data written. This will result in OOB reads when large OIDs are
311 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
315 *) Pointer arithmetic undefined behaviour
317 Avoid some undefined pointer arithmetic
319 A common idiom in the codebase is to check limits in the following manner:
322 Where "p" points to some malloc'd data of SIZE bytes and
325 "len" here could be from some externally supplied data (e.g. from a TLS
328 The rules of C pointer arithmetic are such that "p + len" is only well
329 defined where len <= SIZE. Therefore the above idiom is actually
332 For example this could cause problems if some malloc implementation
333 provides an address for "p" such that "p + len" actually overflows for
334 values of len that are too big and therefore p + len < limit.
336 This issue was reported to OpenSSL by Guido Vranken
340 *) Constant time flag not preserved in DSA signing
342 Operations in the DSA signing algorithm should run in constant time in
343 order to avoid side channel attacks. A flaw in the OpenSSL DSA
344 implementation means that a non-constant time codepath is followed for
345 certain operations. This has been demonstrated through a cache-timing
346 attack to be sufficient for an attacker to recover the private DSA key.
348 This issue was reported by César Pereida (Aalto University), Billy Brumley
349 (Tampere University of Technology), and Yuval Yarom (The University of
354 *) DTLS buffered message DoS
356 In a DTLS connection where handshake messages are delivered out-of-order
357 those messages that OpenSSL is not yet ready to process will be buffered
358 for later use. Under certain circumstances, a flaw in the logic means that
359 those messages do not get removed from the buffer even though the handshake
360 has been completed. An attacker could force up to approx. 15 messages to
361 remain in the buffer when they are no longer required. These messages will
362 be cleared when the DTLS connection is closed. The default maximum size for
363 a message is 100k. Therefore the attacker could force an additional 1500k
364 to be consumed per connection. By opening many simulataneous connections an
365 attacker could cause a DoS attack through memory exhaustion.
367 This issue was reported to OpenSSL by Quan Luo.
371 *) DTLS replay protection DoS
373 A flaw in the DTLS replay attack protection mechanism means that records
374 that arrive for future epochs update the replay protection "window" before
375 the MAC for the record has been validated. This could be exploited by an
376 attacker by sending a record for the next epoch (which does not have to
377 decrypt or have a valid MAC), with a very large sequence number. This means
378 that all subsequent legitimate packets are dropped causing a denial of
379 service for a specific DTLS connection.
381 This issue was reported to OpenSSL by the OCAP audit team.
385 *) Certificate message OOB reads
387 In OpenSSL 1.0.2 and earlier some missing message length checks can result
388 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
389 theoretical DoS risk but this has not been observed in practice on common
392 The messages affected are client certificate, client certificate request
393 and server certificate. As a result the attack can only be performed
394 against a client or a server which enables client authentication.
396 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
400 Changes between 1.0.2g and 1.0.2h [3 May 2016]
402 *) Prevent padding oracle in AES-NI CBC MAC check
404 A MITM attacker can use a padding oracle attack to decrypt traffic
405 when the connection uses an AES CBC cipher and the server support
408 This issue was introduced as part of the fix for Lucky 13 padding
409 attack (CVE-2013-0169). The padding check was rewritten to be in
410 constant time by making sure that always the same bytes are read and
411 compared against either the MAC or padding bytes. But it no longer
412 checked that there was enough data to have both the MAC and padding
415 This issue was reported by Juraj Somorovsky using TLS-Attacker.
419 *) Fix EVP_EncodeUpdate overflow
421 An overflow can occur in the EVP_EncodeUpdate() function which is used for
422 Base64 encoding of binary data. If an attacker is able to supply very large
423 amounts of input data then a length check can overflow resulting in a heap
426 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
427 the PEM_write_bio* family of functions. These are mainly used within the
428 OpenSSL command line applications, so any application which processes data
429 from an untrusted source and outputs it as a PEM file should be considered
430 vulnerable to this issue. User applications that call these APIs directly
431 with large amounts of untrusted data may also be vulnerable.
433 This issue was reported by Guido Vranken.
437 *) Fix EVP_EncryptUpdate overflow
439 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
440 is able to supply very large amounts of input data after a previous call to
441 EVP_EncryptUpdate() with a partial block then a length check can overflow
442 resulting in a heap corruption. Following an analysis of all OpenSSL
443 internal usage of the EVP_EncryptUpdate() function all usage is one of two
444 forms. The first form is where the EVP_EncryptUpdate() call is known to be
445 the first called function after an EVP_EncryptInit(), and therefore that
446 specific call must be safe. The second form is where the length passed to
447 EVP_EncryptUpdate() can be seen from the code to be some small value and
448 therefore there is no possibility of an overflow. Since all instances are
449 one of these two forms, it is believed that there can be no overflows in
450 internal code due to this problem. It should be noted that
451 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
452 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
453 of these calls have also been analysed too and it is believed there are no
454 instances in internal usage where an overflow could occur.
456 This issue was reported by Guido Vranken.
460 *) Prevent ASN.1 BIO excessive memory allocation
462 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
463 a short invalid encoding can casuse allocation of large amounts of memory
464 potentially consuming excessive resources or exhausting memory.
466 Any application parsing untrusted data through d2i BIO functions is
467 affected. The memory based functions such as d2i_X509() are *not* affected.
468 Since the memory based functions are used by the TLS library, TLS
469 applications are not affected.
471 This issue was reported by Brian Carpenter.
477 ASN1 Strings that are over 1024 bytes can cause an overread in applications
478 using the X509_NAME_oneline() function on EBCDIC systems. This could result
479 in arbitrary stack data being returned in the buffer.
481 This issue was reported by Guido Vranken.
485 *) Modify behavior of ALPN to invoke callback after SNI/servername
486 callback, such that updates to the SSL_CTX affect ALPN.
489 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
493 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
494 methods are enabled and ssl2 is disabled the methods return NULL.
497 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
499 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
500 Builds that are not configured with "enable-weak-ssl-ciphers" will not
501 provide any "EXPORT" or "LOW" strength ciphers.
504 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
505 is by default disabled at build-time. Builds that are not configured with
506 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
507 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
508 will need to explicitly call either of:
510 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
512 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
514 as appropriate. Even if either of those is used, or the application
515 explicitly uses the version-specific SSLv2_method() or its client and
516 server variants, SSLv2 ciphers vulnerable to exhaustive search key
517 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
518 ciphers, and SSLv2 56-bit DES are no longer available.
522 *) Fix a double-free in DSA code
524 A double free bug was discovered when OpenSSL parses malformed DSA private
525 keys and could lead to a DoS attack or memory corruption for applications
526 that receive DSA private keys from untrusted sources. This scenario is
529 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
534 *) Disable SRP fake user seed to address a server memory leak.
536 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
538 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
539 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
540 was changed to ignore the "fake user" SRP seed, even if the seed
543 Users should use SRP_VBASE_get1_by_user instead. Note that in
544 SRP_VBASE_get1_by_user, caller must free the returned value. Note
545 also that even though configuring the SRP seed attempts to hide
546 invalid usernames by continuing the handshake with fake
547 credentials, this behaviour is not constant time and no strong
548 guarantees are made that the handshake is indistinguishable from
549 that of a valid user.
553 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
555 In the BN_hex2bn function the number of hex digits is calculated using an
556 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
557 large values of |i| this can result in |bn_expand| not allocating any
558 memory because |i * 4| is negative. This can leave the internal BIGNUM data
559 field as NULL leading to a subsequent NULL ptr deref. For very large values
560 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
561 In this case memory is allocated to the internal BIGNUM data field, but it
562 is insufficiently sized leading to heap corruption. A similar issue exists
563 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
564 is ever called by user applications with very large untrusted hex/dec data.
565 This is anticipated to be a rare occurrence.
567 All OpenSSL internal usage of these functions use data that is not expected
568 to be untrusted, e.g. config file data or application command line
569 arguments. If user developed applications generate config file data based
570 on untrusted data then it is possible that this could also lead to security
571 consequences. This is also anticipated to be rare.
573 This issue was reported to OpenSSL by Guido Vranken.
577 *) Fix memory issues in BIO_*printf functions
579 The internal |fmtstr| function used in processing a "%s" format string in
580 the BIO_*printf functions could overflow while calculating the length of a
581 string and cause an OOB read when printing very long strings.
583 Additionally the internal |doapr_outch| function can attempt to write to an
584 OOB memory location (at an offset from the NULL pointer) in the event of a
585 memory allocation failure. In 1.0.2 and below this could be caused where
586 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
587 could be in processing a very long "%s" format string. Memory leaks can
590 The first issue may mask the second issue dependent on compiler behaviour.
591 These problems could enable attacks where large amounts of untrusted data
592 is passed to the BIO_*printf functions. If applications use these functions
593 in this way then they could be vulnerable. OpenSSL itself uses these
594 functions when printing out human-readable dumps of ASN.1 data. Therefore
595 applications that print this data could be vulnerable if the data is from
596 untrusted sources. OpenSSL command line applications could also be
597 vulnerable where they print out ASN.1 data, or if untrusted data is passed
598 as command line arguments.
600 Libssl is not considered directly vulnerable. Additionally certificates etc
601 received via remote connections via libssl are also unlikely to be able to
602 trigger these issues because of message size limits enforced within libssl.
604 This issue was reported to OpenSSL Guido Vranken.
608 *) Side channel attack on modular exponentiation
610 A side-channel attack was found which makes use of cache-bank conflicts on
611 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
612 of RSA keys. The ability to exploit this issue is limited as it relies on
613 an attacker who has control of code in a thread running on the same
614 hyper-threaded core as the victim thread which is performing decryptions.
616 This issue was reported to OpenSSL by Yuval Yarom, The University of
617 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
618 Nadia Heninger, University of Pennsylvania with more information at
619 http://cachebleed.info.
623 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
624 if no keysize is specified with default_bits. This fixes an
625 omission in an earlier change that changed all RSA/DSA key generation
626 apps to use 2048 bits by default.
629 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
631 *) DH small subgroups
633 Historically OpenSSL only ever generated DH parameters based on "safe"
634 primes. More recently (in version 1.0.2) support was provided for
635 generating X9.42 style parameter files such as those required for RFC 5114
636 support. The primes used in such files may not be "safe". Where an
637 application is using DH configured with parameters based on primes that are
638 not "safe" then an attacker could use this fact to find a peer's private
639 DH exponent. This attack requires that the attacker complete multiple
640 handshakes in which the peer uses the same private DH exponent. For example
641 this could be used to discover a TLS server's private DH exponent if it's
642 reusing the private DH exponent or it's using a static DH ciphersuite.
644 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
645 TLS. It is not on by default. If the option is not set then the server
646 reuses the same private DH exponent for the life of the server process and
647 would be vulnerable to this attack. It is believed that many popular
648 applications do set this option and would therefore not be at risk.
650 The fix for this issue adds an additional check where a "q" parameter is
651 available (as is the case in X9.42 based parameters). This detects the
652 only known attack, and is the only possible defense for static DH
653 ciphersuites. This could have some performance impact.
655 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
656 default and cannot be disabled. This could have some performance impact.
658 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
662 *) SSLv2 doesn't block disabled ciphers
664 A malicious client can negotiate SSLv2 ciphers that have been disabled on
665 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
666 been disabled, provided that the SSLv2 protocol was not also disabled via
669 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
670 and Sebastian Schinzel.
674 *) Reject DH handshakes with parameters shorter than 1024 bits.
677 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
679 *) BN_mod_exp may produce incorrect results on x86_64
681 There is a carry propagating bug in the x86_64 Montgomery squaring
682 procedure. No EC algorithms are affected. Analysis suggests that attacks
683 against RSA and DSA as a result of this defect would be very difficult to
684 perform and are not believed likely. Attacks against DH are considered just
685 feasible (although very difficult) because most of the work necessary to
686 deduce information about a private key may be performed offline. The amount
687 of resources required for such an attack would be very significant and
688 likely only accessible to a limited number of attackers. An attacker would
689 additionally need online access to an unpatched system using the target
690 private key in a scenario with persistent DH parameters and a private
691 key that is shared between multiple clients. For example this can occur by
692 default in OpenSSL DHE based SSL/TLS ciphersuites.
694 This issue was reported to OpenSSL by Hanno Böck.
698 *) Certificate verify crash with missing PSS parameter
700 The signature verification routines will crash with a NULL pointer
701 dereference if presented with an ASN.1 signature using the RSA PSS
702 algorithm and absent mask generation function parameter. Since these
703 routines are used to verify certificate signature algorithms this can be
704 used to crash any certificate verification operation and exploited in a
705 DoS attack. Any application which performs certificate verification is
706 vulnerable including OpenSSL clients and servers which enable client
709 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
713 *) X509_ATTRIBUTE memory leak
715 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
716 memory. This structure is used by the PKCS#7 and CMS routines so any
717 application which reads PKCS#7 or CMS data from untrusted sources is
718 affected. SSL/TLS is not affected.
720 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
725 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
726 This changes the decoding behaviour for some invalid messages,
727 though the change is mostly in the more lenient direction, and
728 legacy behaviour is preserved as much as possible.
731 *) In DSA_generate_parameters_ex, if the provided seed is too short,
732 use a random seed, as already documented.
733 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
735 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
737 *) Alternate chains certificate forgery
739 During certificate verfification, OpenSSL will attempt to find an
740 alternative certificate chain if the first attempt to build such a chain
741 fails. An error in the implementation of this logic can mean that an
742 attacker could cause certain checks on untrusted certificates to be
743 bypassed, such as the CA flag, enabling them to use a valid leaf
744 certificate to act as a CA and "issue" an invalid certificate.
746 This issue was reported to OpenSSL by Adam Langley/David Benjamin
751 *) Race condition handling PSK identify hint
753 If PSK identity hints are received by a multi-threaded client then
754 the values are wrongly updated in the parent SSL_CTX structure. This can
755 result in a race condition potentially leading to a double free of the
760 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
762 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
763 incompatibility in the handling of HMAC. The previous ABI has now been
766 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
768 *) Malformed ECParameters causes infinite loop
770 When processing an ECParameters structure OpenSSL enters an infinite loop
771 if the curve specified is over a specially malformed binary polynomial
774 This can be used to perform denial of service against any
775 system which processes public keys, certificate requests or
776 certificates. This includes TLS clients and TLS servers with
777 client authentication enabled.
779 This issue was reported to OpenSSL by Joseph Barr-Pixton.
783 *) Exploitable out-of-bounds read in X509_cmp_time
785 X509_cmp_time does not properly check the length of the ASN1_TIME
786 string and can read a few bytes out of bounds. In addition,
787 X509_cmp_time accepts an arbitrary number of fractional seconds in the
790 An attacker can use this to craft malformed certificates and CRLs of
791 various sizes and potentially cause a segmentation fault, resulting in
792 a DoS on applications that verify certificates or CRLs. TLS clients
793 that verify CRLs are affected. TLS clients and servers with client
794 authentication enabled may be affected if they use custom verification
797 This issue was reported to OpenSSL by Robert Swiecki (Google), and
798 independently by Hanno Böck.
802 *) PKCS7 crash with missing EnvelopedContent
804 The PKCS#7 parsing code does not handle missing inner EncryptedContent
805 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
806 with missing content and trigger a NULL pointer dereference on parsing.
808 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
809 structures from untrusted sources are affected. OpenSSL clients and
810 servers are not affected.
812 This issue was reported to OpenSSL by Michal Zalewski (Google).
816 *) CMS verify infinite loop with unknown hash function
818 When verifying a signedData message the CMS code can enter an infinite loop
819 if presented with an unknown hash function OID. This can be used to perform
820 denial of service against any system which verifies signedData messages using
822 This issue was reported to OpenSSL by Johannes Bauer.
826 *) Race condition handling NewSessionTicket
828 If a NewSessionTicket is received by a multi-threaded client when attempting to
829 reuse a previous ticket then a race condition can occur potentially leading to
830 a double free of the ticket data.
834 *) Removed support for the two export grade static DH ciphersuites
835 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
836 were newly added (along with a number of other static DH ciphersuites) to
837 1.0.2. However the two export ones have *never* worked since they were
838 introduced. It seems strange in any case to be adding new export
839 ciphersuites, and given "logjam" it also does not seem correct to fix them.
842 *) Only support 256-bit or stronger elliptic curves with the
843 'ecdh_auto' setting (server) or by default (client). Of supported
844 curves, prefer P-256 (both).
847 *) Reject DH handshakes with parameters shorter than 768 bits.
848 [Kurt Roeckx and Emilia Kasper]
850 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
852 *) ClientHello sigalgs DoS fix
854 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
855 invalid signature algorithms extension a NULL pointer dereference will
856 occur. This can be exploited in a DoS attack against the server.
858 This issue was was reported to OpenSSL by David Ramos of Stanford
861 [Stephen Henson and Matt Caswell]
863 *) Multiblock corrupted pointer fix
865 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
866 feature only applies on 64 bit x86 architecture platforms that support AES
867 NI instructions. A defect in the implementation of "multiblock" can cause
868 OpenSSL's internal write buffer to become incorrectly set to NULL when
869 using non-blocking IO. Typically, when the user application is using a
870 socket BIO for writing, this will only result in a failed connection.
871 However if some other BIO is used then it is likely that a segmentation
872 fault will be triggered, thus enabling a potential DoS attack.
874 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
878 *) Segmentation fault in DTLSv1_listen fix
880 The DTLSv1_listen function is intended to be stateless and processes the
881 initial ClientHello from many peers. It is common for user code to loop
882 over the call to DTLSv1_listen until a valid ClientHello is received with
883 an associated cookie. A defect in the implementation of DTLSv1_listen means
884 that state is preserved in the SSL object from one invocation to the next
885 that can lead to a segmentation fault. Errors processing the initial
886 ClientHello can trigger this scenario. An example of such an error could be
887 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
890 This issue was reported to OpenSSL by Per Allansson.
894 *) Segmentation fault in ASN1_TYPE_cmp fix
896 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
897 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
898 certificate signature algorithm consistency this can be used to crash any
899 certificate verification operation and exploited in a DoS attack. Any
900 application which performs certificate verification is vulnerable including
901 OpenSSL clients and servers which enable client authentication.
905 *) Segmentation fault for invalid PSS parameters fix
907 The signature verification routines will crash with a NULL pointer
908 dereference if presented with an ASN.1 signature using the RSA PSS
909 algorithm and invalid parameters. Since these routines are used to verify
910 certificate signature algorithms this can be used to crash any
911 certificate verification operation and exploited in a DoS attack. Any
912 application which performs certificate verification is vulnerable including
913 OpenSSL clients and servers which enable client authentication.
915 This issue was was reported to OpenSSL by Brian Carpenter.
919 *) ASN.1 structure reuse memory corruption fix
921 Reusing a structure in ASN.1 parsing may allow an attacker to cause
922 memory corruption via an invalid write. Such reuse is and has been
923 strongly discouraged and is believed to be rare.
925 Applications that parse structures containing CHOICE or ANY DEFINED BY
926 components may be affected. Certificate parsing (d2i_X509 and related
927 functions) are however not affected. OpenSSL clients and servers are
932 *) PKCS7 NULL pointer dereferences fix
934 The PKCS#7 parsing code does not handle missing outer ContentInfo
935 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
936 missing content and trigger a NULL pointer dereference on parsing.
938 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
939 otherwise parse PKCS#7 structures from untrusted sources are
940 affected. OpenSSL clients and servers are not affected.
942 This issue was reported to OpenSSL by Michal Zalewski (Google).
946 *) DoS via reachable assert in SSLv2 servers fix
948 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
949 servers that both support SSLv2 and enable export cipher suites by sending
950 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
952 This issue was discovered by Sean Burford (Google) and Emilia Käsper
953 (OpenSSL development team).
957 *) Empty CKE with client auth and DHE fix
959 If client auth is used then a server can seg fault in the event of a DHE
960 ciphersuite being selected and a zero length ClientKeyExchange message
961 being sent by the client. This could be exploited in a DoS attack.
965 *) Handshake with unseeded PRNG fix
967 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
968 with an unseeded PRNG. The conditions are:
969 - The client is on a platform where the PRNG has not been seeded
970 automatically, and the user has not seeded manually
971 - A protocol specific client method version has been used (i.e. not
972 SSL_client_methodv23)
973 - A ciphersuite is used that does not require additional random data from
974 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
976 If the handshake succeeds then the client random that has been used will
977 have been generated from a PRNG with insufficient entropy and therefore the
978 output may be predictable.
980 For example using the following command with an unseeded openssl will
981 succeed on an unpatched platform:
983 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
987 *) Use After Free following d2i_ECPrivatekey error fix
989 A malformed EC private key file consumed via the d2i_ECPrivateKey function
990 could cause a use after free condition. This, in turn, could cause a double
991 free in several private key parsing functions (such as d2i_PrivateKey
992 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
993 for applications that receive EC private keys from untrusted
994 sources. This scenario is considered rare.
996 This issue was discovered by the BoringSSL project and fixed in their
1001 *) X509_to_X509_REQ NULL pointer deref fix
1003 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1004 the certificate key is invalid. This function is rarely used in practice.
1006 This issue was discovered by Brian Carpenter.
1010 *) Removed the export ciphers from the DEFAULT ciphers
1013 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1015 *) Change RSA and DH/DSA key generation apps to generate 2048-bit
1019 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1020 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1021 So far those who have to target multiple plaforms would compromise
1022 and argue that binary targeting say ARMv5 would still execute on
1023 ARMv8. "Universal" build resolves this compromise by providing
1024 near-optimal performance even on newer platforms.
1027 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1028 (other platforms pending).
1029 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1031 *) Add support for the SignedCertificateTimestampList certificate and
1032 OCSP response extensions from RFC6962.
1035 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1036 for corner cases. (Certain input points at infinity could lead to
1037 bogus results, with non-infinity inputs mapped to infinity too.)
1040 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1041 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1042 common cases are optimized and there still is room for further
1043 improvements. Vector Permutation AES for Altivec is also added.
1046 *) Add support for little-endian ppc64 Linux target.
1047 [Marcelo Cerri (IBM)]
1049 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1050 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1051 are optimized and there still is room for further improvements.
1052 Both 32- and 64-bit modes are supported.
1053 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1055 *) Improved ARMv7 NEON support.
1058 *) Support for SPARC Architecture 2011 crypto extensions, first
1059 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1060 SHA256/512, MD5, GHASH and modular exponentiation.
1061 [Andy Polyakov, David Miller]
1063 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1065 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1067 *) Support for new and upcoming Intel processors, including AVX2,
1068 BMI and SHA ISA extensions. This includes additional "stitched"
1069 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1072 This work was sponsored by Intel Corp.
1075 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1076 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1077 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1080 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1081 this fixes a limiation in previous versions of OpenSSL.
1084 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1085 MGF1 digest and OAEP label.
1088 *) Add EVP support for key wrapping algorithms, to avoid problems with
1089 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1090 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1091 algorithms and include tests cases.
1094 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1096 [Douglas E. Engert, Steve Henson]
1098 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1099 difference in days and seconds between two tm or ASN1_TIME structures.
1102 *) Add -rev test option to s_server to just reverse order of characters
1103 received by client and send back to server. Also prints an abbreviated
1104 summary of the connection parameters.
1107 *) New option -brief for s_client and s_server to print out a brief summary
1108 of connection parameters.
1111 *) Add callbacks for arbitrary TLS extensions.
1112 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1114 *) New option -crl_download in several openssl utilities to download CRLs
1115 from CRLDP extension in certificates.
1118 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1121 *) New function X509_CRL_diff to generate a delta CRL from the difference
1122 of two full CRLs. Add support to "crl" utility.
1125 *) New functions to set lookup_crls function and to retrieve
1126 X509_STORE from X509_STORE_CTX.
1129 *) Print out deprecated issuer and subject unique ID fields in
1133 *) Extend OCSP I/O functions so they can be used for simple general purpose
1134 HTTP as well as OCSP. New wrapper function which can be used to download
1135 CRLs using the OCSP API.
1138 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1141 *) SSL_CONF* functions. These provide a common framework for application
1142 configuration using configuration files or command lines.
1145 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1146 message callback and prints the results. Needs compile time option
1147 "enable-ssl-trace". New options to s_client and s_server to enable
1151 *) New ctrl and macro to retrieve supported points extensions.
1152 Print out extension in s_server and s_client.
1155 *) New functions to retrieve certificate signature and signature
1159 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1163 *) New Suite B modes for TLS code. These use and enforce the requirements
1164 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1165 only use Suite B curves. The Suite B modes can be set by using the
1166 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1169 *) New chain verification flags for Suite B levels of security. Check
1170 algorithms are acceptable when flags are set in X509_verify_cert.
1173 *) Make tls1_check_chain return a set of flags indicating checks passed
1174 by a certificate chain. Add additional tests to handle client
1175 certificates: checks for matching certificate type and issuer name
1179 *) If an attempt is made to use a signature algorithm not in the peer
1180 preference list abort the handshake. If client has no suitable
1181 signature algorithms in response to a certificate request do not
1182 use the certificate.
1185 *) If server EC tmp key is not in client preference list abort handshake.
1188 *) Add support for certificate stores in CERT structure. This makes it
1189 possible to have different stores per SSL structure or one store in
1190 the parent SSL_CTX. Include distint stores for certificate chain
1191 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1192 to build and store a certificate chain in CERT structure: returing
1193 an error if the chain cannot be built: this will allow applications
1194 to test if a chain is correctly configured.
1196 Note: if the CERT based stores are not set then the parent SSL_CTX
1197 store is used to retain compatibility with existing behaviour.
1201 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1202 mask based on the current session, check mask when sending client
1203 hello and checking the requested ciphersuite.
1206 *) New ctrls to retrieve and set certificate types in a certificate
1207 request message. Print out received values in s_client. If certificate
1208 types is not set with custom values set sensible values based on
1209 supported signature algorithms.
1212 *) Support for distinct client and server supported signature algorithms.
1215 *) Add certificate callback. If set this is called whenever a certificate
1216 is required by client or server. An application can decide which
1217 certificate chain to present based on arbitrary criteria: for example
1218 supported signature algorithms. Add very simple example to s_server.
1219 This fixes many of the problems and restrictions of the existing client
1220 certificate callback: for example you can now clear an existing
1221 certificate and specify the whole chain.
1224 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1225 the certificate can be used for (if anything). Set valid_flags field
1226 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1227 to have similar checks in it.
1229 Add new "cert_flags" field to CERT structure and include a "strict mode".
1230 This enforces some TLS certificate requirements (such as only permitting
1231 certificate signature algorithms contained in the supported algorithms
1232 extension) which some implementations ignore: this option should be used
1233 with caution as it could cause interoperability issues.
1236 *) Update and tidy signature algorithm extension processing. Work out
1237 shared signature algorithms based on preferences and peer algorithms
1238 and print them out in s_client and s_server. Abort handshake if no
1239 shared signature algorithms.
1242 *) Add new functions to allow customised supported signature algorithms
1243 for SSL and SSL_CTX structures. Add options to s_client and s_server
1247 *) New function SSL_certs_clear() to delete all references to certificates
1248 from an SSL structure. Before this once a certificate had been added
1249 it couldn't be removed.
1252 *) Integrate hostname, email address and IP address checking with certificate
1253 verification. New verify options supporting checking in opensl utility.
1256 *) Fixes and wildcard matching support to hostname and email checking
1257 functions. Add manual page.
1258 [Florian Weimer (Red Hat Product Security Team)]
1260 *) New functions to check a hostname email or IP address against a
1261 certificate. Add options x509 utility to print results of checks against
1265 *) Fix OCSP checking.
1266 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1268 *) Initial experimental support for explicitly trusted non-root CAs.
1269 OpenSSL still tries to build a complete chain to a root but if an
1270 intermediate CA has a trust setting included that is used. The first
1271 setting is used: whether to trust (e.g., -addtrust option to the x509
1275 *) Add -trusted_first option which attempts to find certificates in the
1276 trusted store even if an untrusted chain is also supplied.
1279 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1280 platform support for Linux and Android.
1283 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1286 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1287 When in FIPS mode the approved implementations are used as normal,
1288 when not in FIPS mode the internal unapproved versions are used instead.
1289 This means that the FIPS capable OpenSSL isn't forced to use the
1290 (often lower perfomance) FIPS implementations outside FIPS mode.
1293 *) Transparently support X9.42 DH parameters when calling
1294 PEM_read_bio_DHparameters. This means existing applications can handle
1295 the new parameter format automatically.
1298 *) Initial experimental support for X9.42 DH parameter format: mainly
1299 to support use of 'q' parameter for RFC5114 parameters.
1302 *) Add DH parameters from RFC5114 including test data to dhtest.
1305 *) Support for automatic EC temporary key parameter selection. If enabled
1306 the most preferred EC parameters are automatically used instead of
1307 hardcoded fixed parameters. Now a server just has to call:
1308 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1309 support ECDH and use the most appropriate parameters.
1312 *) Enhance and tidy EC curve and point format TLS extension code. Use
1313 static structures instead of allocation if default values are used.
1314 New ctrls to set curves we wish to support and to retrieve shared curves.
1315 Print out shared curves in s_server. New options to s_server and s_client
1316 to set list of supported curves.
1319 *) New ctrls to retrieve supported signature algorithms and
1320 supported curve values as an array of NIDs. Extend openssl utility
1321 to print out received values.
1324 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1325 between NIDs and the more common NIST names such as "P-256". Enhance
1326 ecparam utility and ECC method to recognise the NIST names for curves.
1329 *) Enhance SSL/TLS certificate chain handling to support different
1330 chains for each certificate instead of one chain in the parent SSL_CTX.
1333 *) Support for fixed DH ciphersuite client authentication: where both
1334 server and client use DH certificates with common parameters.
1337 *) Support for fixed DH ciphersuites: those requiring DH server
1341 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1343 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1344 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1345 X509_CINF_get_signature were reverted post internal team review.
1347 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1349 *) Build fixes for the Windows and OpenVMS platforms
1350 [Matt Caswell and Richard Levitte]
1352 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1354 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1355 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1356 dereference. This could lead to a Denial Of Service attack. Thanks to
1357 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1361 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1362 dtls1_buffer_record function under certain conditions. In particular this
1363 could occur if an attacker sent repeated DTLS records with the same
1364 sequence number but for the next epoch. The memory leak could be exploited
1365 by an attacker in a Denial of Service attack through memory exhaustion.
1366 Thanks to Chris Mueller for reporting this issue.
1370 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1371 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1372 method would be set to NULL which could later result in a NULL pointer
1373 dereference. Thanks to Frank Schmirler for reporting this issue.
1377 *) Abort handshake if server key exchange message is omitted for ephemeral
1380 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1381 reporting this issue.
1385 *) Remove non-export ephemeral RSA code on client and server. This code
1386 violated the TLS standard by allowing the use of temporary RSA keys in
1387 non-export ciphersuites and could be used by a server to effectively
1388 downgrade the RSA key length used to a value smaller than the server
1389 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1390 INRIA or reporting this issue.
1394 *) Fixed issue where DH client certificates are accepted without verification.
1395 An OpenSSL server will accept a DH certificate for client authentication
1396 without the certificate verify message. This effectively allows a client to
1397 authenticate without the use of a private key. This only affects servers
1398 which trust a client certificate authority which issues certificates
1399 containing DH keys: these are extremely rare and hardly ever encountered.
1400 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1405 *) Ensure that the session ID context of an SSL is updated when its
1406 SSL_CTX is updated via SSL_set_SSL_CTX.
1408 The session ID context is typically set from the parent SSL_CTX,
1409 and can vary with the CTX.
1412 *) Fix various certificate fingerprint issues.
1414 By using non-DER or invalid encodings outside the signed portion of a
1415 certificate the fingerprint can be changed without breaking the signature.
1416 Although no details of the signed portion of the certificate can be changed
1417 this can cause problems with some applications: e.g. those using the
1418 certificate fingerprint for blacklists.
1420 1. Reject signatures with non zero unused bits.
1422 If the BIT STRING containing the signature has non zero unused bits reject
1423 the signature. All current signature algorithms require zero unused bits.
1425 2. Check certificate algorithm consistency.
1427 Check the AlgorithmIdentifier inside TBS matches the one in the
1428 certificate signature. NB: this will result in signature failure
1429 errors for some broken certificates.
1431 Thanks to Konrad Kraszewski from Google for reporting this issue.
1433 3. Check DSA/ECDSA signatures use DER.
1435 Reencode DSA/ECDSA signatures and compare with the original received
1436 signature. Return an error if there is a mismatch.
1438 This will reject various cases including garbage after signature
1439 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1440 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1441 (negative or with leading zeroes).
1443 Further analysis was conducted and fixes were developed by Stephen Henson
1444 of the OpenSSL core team.
1449 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1450 results on some platforms, including x86_64. This bug occurs at random
1451 with a very low probability, and is not known to be exploitable in any
1452 way, though its exact impact is difficult to determine. Thanks to Pieter
1453 Wuille (Blockstream) who reported this issue and also suggested an initial
1454 fix. Further analysis was conducted by the OpenSSL development team and
1455 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1456 the OpenSSL core team.
1460 *) Do not resume sessions on the server if the negotiated protocol
1461 version does not match the session's version. Resuming with a different
1462 version, while not strictly forbidden by the RFC, is of questionable
1463 sanity and breaks all known clients.
1464 [David Benjamin, Emilia Käsper]
1466 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1467 early CCS messages during renegotiation. (Note that because
1468 renegotiation is encrypted, this early CCS was not exploitable.)
1471 *) Tighten client-side session ticket handling during renegotiation:
1472 ensure that the client only accepts a session ticket if the server sends
1473 the extension anew in the ServerHello. Previously, a TLS client would
1474 reuse the old extension state and thus accept a session ticket if one was
1475 announced in the initial ServerHello.
1477 Similarly, ensure that the client requires a session ticket if one
1478 was advertised in the ServerHello. Previously, a TLS client would
1479 ignore a missing NewSessionTicket message.
1482 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1484 *) SRTP Memory Leak.
1486 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1487 sends a carefully crafted handshake message, to cause OpenSSL to fail
1488 to free up to 64k of memory causing a memory leak. This could be
1489 exploited in a Denial Of Service attack. This issue affects OpenSSL
1490 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1491 whether SRTP is used or configured. Implementations of OpenSSL that
1492 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1494 The fix was developed by the OpenSSL team.
1498 *) Session Ticket Memory Leak.
1500 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1501 integrity of that ticket is first verified. In the event of a session
1502 ticket integrity check failing, OpenSSL will fail to free memory
1503 causing a memory leak. By sending a large number of invalid session
1504 tickets an attacker could exploit this issue in a Denial Of Service
1509 *) Build option no-ssl3 is incomplete.
1511 When OpenSSL is configured with "no-ssl3" as a build option, servers
1512 could accept and complete a SSL 3.0 handshake, and clients could be
1513 configured to send them.
1515 [Akamai and the OpenSSL team]
1517 *) Add support for TLS_FALLBACK_SCSV.
1518 Client applications doing fallback retries should call
1519 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1521 [Adam Langley, Bodo Moeller]
1523 *) Add additional DigestInfo checks.
1525 Reencode DigestInto in DER and check against the original when
1526 verifying RSA signature: this will reject any improperly encoded
1527 DigestInfo structures.
1529 Note: this is a precautionary measure and no attacks are currently known.
1533 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1535 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1536 SRP code can be overrun an internal buffer. Add sanity check that
1537 g, A, B < N to SRP code.
1539 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1540 Group for discovering this issue.
1544 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1545 TLS 1.0 instead of higher protocol versions when the ClientHello message
1546 is badly fragmented. This allows a man-in-the-middle attacker to force a
1547 downgrade to TLS 1.0 even if both the server and the client support a
1548 higher protocol version, by modifying the client's TLS records.
1550 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1551 researching this issue.
1555 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1556 to a denial of service attack. A malicious server can crash the client
1557 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1558 ciphersuite and sending carefully crafted handshake messages.
1560 Thanks to Felix Gröbert (Google) for discovering and researching this
1565 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1566 to leak memory. This can be exploited through a Denial of Service attack.
1567 Thanks to Adam Langley for discovering and researching this issue.
1571 *) An attacker can force openssl to consume large amounts of memory whilst
1572 processing DTLS handshake messages. This can be exploited through a
1573 Denial of Service attack.
1574 Thanks to Adam Langley for discovering and researching this issue.
1578 *) An attacker can force an error condition which causes openssl to crash
1579 whilst processing DTLS packets due to memory being freed twice. This
1580 can be exploited through a Denial of Service attack.
1581 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1586 *) If a multithreaded client connects to a malicious server using a resumed
1587 session and the server sends an ec point format extension it could write
1588 up to 255 bytes to freed memory.
1590 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1595 *) A malicious server can crash an OpenSSL client with a null pointer
1596 dereference (read) by specifying an SRP ciphersuite even though it was not
1597 properly negotiated with the client. This can be exploited through a
1598 Denial of Service attack.
1600 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1601 discovering and researching this issue.
1605 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1606 X509_name_oneline, X509_name_print_ex et al. to leak some information
1607 from the stack. Applications may be affected if they echo pretty printing
1608 output to the attacker.
1610 Thanks to Ivan Fratric (Google) for discovering this issue.
1612 [Emilia Käsper, and Steve Henson]
1614 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1615 for corner cases. (Certain input points at infinity could lead to
1616 bogus results, with non-infinity inputs mapped to infinity too.)
1619 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1621 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1622 handshake can force the use of weak keying material in OpenSSL
1623 SSL/TLS clients and servers.
1625 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1626 researching this issue. (CVE-2014-0224)
1627 [KIKUCHI Masashi, Steve Henson]
1629 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1630 OpenSSL DTLS client the code can be made to recurse eventually crashing
1633 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1635 [Imre Rad, Steve Henson]
1637 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1638 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1639 client or server. This is potentially exploitable to run arbitrary
1640 code on a vulnerable client or server.
1642 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1643 [Jüri Aedla, Steve Henson]
1645 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1646 are subject to a denial of service attack.
1648 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1649 this issue. (CVE-2014-3470)
1650 [Felix Gröbert, Ivan Fratric, Steve Henson]
1652 *) Harmonize version and its documentation. -f flag is used to display
1654 [mancha <mancha1@zoho.com>]
1656 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1657 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1658 [mancha <mancha1@zoho.com>]
1660 *) Fix some double frees. These are not thought to be exploitable.
1661 [mancha <mancha1@zoho.com>]
1663 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1665 *) A missing bounds check in the handling of the TLS heartbeat extension
1666 can be used to reveal up to 64k of memory to a connected client or
1669 Thanks for Neel Mehta of Google Security for discovering this bug and to
1670 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1671 preparing the fix (CVE-2014-0160)
1672 [Adam Langley, Bodo Moeller]
1674 *) Fix for the attack described in the paper "Recovering OpenSSL
1675 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1676 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1677 http://eprint.iacr.org/2014/140
1679 Thanks to Yuval Yarom and Naomi Benger for discovering this
1680 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1681 [Yuval Yarom and Naomi Benger]
1683 *) TLS pad extension: draft-agl-tls-padding-03
1685 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1686 TLS client Hello record length value would otherwise be > 255 and
1687 less that 512 pad with a dummy extension containing zeroes so it
1688 is at least 512 bytes long.
1690 [Adam Langley, Steve Henson]
1692 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1694 *) Fix for TLS record tampering bug. A carefully crafted invalid
1695 handshake could crash OpenSSL with a NULL pointer exception.
1696 Thanks to Anton Johansson for reporting this issues.
1699 *) Keep original DTLS digest and encryption contexts in retransmission
1700 structures so we can use the previous session parameters if they need
1701 to be resent. (CVE-2013-6450)
1704 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1705 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1706 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1707 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1708 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1709 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1710 [Rob Stradling, Adam Langley]
1712 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1714 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1715 supporting platforms or when small records were transferred.
1716 [Andy Polyakov, Steve Henson]
1718 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1720 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1722 This addresses the flaw in CBC record processing discovered by
1723 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1724 at: http://www.isg.rhul.ac.uk/tls/
1726 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1727 Security Group at Royal Holloway, University of London
1728 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1729 Emilia Käsper for the initial patch.
1731 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1733 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1734 ciphersuites which can be exploited in a denial of service attack.
1735 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1736 and detecting this bug and to Wolfgang Ettlinger
1737 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1741 *) Return an error when checking OCSP signatures when key is NULL.
1742 This fixes a DoS attack. (CVE-2013-0166)
1745 *) Make openssl verify return errors.
1746 [Chris Palmer <palmer@google.com> and Ben Laurie]
1748 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1749 the right response is stapled. Also change SSL_get_certificate()
1750 so it returns the certificate actually sent.
1751 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1752 [Rob Stradling <rob.stradling@comodo.com>]
1754 *) Fix possible deadlock when decoding public keys.
1757 *) Don't use TLS 1.0 record version number in initial client hello
1761 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1763 *) Sanity check record length before skipping explicit IV in TLS
1764 1.2, 1.1 and DTLS to fix DoS attack.
1766 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1767 fuzzing as a service testing platform.
1771 *) Initialise tkeylen properly when encrypting CMS messages.
1772 Thanks to Solar Designer of Openwall for reporting this issue.
1775 *) In FIPS mode don't try to use composite ciphers as they are not
1779 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1781 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1782 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1783 mean any application compiled against OpenSSL 1.0.0 headers setting
1784 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1785 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1786 0x10000000L Any application which was previously compiled against
1787 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1788 will need to be recompiled as a result. Letting be results in
1789 inability to disable specifically TLS 1.1 and in client context,
1790 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1793 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1794 disable just protocol X, but all protocols above X *if* there are
1795 protocols *below* X still enabled. In more practical terms it means
1796 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1797 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1798 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1802 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1804 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1805 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1806 in CRYPTO_realloc_clean.
1808 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1809 issue and to Adam Langley <agl@chromium.org> for fixing it.
1811 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1813 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1816 *) Workarounds for some broken servers that "hang" if a client hello
1817 record length exceeds 255 bytes.
1819 1. Do not use record version number > TLS 1.0 in initial client
1820 hello: some (but not all) hanging servers will now work.
1821 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1822 the number of ciphers sent in the client hello. This should be
1823 set to an even number, such as 50, for example by passing:
1824 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1825 Most broken servers should now work.
1826 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1827 TLS 1.2 client support entirely.
1830 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1833 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1835 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1836 STRING form instead of a DigestInfo.
1839 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1840 and the RSA_sign/RSA_verify functions. This was made more apparent when
1841 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1842 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1843 the correct format in RSA_verify so both forms transparently work.
1846 *) Some servers which support TLS 1.0 can choke if we initially indicate
1847 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1848 encrypted premaster secret. As a workaround use the maximum pemitted
1849 client version in client hello, this should keep such servers happy
1850 and still work with previous versions of OpenSSL.
1853 *) Add support for TLS/DTLS heartbeats.
1854 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1856 *) Add support for SCTP.
1857 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1859 *) Improved PRNG seeding for VOS.
1860 [Paul Green <Paul.Green@stratus.com>]
1862 *) Extensive assembler packs updates, most notably:
1864 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1865 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1866 - x86_64: bit-sliced AES implementation;
1867 - ARM: NEON support, contemporary platforms optimizations;
1868 - s390x: z196 support;
1869 - *: GHASH and GF(2^m) multiplication implementations;
1873 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1874 (removal of unnecessary code)
1875 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1877 *) Add TLS key material exporter from RFC 5705.
1880 *) Add DTLS-SRTP negotiation from RFC 5764.
1883 *) Add Next Protocol Negotiation,
1884 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1885 disabled with a no-npn flag to config or Configure. Code donated
1887 [Adam Langley <agl@google.com> and Ben Laurie]
1889 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1890 NIST-P256, NIST-P521, with constant-time single point multiplication on
1891 typical inputs. Compiler support for the nonstandard type __uint128_t is
1892 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1893 Code made available under Apache License version 2.0.
1895 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1896 line to include this in your build of OpenSSL, and run "make depend" (or
1897 "make update"). This enables the following EC_METHODs:
1899 EC_GFp_nistp224_method()
1900 EC_GFp_nistp256_method()
1901 EC_GFp_nistp521_method()
1903 EC_GROUP_new_by_curve_name() will automatically use these (while
1904 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1906 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1908 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1909 all platforms. Move ssize_t definition from e_os.h to the public
1910 header file e_os2.h as it now appears in public header file cms.h
1913 *) New -sigopt option to the ca, req and x509 utilities. Additional
1914 signature parameters can be passed using this option and in
1918 *) Add RSA PSS signing function. This will generate and set the
1919 appropriate AlgorithmIdentifiers for PSS based on those in the
1920 corresponding EVP_MD_CTX structure. No application support yet.
1923 *) Support for companion algorithm specific ASN1 signing routines.
1924 New function ASN1_item_sign_ctx() signs a pre-initialised
1925 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1926 the appropriate parameters.
1929 *) Add new algorithm specific ASN1 verification initialisation function
1930 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1931 handling will be the same no matter what EVP_PKEY_METHOD is used.
1932 Add a PSS handler to support verification of PSS signatures: checked
1933 against a number of sample certificates.
1936 *) Add signature printing for PSS. Add PSS OIDs.
1937 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1939 *) Add algorithm specific signature printing. An individual ASN1 method
1940 can now print out signatures instead of the standard hex dump.
1942 More complex signatures (e.g. PSS) can print out more meaningful
1943 information. Include DSA version that prints out the signature
1947 *) Password based recipient info support for CMS library: implementing
1951 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1952 neatly separates the code into cipher and PBE sections and is required
1953 for some algorithms that split PBES2 into separate pieces (such as
1954 password based CMS).
1957 *) Session-handling fixes:
1958 - Fix handling of connections that are resuming with a session ID,
1959 but also support Session Tickets.
1960 - Fix a bug that suppressed issuing of a new ticket if the client
1961 presented a ticket with an expired session.
1962 - Try to set the ticket lifetime hint to something reasonable.
1963 - Make tickets shorter by excluding irrelevant information.
1964 - On the client side, don't ignore renewed tickets.
1965 [Adam Langley, Bodo Moeller (Google)]
1967 *) Fix PSK session representation.
1970 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1972 This work was sponsored by Intel.
1975 *) Add GCM support to TLS library. Some custom code is needed to split
1976 the IV between the fixed (from PRF) and explicit (from TLS record)
1977 portions. This adds all GCM ciphersuites supported by RFC5288 and
1978 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
1979 add a special AESGCM string for GCM only.
1982 *) Expand range of ctrls for AES GCM. Permit setting invocation
1983 field on decrypt and retrieval of invocation field only on encrypt.
1986 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1987 As required by RFC5289 these ciphersuites cannot be used if for
1988 versions of TLS earlier than 1.2.
1991 *) For FIPS capable OpenSSL interpret a NULL default public key method
1992 as unset and return the appopriate default but do *not* set the default.
1993 This means we can return the appopriate method in applications that
1994 swicth between FIPS and non-FIPS modes.
1997 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1998 ENGINE is used then we cannot handle that in the FIPS module so we
1999 keep original code iff non-FIPS operations are allowed.
2002 *) Add -attime option to openssl utilities.
2003 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2005 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2008 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2009 FIPS EC methods unconditionally for now.
2012 *) New build option no-ec2m to disable characteristic 2 code.
2015 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2016 all cases can be covered as some introduce binary incompatibilities.
2019 *) Redirect RSA operations to FIPS module including keygen,
2020 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2023 *) Add similar low level API blocking to ciphers.
2026 *) Low level digest APIs are not approved in FIPS mode: any attempt
2027 to use these will cause a fatal error. Applications that *really* want
2028 to use them can use the private_* version instead.
2031 *) Redirect cipher operations to FIPS module for FIPS builds.
2034 *) Redirect digest operations to FIPS module for FIPS builds.
2037 *) Update build system to add "fips" flag which will link in fipscanister.o
2038 for static and shared library builds embedding a signature if needed.
2041 *) Output TLS supported curves in preference order instead of numerical
2042 order. This is currently hardcoded for the highest order curves first.
2043 This should be configurable so applications can judge speed vs strength.
2046 *) Add TLS v1.2 server support for client authentication.
2049 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2053 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2054 FIPS modules versions.
2057 *) Add TLS v1.2 client side support for client authentication. Keep cache
2058 of handshake records longer as we don't know the hash algorithm to use
2059 until after the certificate request message is received.
2062 *) Initial TLS v1.2 client support. Add a default signature algorithms
2063 extension including all the algorithms we support. Parse new signature
2064 format in client key exchange. Relax some ECC signing restrictions for
2065 TLS v1.2 as indicated in RFC5246.
2068 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2069 to new signature format when needed using client digest preference.
2070 All server ciphersuites should now work correctly in TLS v1.2. No client
2071 support yet and no support for client certificates.
2074 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2075 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2076 ciphersuites. At present only RSA key exchange ciphersuites work with
2077 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2078 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2079 and version checking.
2082 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2083 with this defined it will not be affected by any changes to ssl internal
2084 structures. Add several utility functions to allow openssl application
2085 to work with OPENSSL_NO_SSL_INTERN defined.
2088 *) A long standing patch to add support for SRP from EdelWeb (Peter
2089 Sylvester and Christophe Renou) was integrated.
2090 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
2091 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
2094 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2097 *) Permit abbreviated handshakes when renegotiating using the function
2098 SSL_renegotiate_abbreviated().
2099 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2101 *) Add call to ENGINE_register_all_complete() to
2102 ENGINE_load_builtin_engines(), so some implementations get used
2103 automatically instead of needing explicit application support.
2106 *) Add support for TLS key exporter as described in RFC5705.
2107 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2109 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2110 a few changes are required:
2112 Add SSL_OP_NO_TLSv1_1 flag.
2113 Add TLSv1_1 methods.
2114 Update version checking logic to handle version 1.1.
2115 Add explicit IV handling (ported from DTLS code).
2116 Add command line options to s_client/s_server.
2119 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2121 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2122 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2123 content decryption and always return the same error. Note: this attack
2124 needs on average 2^20 messages so it only affects automated senders. The
2125 old behaviour can be reenabled in the CMS code by setting the
2126 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2127 an MMA defence is not necessary.
2128 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2129 this issue. (CVE-2012-0884)
2132 *) Fix CVE-2011-4619: make sure we really are receiving a
2133 client hello before rejecting multiple SGC restarts. Thanks to
2134 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2137 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2139 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2140 Thanks to Antonio Martin, Enterprise Secure Access Research and
2141 Development, Cisco Systems, Inc. for discovering this bug and
2142 preparing a fix. (CVE-2012-0050)
2145 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2147 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2148 of the Vaudenay padding oracle attack on CBC mode encryption
2149 which enables an efficient plaintext recovery attack against
2150 the OpenSSL implementation of DTLS. Their attack exploits timing
2151 differences arising during decryption processing. A research
2152 paper describing this attack can be found at:
2153 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2154 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2155 Security Group at Royal Holloway, University of London
2156 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2157 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2158 for preparing the fix. (CVE-2011-4108)
2159 [Robin Seggelmann, Michael Tuexen]
2161 *) Clear bytes used for block padding of SSL 3.0 records.
2163 [Adam Langley (Google)]
2165 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2166 Kadianakis <desnacked@gmail.com> for discovering this issue and
2167 Adam Langley for preparing the fix. (CVE-2011-4619)
2168 [Adam Langley (Google)]
2170 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2171 [Andrey Kulikov <amdeich@gmail.com>]
2173 *) Prevent malformed RFC3779 data triggering an assertion failure.
2174 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2175 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2176 [Rob Austein <sra@hactrn.net>]
2178 *) Improved PRNG seeding for VOS.
2179 [Paul Green <Paul.Green@stratus.com>]
2181 *) Fix ssl_ciph.c set-up race.
2182 [Adam Langley (Google)]
2184 *) Fix spurious failures in ecdsatest.c.
2185 [Emilia Käsper (Google)]
2187 *) Fix the BIO_f_buffer() implementation (which was mixing different
2188 interpretations of the '..._len' fields).
2189 [Adam Langley (Google)]
2191 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2192 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2193 threads won't reuse the same blinding coefficients.
2195 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2196 lock to call BN_BLINDING_invert_ex, and avoids one use of
2197 BN_BLINDING_update for each BN_BLINDING structure (previously,
2198 the last update always remained unused).
2199 [Emilia Käsper (Google)]
2201 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2202 [Bob Buckholz (Google)]
2204 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2206 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2207 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2208 [Kaspar Brand <ossl@velox.ch>]
2210 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2211 for multi-threaded use of ECDH. (CVE-2011-3210)
2212 [Adam Langley (Google)]
2214 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2217 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2218 signature public key algorithm by using OID xref utilities instead.
2219 Before this you could only use some ECC ciphersuites with SHA1 only.
2222 *) Add protection against ECDSA timing attacks as mentioned in the paper
2223 by Billy Bob Brumley and Nicola Tuveri, see:
2225 http://eprint.iacr.org/2011/232.pdf
2227 [Billy Bob Brumley and Nicola Tuveri]
2229 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2231 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2232 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2234 *) Fix bug in string printing code: if *any* escaping is enabled we must
2235 escape the escape character (backslash) or the resulting string is
2239 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2241 *) Disable code workaround for ancient and obsolete Netscape browsers
2242 and servers: an attacker can use it in a ciphersuite downgrade attack.
2243 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2246 *) Fixed J-PAKE implementation error, originally discovered by
2247 Sebastien Martini, further info and confirmation from Stefan
2248 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2251 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2253 *) Fix extension code to avoid race conditions which can result in a buffer
2254 overrun vulnerability: resumed sessions must not be modified as they can
2255 be shared by multiple threads. CVE-2010-3864
2258 *) Fix WIN32 build system to correctly link an ENGINE directory into
2262 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2264 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2266 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2268 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2270 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2271 context. The operation can be customised via the ctrl mechanism in
2272 case ENGINEs want to include additional functionality.
2275 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2278 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2279 output hashes compatible with older versions of OpenSSL.
2280 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2282 *) Fix compression algorithm handling: if resuming a session use the
2283 compression algorithm of the resumed session instead of determining
2284 it from client hello again. Don't allow server to change algorithm.
2287 *) Add load_crls() function to apps tidying load_certs() too. Add option
2288 to verify utility to allow additional CRLs to be included.
2291 *) Update OCSP request code to permit adding custom headers to the request:
2292 some responders need this.
2295 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2297 [Julia Lawall <julia@diku.dk>]
2299 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2300 needlessly dereferenced structures, used obsolete functions and
2301 didn't handle all updated verify codes correctly.
2304 *) Disable MD2 in the default configuration.
2307 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2308 indicate the initial BIO being pushed or popped. This makes it possible
2309 to determine whether the BIO is the one explicitly called or as a result
2310 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2311 it handles reference counts correctly and doesn't zero out the I/O bio
2312 when it is not being explicitly popped. WARNING: applications which
2313 included workarounds for the old buggy behaviour will need to be modified
2314 or they could free up already freed BIOs.
2317 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2318 renaming to all platforms (within the 0.9.8 branch, this was
2319 done conditionally on Netware platforms to avoid a name clash).
2320 [Guenter <lists@gknw.net>]
2322 *) Add ECDHE and PSK support to DTLS.
2323 [Michael Tuexen <tuexen@fh-muenster.de>]
2325 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2329 *) Add "missing" function EVP_MD_flags() (without this the only way to
2330 retrieve a digest flags is by accessing the structure directly. Update
2331 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2332 or cipher is registered as in the "from" argument. Print out all
2333 registered digests in the dgst usage message instead of manually
2334 attempting to work them out.
2337 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2338 this allows the use of compression and extensions. Change default cipher
2339 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2340 by default unless an application cipher string requests it.
2343 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2344 key ids to find matching certificates and keys but some PKCS#12 files
2345 don't follow the (somewhat unwritten) rules and this strategy fails.
2346 Now just gather all certificates together and the first private key
2347 then look for the first certificate that matches the key.
2350 *) Support use of registered digest and cipher names for dgst and cipher
2351 commands instead of having to add each one as a special case. So now
2358 openssl dgst -sha256 foo
2360 and this works for ENGINE based algorithms too.
2364 *) Update Gost ENGINE to support parameter files.
2365 [Victor B. Wagner <vitus@cryptocom.ru>]
2367 *) Support GeneralizedTime in ca utility.
2368 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2370 *) Enhance the hash format used for certificate directory links. The new
2371 form uses the canonical encoding (meaning equivalent names will work
2372 even if they aren't identical) and uses SHA1 instead of MD5. This form
2373 is incompatible with the older format and as a result c_rehash should
2374 be used to rebuild symbolic links.
2377 *) Make PKCS#8 the default write format for private keys, replacing the
2378 traditional format. This form is standardised, more secure and doesn't
2379 include an implicit MD5 dependency.
2382 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2383 committed to OpenSSL should pass this lot as a minimum.
2386 *) Add session ticket override functionality for use by EAP-FAST.
2387 [Jouni Malinen <j@w1.fi>]
2389 *) Modify HMAC functions to return a value. Since these can be implemented
2390 in an ENGINE errors can occur.
2393 *) Type-checked OBJ_bsearch_ex.
2396 *) Type-checked OBJ_bsearch. Also some constification necessitated
2397 by type-checking. Still to come: TXT_DB, bsearch(?),
2398 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2402 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2403 seconds to a tm structure directly, instead of going through OS
2404 specific date routines. This avoids any issues with OS routines such
2405 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2406 and X509_time_adj_ex() to cover the extended range. The existing
2407 X509_time_adj() is still usable and will no longer have any date issues.
2410 *) Delta CRL support. New use deltas option which will attempt to locate
2411 and search any appropriate delta CRLs available.
2413 This work was sponsored by Google.
2416 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2417 code and add additional score elements. Validate alternate CRL paths
2418 as part of the CRL checking and indicate a new error "CRL path validation
2419 error" in this case. Applications wanting additional details can use
2420 the verify callback and check the new "parent" field. If this is not
2421 NULL CRL path validation is taking place. Existing applications wont
2422 see this because it requires extended CRL support which is off by
2425 This work was sponsored by Google.
2428 *) Support for freshest CRL extension.
2430 This work was sponsored by Google.
2433 *) Initial indirect CRL support. Currently only supported in the CRLs
2434 passed directly and not via lookup. Process certificate issuer
2435 CRL entry extension and lookup CRL entries by bother issuer name
2436 and serial number. Check and process CRL issuer entry in IDP extension.
2438 This work was sponsored by Google.
2441 *) Add support for distinct certificate and CRL paths. The CRL issuer
2442 certificate is validated separately in this case. Only enabled if
2443 an extended CRL support flag is set: this flag will enable additional
2444 CRL functionality in future.
2446 This work was sponsored by Google.
2449 *) Add support for policy mappings extension.
2451 This work was sponsored by Google.
2454 *) Fixes to pathlength constraint, self issued certificate handling,
2455 policy processing to align with RFC3280 and PKITS tests.
2457 This work was sponsored by Google.
2460 *) Support for name constraints certificate extension. DN, email, DNS
2461 and URI types are currently supported.
2463 This work was sponsored by Google.
2466 *) To cater for systems that provide a pointer-based thread ID rather
2467 than numeric, deprecate the current numeric thread ID mechanism and
2468 replace it with a structure and associated callback type. This
2469 mechanism allows a numeric "hash" to be extracted from a thread ID in
2470 either case, and on platforms where pointers are larger than 'long',
2471 mixing is done to help ensure the numeric 'hash' is usable even if it
2472 can't be guaranteed unique. The default mechanism is to use "&errno"
2473 as a pointer-based thread ID to distinguish between threads.
2475 Applications that want to provide their own thread IDs should now use
2476 CRYPTO_THREADID_set_callback() to register a callback that will call
2477 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2479 Note that ERR_remove_state() is now deprecated, because it is tied
2480 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2481 to free the current thread's error state should be replaced by
2482 ERR_remove_thread_state(NULL).
2484 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2485 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2486 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2487 application was previously providing a numeric thread callback that
2488 was inappropriate for distinguishing threads, then uniqueness might
2489 have been obtained with &errno that happened immediately in the
2490 intermediate development versions of OpenSSL; this is no longer the
2491 case, the numeric thread callback will now override the automatic use
2493 [Geoff Thorpe, with help from Bodo Moeller]
2495 *) Initial support for different CRL issuing certificates. This covers a
2496 simple case where the self issued certificates in the chain exist and
2497 the real CRL issuer is higher in the existing chain.
2499 This work was sponsored by Google.
2502 *) Removed effectively defunct crypto/store from the build.
2505 *) Revamp of STACK to provide stronger type-checking. Still to come:
2506 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2507 ASN1_STRING, CONF_VALUE.
2510 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2511 RAM on SSL connections. This option can save about 34k per idle SSL.
2514 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2515 STACK, TXT_DB, bsearch, qsort.
2518 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2519 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2520 support for data, signedData, compressedData, digestedData and
2521 encryptedData, envelopedData types included. Scripts to check against
2522 RFC4134 examples draft and interop and consistency checks of many
2523 content types and variants.
2526 *) Add options to enc utility to support use of zlib compression BIO.
2529 *) Extend mk1mf to support importing of options and assembly language
2530 files from Configure script, currently only included in VC-WIN32.
2531 The assembly language rules can now optionally generate the source
2532 files from the associated perl scripts.
2535 *) Implement remaining functionality needed to support GOST ciphersuites.
2536 Interop testing has been performed using CryptoPro implementations.
2537 [Victor B. Wagner <vitus@cryptocom.ru>]
2539 *) s390x assembler pack.
2542 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2546 *) Implement Opaque PRF Input TLS extension as specified in
2547 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2548 official specification yet and no extension type assignment by
2549 IANA exists, this extension (for now) will have to be explicitly
2550 enabled when building OpenSSL by providing the extension number
2551 to use. For example, specify an option
2553 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2555 to the "config" or "Configure" script to enable the extension,
2556 assuming extension number 0x9527 (which is a completely arbitrary
2557 and unofficial assignment based on the MD5 hash of the Internet
2558 Draft). Note that by doing so, you potentially lose
2559 interoperability with other TLS implementations since these might
2560 be using the same extension number for other purposes.
2562 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2563 opaque PRF input value to use in the handshake. This will create
2564 an interal copy of the length-'len' string at 'src', and will
2565 return non-zero for success.
2567 To get more control and flexibility, provide a callback function
2570 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2571 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2575 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2578 Callback function 'cb' will be called in handshakes, and is
2579 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2580 Argument 'arg' is for application purposes (the value as given to
2581 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2582 be provided to the callback function). The callback function
2583 has to return non-zero to report success: usually 1 to use opaque
2584 PRF input just if possible, or 2 to enforce use of the opaque PRF
2585 input. In the latter case, the library will abort the handshake
2586 if opaque PRF input is not successfully negotiated.
2588 Arguments 'peerinput' and 'len' given to the callback function
2589 will always be NULL and 0 in the case of a client. A server will
2590 see the client's opaque PRF input through these variables if
2591 available (NULL and 0 otherwise). Note that if the server
2592 provides an opaque PRF input, the length must be the same as the
2593 length of the client's opaque PRF input.
2595 Note that the callback function will only be called when creating
2596 a new session (session resumption can resume whatever was
2597 previously negotiated), and will not be called in SSL 2.0
2598 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2599 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2600 for applications that need to enforce opaque PRF input.
2604 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2607 [Victor B. Wagner <vitus@cryptocom.ru>]
2609 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2610 RFC4507bis. The encrypted ticket format is an encrypted encoded
2611 SSL_SESSION structure, that way new session features are automatically
2614 If a client application caches session in an SSL_SESSION structure
2615 support is transparent because tickets are now stored in the encoded
2618 The SSL_CTX structure automatically generates keys for ticket
2619 protection in servers so again support should be possible
2620 with no application modification.
2622 If a client or server wishes to disable RFC4507 support then the option
2623 SSL_OP_NO_TICKET can be set.
2625 Add a TLS extension debugging callback to allow the contents of any client
2626 or server extensions to be examined.
2628 This work was sponsored by Google.
2631 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2632 OpenSSL should now compile cleanly on gcc 4.2
2633 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2635 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2636 support including streaming MAC support: this is required for GOST
2637 ciphersuite support.
2638 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2640 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2641 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2642 to output in BER and PEM format.
2645 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2646 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2647 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2648 ENGINE support for HMAC keys which are unextractable. New -mac and
2649 -macopt options to dgst utility.
2652 *) New option -sigopt to dgst utility. Update dgst to use
2653 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2654 alternative signing paramaters such as X9.31 or PSS in the dgst
2658 *) Change ssl_cipher_apply_rule(), the internal function that does
2659 the work each time a ciphersuite string requests enabling
2660 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2661 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2662 the order of disabled ciphersuites such that those ciphersuites
2663 that most recently went from enabled to disabled not only stay
2664 in order with respect to each other, but also have higher priority
2665 than other disabled ciphersuites the next time ciphersuites are
2668 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2669 the same ciphersuites as with "HIGH" alone, but in a specific
2670 order where the PSK ciphersuites come first (since they are the
2671 most recently disabled ciphersuites when "HIGH" is parsed).
2673 Also, change ssl_create_cipher_list() (using this new
2674 funcionality) such that between otherwise identical
2675 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2679 *) Change ssl_create_cipher_list() so that it automatically
2680 arranges the ciphersuites in reasonable order before starting
2681 to process the rule string. Thus, the definition for "DEFAULT"
2682 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2683 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2684 This makes it much easier to arrive at a reasonable default order
2685 in applications for which anonymous ciphers are OK (meaning
2686 that you can't actually use DEFAULT).
2687 [Bodo Moeller; suggested by Victor Duchovni]
2689 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2690 processing) into multiple integers instead of setting
2691 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2692 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2693 (These masks as well as the individual bit definitions are hidden
2694 away into the non-exported interface ssl/ssl_locl.h, so this
2695 change to the definition of the SSL_CIPHER structure shouldn't
2696 affect applications.) This give us more bits for each of these
2697 categories, so there is no longer a need to coagulate AES128 and
2698 AES256 into a single algorithm bit, and to coagulate Camellia128
2699 and Camellia256 into a single algorithm bit, which has led to all
2702 Thus, among other things, the kludge introduced in 0.9.7m and
2703 0.9.8e for masking out AES256 independently of AES128 or masking
2704 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2706 With the change, we also introduce new ciphersuite aliases that
2707 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2711 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2712 Use the leftmost N bytes of the signature input if the input is
2713 larger than the prime q (with N being the size in bytes of q).
2716 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2717 it yet and it is largely untested.
2720 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2723 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2724 some compilers (gcc 4.2 and later) reject their use. Safestack is
2725 reimplemented. Update ASN1 to avoid use of legacy functions.
2728 *) Win32/64 targets are linked with Winsock2.
2731 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2732 to external functions. This can be used to increase CRL handling
2733 efficiency especially when CRLs are very large by (for example) storing
2734 the CRL revoked certificates in a database.
2737 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2738 new CRLs added to a directory can be used. New command line option
2739 -verify_return_error to s_client and s_server. This causes real errors
2740 to be returned by the verify callback instead of carrying on no matter
2741 what. This reflects the way a "real world" verify callback would behave.
2744 *) GOST engine, supporting several GOST algorithms and public key formats.
2745 Kindly donated by Cryptocom.
2748 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2749 partitioned by DP are handled but no indirect CRL or reason partitioning
2750 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2751 selected via a scoring technique which handles IDP and AKID in CRLs.
2754 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2755 will ultimately be used for all verify operations: this will remove the
2756 X509_STORE dependency on certificate verification and allow alternative
2757 lookup methods. X509_STORE based implementations of these two callbacks.
2760 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2761 Modify get_crl() to find a valid (unexpired) CRL if possible.
2764 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2765 this would be called X509_CRL_cmp() but that name is already used by
2766 a function that just compares CRL issuer names. Cache several CRL
2767 extensions in X509_CRL structure and cache CRLDP in X509.
2770 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2771 this maps equivalent X509_NAME structures into a consistent structure.
2772 Name comparison can then be performed rapidly using memcmp().
2775 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2779 *) Allow digests to supply their own micalg string for S/MIME type using
2780 the ctrl EVP_MD_CTRL_MICALG.
2783 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2784 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2785 ctrl. It can then customise the structure before and/or after signing
2789 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2790 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2791 to free up any added signature OIDs.
2794 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2795 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2796 digest and cipher tables. New options added to openssl utility:
2797 list-message-digest-algorithms and list-cipher-algorithms.
2800 *) Change the array representation of binary polynomials: the list
2801 of degrees of non-zero coefficients is now terminated with -1.
2802 Previously it was terminated with 0, which was also part of the
2803 value; thus, the array representation was not applicable to
2804 polynomials where t^0 has coefficient zero. This change makes
2805 the array representation useful in a more general context.
2808 *) Various modifications and fixes to SSL/TLS cipher string
2809 handling. For ECC, the code now distinguishes between fixed ECDH
2810 with RSA certificates on the one hand and with ECDSA certificates
2811 on the other hand, since these are separate ciphersuites. The
2812 unused code for Fortezza ciphersuites has been removed.
2814 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2815 (not "ECDHE"). For consistency with the code for DH
2816 certificates, use of ECDH certificates is now considered ECDH
2817 authentication, not RSA or ECDSA authentication (the latter is
2818 merely the CA's signing algorithm and not actively used in the
2821 The temporary ciphersuite alias "ECCdraft" is no longer
2822 available, and ECC ciphersuites are no longer excluded from "ALL"
2823 and "DEFAULT". The following aliases now exist for RFC 4492
2824 ciphersuites, most of these by analogy with the DH case:
2826 kECDHr - ECDH cert, signed with RSA
2827 kECDHe - ECDH cert, signed with ECDSA
2828 kECDH - ECDH cert (signed with either RSA or ECDSA)
2829 kEECDH - ephemeral ECDH
2830 ECDH - ECDH cert or ephemeral ECDH
2836 AECDH - anonymous ECDH
2837 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2841 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2842 Use correct micalg parameters depending on digest(s) in signed message.
2845 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2846 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2849 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2850 an engine to register a method. Add ENGINE lookups for methods and
2851 functional reference processing.
2854 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2855 EVP_{Sign,Verify}* which allow an application to customise the signature
2859 *) New -resign option to smime utility. This adds one or more signers
2860 to an existing PKCS#7 signedData structure. Also -md option to use an
2861 alternative message digest algorithm for signing.
2864 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2865 create PKCS7 structures containing multiple signers. Update smime
2866 application to support multiple signers.
2869 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2873 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2874 Reorganize PBE internals to lookup from a static table using NIDs,
2875 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2876 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2877 PRF which will be automatically used with PBES2.
2880 *) Replace the algorithm specific calls to generate keys in "req" with the
2884 *) Update PKCS#7 enveloped data routines to use new API. This is now
2885 supported by any public key method supporting the encrypt operation. A
2886 ctrl is added to allow the public key algorithm to examine or modify
2887 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2891 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2892 a default digest type to use. In most cases this will be SHA1 but some
2893 algorithms (such as GOST) need to specify an alternative digest. The
2894 return value indicates how strong the prefernce is 1 means optional and
2895 2 is mandatory (that is it is the only supported type). Modify
2896 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2897 use the default md. Update openssl utilities to use the default digest
2898 type for signing if it is not explicitly indicated.
2901 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2902 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2903 signing method from the key type. This effectively removes the link
2904 between digests and public key types.
2907 *) Add an OID cross reference table and utility functions. Its purpose is to
2908 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2909 rsaEncryption. This will allow some of the algorithm specific hackery
2910 needed to use the correct OID to be removed.
2913 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2914 structures for PKCS7_sign(). They are now set up by the relevant public
2918 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2921 *) Add support for key derivation (agreement) in the API, DH method and
2925 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2926 public and private key formats. As a side effect these add additional
2927 command line functionality not previously available: DSA signatures can be
2928 generated and verified using pkeyutl and DH key support and generation in
2933 [Oliver Tappe <zooey@hirschkaefer.de>]
2935 *) New make target "install_html_docs" installs HTML renditions of the
2937 [Oliver Tappe <zooey@hirschkaefer.de>]
2939 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
2940 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2941 support key and parameter generation and add initial key generation
2942 functionality for RSA.
2945 *) Add functions for main EVP_PKEY_method operations. The undocumented
2946 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2947 EVP_PKEY_{encrypt,decrypt}_old.
2950 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2951 key API, doesn't do much yet.
2954 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2955 public key algorithms. New option to openssl utility:
2956 "list-public-key-algorithms" to print out info.
2959 *) Implement the Supported Elliptic Curves Extension for
2960 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2963 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2964 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2967 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2968 utilities such as rsa, dsa, dsaparam etc except they process any key
2972 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2973 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2974 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2978 *) Initial support for pluggable public key ASN1.
2979 De-spaghettify the public key ASN1 handling. Move public and private
2980 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2981 algorithm specific handling to a single module within the relevant
2982 algorithm directory. Add functions to allow (near) opaque processing
2983 of public and private key structures.
2986 *) Implement the Supported Point Formats Extension for
2987 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2990 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2991 for the psk identity [hint] and the psk callback functions to the
2992 SSL_SESSION, SSL and SSL_CTX structure.
2995 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2999 SSL_CTX_use_psk_identity_hint
3000 SSL_get_psk_identity_hint
3001 SSL_get_psk_identity
3002 SSL_use_psk_identity_hint
3004 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3006 *) Add RFC 3161 compliant time stamp request creation, response generation
3007 and response verification functionality.
3008 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3010 *) Add initial support for TLS extensions, specifically for the server_name
3011 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3012 have new members for a host name. The SSL data structure has an
3013 additional member SSL_CTX *initial_ctx so that new sessions can be
3014 stored in that context to allow for session resumption, even after the
3015 SSL has been switched to a new SSL_CTX in reaction to a client's
3016 server_name extension.
3018 New functions (subject to change):
3020 SSL_get_servername()
3021 SSL_get_servername_type()
3024 New CTRL codes and macros (subject to change):
3026 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3027 - SSL_CTX_set_tlsext_servername_callback()
3028 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3029 - SSL_CTX_set_tlsext_servername_arg()
3030 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3032 openssl s_client has a new '-servername ...' option.
3034 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3035 '-key2 ...', '-servername_fatal' (subject to change). This allows
3036 testing the HostName extension for a specific single host name ('-cert'
3037 and '-key' remain fallbacks for handshakes without HostName
3038 negotiation). If the unrecogninzed_name alert has to be sent, this by
3039 default is a warning; it becomes fatal with the '-servername_fatal'
3042 [Peter Sylvester, Remy Allais, Christophe Renou]
3044 *) Whirlpool hash implementation is added.
3047 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3048 bn(64,32). Because of instruction set limitations it doesn't have
3049 any negative impact on performance. This was done mostly in order
3050 to make it possible to share assembler modules, such as bn_mul_mont
3051 implementations, between 32- and 64-bit builds without hassle.
3054 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3055 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3059 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3060 dedicated Montgomery multiplication procedure, is introduced.
3061 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3062 "64-bit" performance on certain 32-bit targets.
3065 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3066 in SSL structures. New SSL ctrl to set maximum send fragment size.
3067 Save memory by seeting the I/O buffer sizes dynamically instead of
3068 using the maximum available value.
3071 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3072 in addition to the text details.
3075 *) Very, very preliminary EXPERIMENTAL support for printing of general
3076 ASN1 structures. This currently produces rather ugly output and doesn't
3077 handle several customised structures at all.
3080 *) Integrated support for PVK file format and some related formats such
3081 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3082 these in the 'rsa' and 'dsa' utilities.
3085 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3088 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3089 place for the (very old) "NETSCAPE" format certificates which are now
3090 handled using new ASN1 code equivalents.
3093 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3094 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3095 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3098 *) Modify CRL distribution points extension code to print out previously
3099 unsupported fields. Enhance extension setting code to allow setting of
3103 *) Add print and set support for Issuing Distribution Point CRL extension.
3106 *) Change 'Configure' script to enable Camellia by default.
3109 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3111 *) When rejecting SSL/TLS records due to an incorrect version number, never
3112 update s->server with a new major version number. As of
3113 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3114 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3115 the previous behavior could result in a read attempt at NULL when
3116 receiving specific incorrect SSL/TLS records once record payload
3117 protection is active. (CVE-2010-0740)
3118 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3120 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3121 could be crashed if the relevant tables were not present (e.g. chrooted).
3122 [Tomas Hoger <thoger@redhat.com>]
3124 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3126 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3127 [Martin Olsson, Neel Mehta]
3129 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3130 accommodate for stack sorting, always a write lock!).
3133 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3134 excessive delays in the RAND_poll(): over a minute. As a workaround
3135 include a time check in the inner Heap32Next loop too.
3138 *) The code that handled flushing of data in SSL/TLS originally used the
3139 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3140 the problem outlined in PR#1949. The fix suggested there however can
3141 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3142 of Apache). So instead simplify the code to flush unconditionally.
3143 This should be fine since flushing with no data to flush is a no op.
3146 *) Handle TLS versions 2.0 and later properly and correctly use the
3147 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3148 off ancient servers have a habit of sticking around for a while...
3151 *) Modify compression code so it frees up structures without using the
3152 ex_data callbacks. This works around a problem where some applications
3153 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3154 restarting) then use compression (e.g. SSL with compression) later.
3155 This results in significant per-connection memory leaks and
3156 has caused some security issues including CVE-2008-1678 and
3160 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3161 change when encrypting or decrypting.
3164 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3165 connect and renegotiate with servers which do not support RI.
3166 Until RI is more widely deployed this option is enabled by default.
3169 *) Add "missing" ssl ctrls to clear options and mode.
3172 *) If client attempts to renegotiate and doesn't support RI respond with
3173 a no_renegotiation alert as required by RFC5746. Some renegotiating
3174 TLS clients will continue a connection gracefully when they receive
3175 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3176 waiting for a server hello which it will never receive. Now we treat a
3177 received no_renegotiation alert as a fatal error. This is because
3178 applications requesting a renegotiation might well expect it to succeed
3179 and would have no code in place to handle the server denying it so the
3180 only safe thing to do is to terminate the connection.
3183 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3184 peer supports secure renegotiation and 0 otherwise. Print out peer
3185 renegotiation support in s_client/s_server.
3188 *) Replace the highly broken and deprecated SPKAC certification method with
3189 the updated NID creation version. This should correctly handle UTF8.
3192 *) Implement RFC5746. Re-enable renegotiation but require the extension
3193 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3194 turns out to be a bad idea. It has been replaced by
3195 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3196 SSL_CTX_set_options(). This is really not recommended unless you
3197 know what you are doing.
3198 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3200 *) Fixes to stateless session resumption handling. Use initial_ctx when
3201 issuing and attempting to decrypt tickets in case it has changed during
3202 servername handling. Use a non-zero length session ID when attempting
3203 stateless session resumption: this makes it possible to determine if
3204 a resumption has occurred immediately after receiving server hello
3205 (several places in OpenSSL subtly assume this) instead of later in
3209 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3210 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3211 fixes for a few places where the return code is not checked
3213 [Julia Lawall <julia@diku.dk>]
3215 *) Add --strict-warnings option to Configure script to include devteam
3216 warnings in other configurations.
3219 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3220 makes it possible to install openssl libraries in locations which
3221 have names other than "lib", for example "/usr/lib64" which some
3223 [Steve Henson, based on patch from Jeremy Utley]
3225 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3226 X690 8.9.12 and can produce some misleading textual output of OIDs.
3227 [Steve Henson, reported by Dan Kaminsky]
3229 *) Delete MD2 from algorithm tables. This follows the recommendation in
3230 several standards that it is not used in new applications due to
3231 several cryptographic weaknesses. For binary compatibility reasons
3232 the MD2 API is still compiled in by default.
3235 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3239 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3240 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3242 [Guenter <lists@gknw.net>]
3244 *) Fix the server certificate chain building code to use X509_verify_cert(),
3245 it used to have an ad-hoc builder which was unable to cope with anything
3246 other than a simple chain.
3247 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3249 *) Don't check self signed certificate signatures in X509_verify_cert()
3250 by default (a flag can override this): it just wastes time without
3251 adding any security. As a useful side effect self signed root CAs
3252 with non-FIPS digests are now usable in FIPS mode.
3255 *) In dtls1_process_out_of_seq_message() the check if the current message
3256 is already buffered was missing. For every new message was memory
3257 allocated, allowing an attacker to perform an denial of service attack
3258 with sending out of seq handshake messages until there is no memory
3259 left. Additionally every future messege was buffered, even if the
3260 sequence number made no sense and would be part of another handshake.
3261 So only messages with sequence numbers less than 10 in advance will be
3262 buffered. (CVE-2009-1378)
3263 [Robin Seggelmann, discovered by Daniel Mentz]
3265 *) Records are buffered if they arrive with a future epoch to be
3266 processed after finishing the corresponding handshake. There is
3267 currently no limitation to this buffer allowing an attacker to perform
3268 a DOS attack with sending records with future epochs until there is no
3269 memory left. This patch adds the pqueue_size() function to detemine
3270 the size of a buffer and limits the record buffer to 100 entries.
3272 [Robin Seggelmann, discovered by Daniel Mentz]
3274 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3275 parent structure is freed. (CVE-2009-1379)
3278 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3279 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3282 [Ilya O. <vrghost@gmail.com>]
3284 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3286 *) Disable renegotiation completely - this fixes a severe security
3287 problem (CVE-2009-3555) at the cost of breaking all
3288 renegotiation. Renegotiation can be re-enabled by setting
3289 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3290 run-time. This is really not recommended unless you know what
3294 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3296 *) Don't set val to NULL when freeing up structures, it is freed up by
3297 underlying code. If sizeof(void *) > sizeof(long) this can result in
3298 zeroing past the valid field. (CVE-2009-0789)
3299 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3301 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3302 checked correctly. This would allow some invalid signed attributes to
3303 appear to verify correctly. (CVE-2009-0591)
3304 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3306 *) Reject UniversalString and BMPString types with invalid lengths. This
3307 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3308 a legal length. (CVE-2009-0590)
3311 *) Set S/MIME signing as the default purpose rather than setting it
3312 unconditionally. This allows applications to override it at the store
3316 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3317 to handle some structures.
3320 *) Improve efficiency of mem_gets: don't search whole buffer each time
3322 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3324 *) New -hex option for openssl rand.
3327 *) Print out UTF8String and NumericString when parsing ASN1.
3330 *) Support NumericString type for name components.
3333 *) Allow CC in the environment to override the automatically chosen
3334 compiler. Note that nothing is done to ensure flags work with the
3338 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3340 *) Properly check EVP_VerifyFinal() and similar return values
3342 [Ben Laurie, Bodo Moeller, Google Security Team]
3344 *) Enable TLS extensions by default.
3347 *) Allow the CHIL engine to be loaded, whether the application is
3348 multithreaded or not. (This does not release the developer from the
3349 obligation to set up the dynamic locking callbacks.)
3350 [Sander Temme <sander@temme.net>]
3352 *) Use correct exit code if there is an error in dgst command.
3353 [Steve Henson; problem pointed out by Roland Dirlewanger]
3355 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3356 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3359 *) Add experimental JPAKE support, including demo authentication in
3360 s_client and s_server.
3363 *) Set the comparison function in v3_addr_canonize().
3364 [Rob Austein <sra@hactrn.net>]
3366 *) Add support for XMPP STARTTLS in s_client.
3367 [Philip Paeps <philip@freebsd.org>]
3369 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3370 to ensure that even with this option, only ciphersuites in the
3371 server's preference list will be accepted. (Note that the option
3372 applies only when resuming a session, so the earlier behavior was
3373 just about the algorithm choice for symmetric cryptography.)
3376 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3378 *) Fix NULL pointer dereference if a DTLS server received
3379 ChangeCipherSpec as first record (CVE-2009-1386).
3382 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
3383 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3386 *) The fix in 0.9.8c that supposedly got rid of unsafe
3387 double-checked locking was incomplete for RSA blinding,
3388 addressing just one layer of what turns out to have been
3389 doubly unsafe triple-checked locking.
3391 So now fix this for real by retiring the MONT_HELPER macro
3392 in crypto/rsa/rsa_eay.c.
3394 [Bodo Moeller; problem pointed out by Marius Schilder]
3396 *) Various precautionary measures:
3398 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3400 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3401 (NB: This would require knowledge of the secret session ticket key
3402 to exploit, in which case you'd be SOL either way.)
3404 - Change bn_nist.c so that it will properly handle input BIGNUMs
3405 outside the expected range.
3407 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3410 [Neel Mehta, Bodo Moeller]
3412 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3413 the load fails. Useful for distros.
3414 [Ben Laurie and the FreeBSD team]
3416 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3419 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3422 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3424 This work was sponsored by Logica.
3427 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3428 keystores. Support for SSL/TLS client authentication too.
3429 Not compiled unless enable-capieng specified to Configure.
3431 This work was sponsored by Logica.
3434 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3435 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3436 attribute creation routines such as certifcate requests and PKCS#12
3440 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3442 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3443 handshake which could lead to a cilent crash as found using the
3444 Codenomicon TLS test suite (CVE-2008-1672)
3445 [Steve Henson, Mark Cox]
3447 *) Fix double free in TLS server name extensions which could lead to
3448 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3451 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3453 Clear the error queue to ensure that error entries left from
3454 older function calls do not interfere with the correct operation.
3455 [Lutz Jaenicke, Erik de Castro Lopo]
3457 *) Remove root CA certificates of commercial CAs:
3459 The OpenSSL project does not recommend any specific CA and does not
3460 have any policy with respect to including or excluding any CA.
3461 Therefore it does not make any sense to ship an arbitrary selection
3462 of root CA certificates with the OpenSSL software.
3465 *) RSA OAEP patches to fix two separate invalid memory reads.
3466 The first one involves inputs when 'lzero' is greater than
3467 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3468 before the beginning of from). The second one involves inputs where
3469 the 'db' section contains nothing but zeroes (there is a one-byte
3470 invalid read after the end of 'db').
3471 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3473 *) Partial backport from 0.9.9-dev:
3475 Introduce bn_mul_mont (dedicated Montgomery multiplication
3476 procedure) as a candidate for BIGNUM assembler implementation.
3477 While 0.9.9-dev uses assembler for various architectures, only
3478 x86_64 is available by default here in the 0.9.8 branch, and
3479 32-bit x86 is available through a compile-time setting.
3481 To try the 32-bit x86 assembler implementation, use Configure
3482 option "enable-montasm" (which exists only for this backport).
3484 As "enable-montasm" for 32-bit x86 disclaims code stability
3485 anyway, in this constellation we activate additional code
3486 backported from 0.9.9-dev for further performance improvements,
3487 namely BN_from_montgomery_word. (To enable this otherwise,
3488 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3490 [Andy Polyakov (backport partially by Bodo Moeller)]
3492 *) Add TLS session ticket callback. This allows an application to set
3493 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3494 values. This is useful for key rollover for example where several key
3495 sets may exist with different names.
3498 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3499 This was broken until now in 0.9.8 releases, such that the only way
3500 a registered ENGINE could be used (assuming it initialises
3501 successfully on the host) was to explicitly set it as the default
3502 for the relevant algorithms. This is in contradiction with 0.9.7
3503 behaviour and the documentation. With this fix, when an ENGINE is
3504 registered into a given algorithm's table of implementations, the
3505 'uptodate' flag is reset so that auto-discovery will be used next
3506 time a new context for that algorithm attempts to select an
3508 [Ian Lister (tweaked by Geoff Thorpe)]
3510 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3511 implemention in the following ways:
3513 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3516 Lack of BER streaming support means one pass streaming processing is
3517 only supported if data is detached: setting the streaming flag is
3518 ignored for embedded content.
3520 CMS support is disabled by default and must be explicitly enabled
3521 with the enable-cms configuration option.
3524 *) Update the GMP engine glue to do direct copies between BIGNUM and
3525 mpz_t when openssl and GMP use the same limb size. Otherwise the
3526 existing "conversion via a text string export" trick is still used.
3527 [Paul Sheer <paulsheer@gmail.com>]
3529 *) Zlib compression BIO. This is a filter BIO which compressed and
3530 uncompresses any data passed through it.
3533 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3534 RFC3394 compatible AES key wrapping.
3537 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3538 sets string data without copying. X509_ALGOR_set0() and
3539 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3540 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3541 from an X509_ATTRIBUTE structure optionally checking it occurs only
3542 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3546 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3547 to get the expected BN_FLG_CONSTTIME behavior.
3548 [Bodo Moeller (Google)]
3552 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3553 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3554 - added some more tests to do_tests.pl
3555 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3556 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3557 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3558 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3559 - various changes to netware.pl to enable gcc-cross builds on Win32
3561 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3562 - various changes to fix missing prototype warnings
3563 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3564 - added AES, WHIRLPOOL and CPUID assembler code to build files
3565 - added missing AES assembler make rules to mk1mf.pl
3566 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3567 [Guenter Knauf <eflash@gmx.net>]
3569 *) Implement certificate status request TLS extension defined in RFC3546.
3570 A client can set the appropriate parameters and receive the encoded
3571 OCSP response via a callback. A server can query the supplied parameters
3572 and set the encoded OCSP response in the callback. Add simplified examples
3573 to s_client and s_server.
3576 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3578 *) Fix various bugs:
3579 + Binary incompatibility of ssl_ctx_st structure
3580 + DTLS interoperation with non-compliant servers
3581 + Don't call get_session_cb() without proposed session
3582 + Fix ia64 assembler code
3583 [Andy Polyakov, Steve Henson]
3585 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3587 *) DTLS Handshake overhaul. There were longstanding issues with
3588 OpenSSL DTLS implementation, which were making it impossible for
3589 RFC 4347 compliant client to communicate with OpenSSL server.
3590 Unfortunately just fixing these incompatibilities would "cut off"
3591 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3592 server keeps tolerating non RFC compliant syntax. The opposite is
3593 not true, 0.9.8f client can not communicate with earlier server.
3594 This update even addresses CVE-2007-4995.
3597 *) Changes to avoid need for function casts in OpenSSL: some compilers
3598 (gcc 4.2 and later) reject their use.
3599 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3602 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3603 RFC4507bis. The encrypted ticket format is an encrypted encoded
3604 SSL_SESSION structure, that way new session features are automatically
3607 If a client application caches session in an SSL_SESSION structure
3608 support is transparent because tickets are now stored in the encoded
3611 The SSL_CTX structure automatically generates keys for ticket
3612 protection in servers so again support should be possible
3613 with no application modification.
3615 If a client or server wishes to disable RFC4507 support then the option
3616 SSL_OP_NO_TICKET can be set.
3618 Add a TLS extension debugging callback to allow the contents of any client
3619 or server extensions to be examined.
3621 This work was sponsored by Google.
3624 *) Add initial support for TLS extensions, specifically for the server_name
3625 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3626 have new members for a host name. The SSL data structure has an
3627 additional member SSL_CTX *initial_ctx so that new sessions can be
3628 stored in that context to allow for session resumption, even after the
3629 SSL has been switched to a new SSL_CTX in reaction to a client's
3630 server_name extension.
3632 New functions (subject to change):
3634 SSL_get_servername()
3635 SSL_get_servername_type()
3638 New CTRL codes and macros (subject to change):
3640 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3641 - SSL_CTX_set_tlsext_servername_callback()
3642 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3643 - SSL_CTX_set_tlsext_servername_arg()
3644 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3646 openssl s_client has a new '-servername ...' option.
3648 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3649 '-key2 ...', '-servername_fatal' (subject to change). This allows
3650 testing the HostName extension for a specific single host name ('-cert'
3651 and '-key' remain fallbacks for handshakes without HostName
3652 negotiation). If the unrecogninzed_name alert has to be sent, this by
3653 default is a warning; it becomes fatal with the '-servername_fatal'
3656 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3658 *) Add AES and SSE2 assembly language support to VC++ build.
3661 *) Mitigate attack on final subtraction in Montgomery reduction.
3664 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3665 (which previously caused an internal error).
3668 *) Squeeze another 10% out of IGE mode when in != out.
3671 *) AES IGE mode speedup.
3672 [Dean Gaudet (Google)]
3674 *) Add the Korean symmetric 128-bit cipher SEED (see
3675 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3676 add SEED ciphersuites from RFC 4162:
3678 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3679 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3680 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3681 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3683 To minimize changes between patchlevels in the OpenSSL 0.9.8
3684 series, SEED remains excluded from compilation unless OpenSSL
3685 is configured with 'enable-seed'.
3686 [KISA, Bodo Moeller]
3688 *) Mitigate branch prediction attacks, which can be practical if a
3689 single processor is shared, allowing a spy process to extract
3690 information. For detailed background information, see
3691 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3692 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3693 and Necessary Software Countermeasures"). The core of the change
3694 are new versions BN_div_no_branch() and
3695 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3696 respectively, which are slower, but avoid the security-relevant
3697 conditional branches. These are automatically called by BN_div()
3698 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3699 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3700 remove a conditional branch.
3702 BN_FLG_CONSTTIME is the new name for the previous
3703 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3704 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3705 in the exponent causes BN_mod_exp_mont() to use the alternative
3706 implementation in BN_mod_exp_mont_consttime().) The old name
3707 remains as a deprecated alias.
3709 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3710 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3711 constant-time implementations for more than just exponentiation.
3712 Here too the old name is kept as a deprecated alias.
3714 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3715 the BN_BLINDING structure gets an independent copy of the
3716 modulus. This means that the previous "BIGNUM *m" argument to
3717 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3718 essentially becomes "const BIGNUM *m", although we can't actually
3719 change this in the header file before 0.9.9. It allows
3720 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3721 enable BN_FLG_CONSTTIME.
3723 [Matthew D Wood (Intel Corp)]
3725 *) In the SSL/TLS server implementation, be strict about session ID
3726 context matching (which matters if an application uses a single
3727 external cache for different purposes). Previously,
3728 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3729 set. This did ensure strict client verification, but meant that,
3730 with applications using a single external cache for quite
3731 different requirements, clients could circumvent ciphersuite
3732 restrictions for a given session ID context by starting a session
3733 in a different context.
3736 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3737 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3738 authentication-only ciphersuites.
3741 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3742 not complete and could lead to a possible single byte overflow
3743 (CVE-2007-5135) [Ben Laurie]
3745 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3747 *) Since AES128 and AES256 (and similarly Camellia128 and
3748 Camellia256) share a single mask bit in the logic of
3749 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3750 kludge to work properly if AES128 is available and AES256 isn't
3751 (or if Camellia128 is available and Camellia256 isn't).
3754 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3755 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3756 When a point or a seed is encoded in a BIT STRING, we need to
3757 prevent the removal of trailing zero bits to get the proper DER
3758 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3759 of a NamedBitList, for which trailing 0 bits need to be removed.)
3762 *) Have SSL/TLS server implementation tolerate "mismatched" record
3763 protocol version while receiving ClientHello even if the
3764 ClientHello is fragmented. (The server can't insist on the
3765 particular protocol version it has chosen before the ServerHello
3766 message has informed the client about his choice.)
3769 *) Add RFC 3779 support.
3770 [Rob Austein for ARIN, Ben Laurie]
3772 *) Load error codes if they are not already present instead of using a
3773 static variable. This allows them to be cleanly unloaded and reloaded.
3774 Improve header file function name parsing.
3777 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3778 or CAPABILITY handshake as required by RFCs.
3781 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3783 *) Introduce limits to prevent malicious keys being able to
3784 cause a denial of service. (CVE-2006-2940)
3785 [Steve Henson, Bodo Moeller]
3787 *) Fix ASN.1 parsing of certain invalid structures that can result
3788 in a denial of service. (CVE-2006-2937) [Steve Henson]
3790 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3791 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3793 *) Fix SSL client code which could crash if connecting to a
3794 malicious SSLv2 server. (CVE-2006-4343)
3795 [Tavis Ormandy and Will Drewry, Google Security Team]
3797 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3798 match only those. Before that, "AES256-SHA" would be interpreted
3799 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3800 the same strength classification in 0.9.7h) as we currently only
3801 have a single AES bit in the ciphersuite description bitmap.
3802 That change, however, also applied to ciphersuite strings such as
3803 "RC4-MD5" that intentionally matched multiple ciphersuites --
3804 namely, SSL 2.0 ciphersuites in addition to the more common ones
3805 from SSL 3.0/TLS 1.0.
3807 So we change the selection algorithm again: Naming an explicit
3808 ciphersuite selects this one ciphersuite, and any other similar
3809 ciphersuite (same bitmap) from *other* protocol versions.
3810 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3811 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3813 Since SSL 2.0 does not have any ciphersuites for which the
3814 128/256 bit distinction would be relevant, this works for now.
3815 The proper fix will be to use different bits for AES128 and
3816 AES256, which would have avoided the problems from the beginning;
3817 however, bits are scarce, so we can only do this in a new release
3818 (not just a patchlevel) when we can change the SSL_CIPHER
3819 definition to split the single 'unsigned long mask' bitmap into
3820 multiple values to extend the available space.
3824 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3826 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3827 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3829 *) Add AES IGE and biIGE modes.
3832 *) Change the Unix randomness entropy gathering to use poll() when
3833 possible instead of select(), since the latter has some
3834 undesirable limitations.
3835 [Darryl Miles via Richard Levitte and Bodo Moeller]
3837 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3838 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3839 cannot be implicitly activated as part of, e.g., the "AES" alias.
3840 However, please upgrade to OpenSSL 0.9.9[-dev] for
3841 non-experimental use of the ECC ciphersuites to get TLS extension
3842 support, which is required for curve and point format negotiation
3843 to avoid potential handshake problems.
3846 *) Disable rogue ciphersuites:
3848 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3849 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3850 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3852 The latter two were purportedly from
3853 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3856 Also deactivate the remaining ciphersuites from
3857 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3858 unofficial, and the ID has long expired.
3861 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3862 dual-core machines) and other potential thread-safety issues.
3865 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3866 versions), which is now available for royalty-free use
3867 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3868 Also, add Camellia TLS ciphersuites from RFC 4132.
3870 To minimize changes between patchlevels in the OpenSSL 0.9.8
3871 series, Camellia remains excluded from compilation unless OpenSSL
3872 is configured with 'enable-camellia'.
3875 *) Disable the padding bug check when compression is in use. The padding
3876 bug check assumes the first packet is of even length, this is not
3877 necessarily true if compresssion is enabled and can result in false
3878 positives causing handshake failure. The actual bug test is ancient
3879 code so it is hoped that implementations will either have fixed it by
3880 now or any which still have the bug do not support compression.
3883 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3885 *) When applying a cipher rule check to see if string match is an explicit
3886 cipher suite and only match that one cipher suite if it is.
3889 *) Link in manifests for VC++ if needed.
3890 [Austin Ziegler <halostatue@gmail.com>]
3892 *) Update support for ECC-based TLS ciphersuites according to
3893 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3894 TLS extensions, which are supported starting with the 0.9.9
3895 branch, not in the OpenSSL 0.9.8 branch).
3898 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3899 opaque EVP_CIPHER_CTX handling.
3902 *) Fixes and enhancements to zlib compression code. We now only use
3903 "zlib1.dll" and use the default __cdecl calling convention on Win32
3904 to conform with the standards mentioned here:
3905 http://www.zlib.net/DLL_FAQ.txt
3906 Static zlib linking now works on Windows and the new --with-zlib-include
3907 --with-zlib-lib options to Configure can be used to supply the location
3908 of the headers and library. Gracefully handle case where zlib library
3912 *) Several fixes and enhancements to the OID generation code. The old code
3913 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3914 handle numbers larger than ULONG_MAX, truncated printing and had a
3915 non standard OBJ_obj2txt() behaviour.
3918 *) Add support for building of engines under engine/ as shared libraries
3919 under VC++ build system.
3922 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3923 Hopefully, we will not see any false combination of paths any more.
3926 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3928 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3929 (part of SSL_OP_ALL). This option used to disable the
3930 countermeasure against man-in-the-middle protocol-version
3931 rollback in the SSL 2.0 server implementation, which is a bad
3932 idea. (CVE-2005-2969)
3934 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3935 for Information Security, National Institute of Advanced Industrial
3936 Science and Technology [AIST], Japan)]
3938 *) Add two function to clear and return the verify parameter flags.
3941 *) Keep cipherlists sorted in the source instead of sorting them at
3942 runtime, thus removing the need for a lock.
3945 *) Avoid some small subgroup attacks in Diffie-Hellman.
3946 [Nick Mathewson and Ben Laurie]
3948 *) Add functions for well-known primes.
3951 *) Extended Windows CE support.
3952 [Satoshi Nakamura and Andy Polyakov]
3954 *) Initialize SSL_METHOD structures at compile time instead of during
3955 runtime, thus removing the need for a lock.
3958 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3959 attempting to decrypt each encrypted key in turn. Add support to
3963 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3965 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3968 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3971 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3972 key into the same file any more.
3975 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3978 *) Add -utf8 command line and config file option to 'ca'.
3979 [Stefan <stf@udoma.org]
3981 *) Removed the macro des_crypt(), as it seems to conflict with some
3982 libraries. Use DES_crypt().
3985 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3986 involves renaming the source and generated shared-libs for
3987 both. The engines will accept the corrected or legacy ids
3988 ('ncipher' and '4758_cca' respectively) when binding. NB,
3989 this only applies when building 'shared'.
3990 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3992 *) Add attribute functions to EVP_PKEY structure. Modify
3993 PKCS12_create() to recognize a CSP name attribute and
3994 use it. Make -CSP option work again in pkcs12 utility.
3997 *) Add new functionality to the bn blinding code:
3998 - automatic re-creation of the BN_BLINDING parameters after
3999 a fixed number of uses (currently 32)
4000 - add new function for parameter creation
4001 - introduce flags to control the update behaviour of the
4002 BN_BLINDING parameters
4003 - hide BN_BLINDING structure
4004 Add a second BN_BLINDING slot to the RSA structure to improve
4005 performance when a single RSA object is shared among several
4009 *) Add support for DTLS.
4010 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
4012 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
4013 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
4016 *) Remove buggy and incompletet DH cert support from
4017 ssl/ssl_rsa.c and ssl/s3_both.c
4020 *) Use SHA-1 instead of MD5 as the default digest algorithm for
4021 the apps/openssl applications.
4024 *) Compile clean with "-Wall -Wmissing-prototypes
4025 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
4026 DEBUG_SAFESTACK must also be set.
4029 *) Change ./Configure so that certain algorithms can be disabled by default.
4030 The new counterpiece to "no-xxx" is "enable-xxx".
4032 The patented RC5 and MDC2 algorithms will now be disabled unless
4033 "enable-rc5" and "enable-mdc2", respectively, are specified.
4035 (IDEA remains enabled despite being patented. This is because IDEA
4036 is frequently required for interoperability, and there is no license
4037 fee for non-commercial use. As before, "no-idea" can be used to
4038 avoid this algorithm.)
4042 *) Add processing of proxy certificates (see RFC 3820). This work was
4043 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
4044 EGEE (Enabling Grids for E-science in Europe).
4047 *) RC4 performance overhaul on modern architectures/implementations, such
4048 as Intel P4, IA-64 and AMD64.
4051 *) New utility extract-section.pl. This can be used specify an alternative
4052 section number in a pod file instead of having to treat each file as
4053 a separate case in Makefile. This can be done by adding two lines to the
4056 =for comment openssl_section:XXX
4058 The blank line is mandatory.
4062 *) New arguments -certform, -keyform and -pass for s_client and s_server
4063 to allow alternative format key and certificate files and passphrase
4067 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4068 update associated structures and add various utility functions.
4070 Add new policy related verify parameters, include policy checking in
4071 standard verify code. Enhance 'smime' application with extra parameters
4072 to support policy checking and print out.
4075 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4076 Nehemiah processors. These extensions support AES encryption in hardware
4077 as well as RNG (though RNG support is currently disabled).
4078 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4080 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4083 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4084 [Andy Polyakov and a number of other people]
4086 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4087 implementation contributed by IBM.
4088 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4090 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4091 exponent rather than 'unsigned long'. There is a corresponding change to
4092 the new 'rsa_keygen' element of the RSA_METHOD structure.
4093 [Jelte Jansen, Geoff Thorpe]
4095 *) Functionality for creating the initial serial number file is now
4096 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4098 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4099 number file to 1, which is bound to cause problems. To avoid
4100 the problems while respecting compatibility between different 0.9.7
4101 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4102 CA.pl for serial number initialization. With the new release 0.9.8,
4103 we can fix the problem directly in the 'ca' utility.)
4106 *) Reduced header interdepencies by declaring more opaque objects in
4107 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4108 give fewer recursive includes, which could break lazy source code - so
4109 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4110 developers should define this symbol when building and using openssl to
4111 ensure they track the recommended behaviour, interfaces, [etc], but
4112 backwards-compatible behaviour prevails when this isn't defined.
4115 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4118 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4119 This will generate a random key of the appropriate length based on the
4120 cipher context. The EVP_CIPHER can provide its own random key generation
4121 routine to support keys of a specific form. This is used in the des and
4122 3des routines to generate a key of the correct parity. Update S/MIME
4123 code to use new functions and hence generate correct parity DES keys.
4124 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4125 valid (weak or incorrect parity).
4128 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4129 as looking them up. This is useful when the verified structure may contain
4130 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4131 present unless the new PKCS7_NO_CRL flag is asserted.
4134 *) Extend ASN1 oid configuration module. It now additionally accepts the
4137 shortName = some long name, 1.2.3.4
4140 *) Reimplemented the BN_CTX implementation. There is now no more static
4141 limitation on the number of variables it can handle nor the depth of the
4142 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4143 information can now expand as required, and rather than having a single
4144 static array of bignums, BN_CTX now uses a linked-list of such arrays
4145 allowing it to expand on demand whilst maintaining the usefulness of
4146 BN_CTX's "bundling".
4149 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4150 to allow all RSA operations to function using a single BN_CTX.
4153 *) Preliminary support for certificate policy evaluation and checking. This
4154 is initially intended to pass the tests outlined in "Conformance Testing
4155 of Relying Party Client Certificate Path Processing Logic" v1.07.
4158 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4159 remained unused and not that useful. A variety of other little bignum
4160 tweaks and fixes have also been made continuing on from the audit (see
4164 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4165 associated ASN1, EVP and SSL functions and old ASN1 macros.
4168 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4169 and this should never fail. So the return value from the use of
4170 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4171 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4174 *) BN_CTX_get() should return zero-valued bignums, providing the same
4175 initialised value as BN_new().
4176 [Geoff Thorpe, suggested by Ulf Möller]
4178 *) Support for inhibitAnyPolicy certificate extension.
4181 *) An audit of the BIGNUM code is underway, for which debugging code is
4182 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4183 is considered valid when processing BIGNUMs, and causes execution to
4184 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4185 further steps are taken to deliberately pollute unused data in BIGNUM
4186 structures to try and expose faulty code further on. For now, openssl will
4187 (in its default mode of operation) continue to tolerate the inconsistent
4188 forms that it has tolerated in the past, but authors and packagers should
4189 consider trying openssl and their own applications when compiled with
4190 these debugging symbols defined. It will help highlight potential bugs in
4191 their own code, and will improve the test coverage for OpenSSL itself. At
4192 some point, these tighter rules will become openssl's default to improve
4193 maintainability, though the assert()s and other overheads will remain only
4194 in debugging configurations. See bn.h for more details.
4195 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4197 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4198 that can only be obtained through BN_CTX_new() (which implicitly
4199 initialises it). The presence of this function only made it possible
4200 to overwrite an existing structure (and cause memory leaks).
4203 *) Because of the callback-based approach for implementing LHASH as a
4204 template type, lh_insert() adds opaque objects to hash-tables and
4205 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4206 to clean up those corresponding objects before destroying the hash table
4207 (and losing the object pointers). So some over-zealous constifications in
4208 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4209 objects as "const" and the lh_doall[_arg] callback wrappers are not
4210 prototyped to have "const" restrictions on the object pointers they are
4211 given (and so aren't required to cast them away any more).
4214 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4215 (speed) prefers to use its own implementation. The two implementations
4216 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4217 its object type properly exposed (MS_TM) instead of casting to/from "char
4218 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4219 aren't necessarily the greatest nomenclatures - but this is what was used
4220 internally to the implementation so I've used that for now.
4223 *) Ensure that deprecated functions do not get compiled when
4224 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4225 the self-tests were still using deprecated key-generation functions so
4226 these have been updated also.
4229 *) Reorganise PKCS#7 code to separate the digest location functionality
4230 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
4231 New function PKCS7_set_digest() to set the digest type for PKCS#7
4232 digestedData type. Add additional code to correctly generate the
4233 digestedData type and add support for this type in PKCS7 initialization
4237 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4238 structure of type "other".
4241 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4242 sure the loop does correctly stop and breaking ("division by zero")
4243 modulus operations are not performed. The (pre-generated) prime
4244 table crypto/bn/bn_prime.h was already correct, but it could not be
4245 re-generated on some platforms because of the "division by zero"
4246 situation in the script.
4247 [Ralf S. Engelschall]
4249 *) Update support for ECC-based TLS ciphersuites according to
4250 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4251 SHA-1 now is only used for "small" curves (where the
4252 representation of a field element takes up to 24 bytes); for
4253 larger curves, the field element resulting from ECDH is directly
4254 used as premaster secret.
4255 [Douglas Stebila (Sun Microsystems Laboratories)]
4257 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4258 curve secp160r1 to the tests.
4259 [Douglas Stebila (Sun Microsystems Laboratories)]
4261 *) Add the possibility to load symbols globally with DSO.
4262 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4264 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4265 control of the error stack.
4268 *) Add support for STORE in ENGINE.
4271 *) Add the STORE type. The intention is to provide a common interface
4272 to certificate and key stores, be they simple file-based stores, or
4273 HSM-type store, or LDAP stores, or...
4274 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4277 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4278 pass a list of arguments to any function as well as provide a way
4279 for a function to pass data back to the caller.
4282 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4283 works like BUF_strdup() but can be used to duplicate a portion of
4284 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4288 *) Add the function sk_find_ex() which works like sk_find(), but will
4289 return an index to an element even if an exact match couldn't be
4290 found. The index is guaranteed to point at the element where the
4291 searched-for key would be inserted to preserve sorting order.
4294 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4295 takes an extra flags argument for optional functionality. Currently,
4296 the following flags are defined:
4298 OBJ_BSEARCH_VALUE_ON_NOMATCH
4299 This one gets OBJ_bsearch_ex() to return a pointer to the first
4300 element where the comparing function returns a negative or zero
4303 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4304 This one gets OBJ_bsearch_ex() to return a pointer to the first
4305 element where the comparing function returns zero. This is useful
4306 if there are more than one element where the comparing function
4310 *) Make it possible to create self-signed certificates with 'openssl ca'
4311 in such a way that the self-signed certificate becomes part of the
4312 CA database and uses the same mechanisms for serial number generation
4313 as all other certificate signing. The new flag '-selfsign' enables
4314 this functionality. Adapt CA.sh and CA.pl.in.
4317 *) Add functionality to check the public key of a certificate request
4318 against a given private. This is useful to check that a certificate
4319 request can be signed by that key (self-signing).
4322 *) Make it possible to have multiple active certificates with the same
4323 subject in the CA index file. This is done only if the keyword
4324 'unique_subject' is set to 'no' in the main CA section (default
4325 if 'CA_default') of the configuration file. The value is saved
4326 with the database itself in a separate index attribute file,
4327 named like the index file with '.attr' appended to the name.
4330 *) Generate muti valued AVAs using '+' notation in config files for
4334 *) Support for nameConstraints certificate extension.
4337 *) Support for policyConstraints certificate extension.
4340 *) Support for policyMappings certificate extension.
4343 *) Make sure the default DSA_METHOD implementation only uses its
4344 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4345 and change its own handlers to be NULL so as to remove unnecessary
4346 indirection. This lets alternative implementations fallback to the
4347 default implementation more easily.
4350 *) Support for directoryName in GeneralName related extensions
4354 *) Make it possible to link applications using Makefile.shared.
4355 Make that possible even when linking against static libraries!
4358 *) Support for single pass processing for S/MIME signing. This now
4359 means that S/MIME signing can be done from a pipe, in addition
4360 cleartext signing (multipart/signed type) is effectively streaming
4361 and the signed data does not need to be all held in memory.
4363 This is done with a new flag PKCS7_STREAM. When this flag is set
4364 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4365 is done after the data is output (and digests calculated) in
4366 SMIME_write_PKCS7().
4369 *) Add full support for -rpath/-R, both in shared libraries and
4370 applications, at least on the platforms where it's known how
4374 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4375 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4376 will now compute a table of multiples of the generator that
4377 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4378 faster (notably in the case of a single point multiplication,
4379 scalar * generator).
4380 [Nils Larsch, Bodo Moeller]
4382 *) IPv6 support for certificate extensions. The various extensions
4383 which use the IP:a.b.c.d can now take IPv6 addresses using the
4384 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4388 *) Added an ENGINE that implements RSA by performing private key
4389 exponentiations with the GMP library. The conversions to and from
4390 GMP's mpz_t format aren't optimised nor are any montgomery forms
4391 cached, and on x86 it appears OpenSSL's own performance has caught up.
4392 However there are likely to be other architectures where GMP could
4393 provide a boost. This ENGINE is not built in by default, but it can be
4394 specified at Configure time and should be accompanied by the necessary
4395 linker additions, eg;
4396 ./config -DOPENSSL_USE_GMP -lgmp
4399 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4400 testing availability of engines with "-t" - the old behaviour is
4401 produced by increasing the feature's verbosity with "-tt".
4404 *) ECDSA routines: under certain error conditions uninitialized BN objects
4405 could be freed. Solution: make sure initialization is performed early
4406 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4410 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4411 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4412 software implementations. For DSA and DH, parameter generation can
4413 also be overriden by providing the appropriate method callbacks.
4416 *) Change the "progress" mechanism used in key-generation and
4417 primality testing to functions that take a new BN_GENCB pointer in
4418 place of callback/argument pairs. The new API functions have "_ex"
4419 postfixes and the older functions are reimplemented as wrappers for
4420 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4421 declarations of the old functions to help (graceful) attempts to
4422 migrate to the new functions. Also, the new key-generation API
4423 functions operate on a caller-supplied key-structure and return
4424 success/failure rather than returning a key or NULL - this is to
4425 help make "keygen" another member function of RSA_METHOD etc.
4427 Example for using the new callback interface:
4429 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4433 BN_GENCB_set(&my_cb, my_callback, my_arg);
4435 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4436 /* For the meaning of a, b in calls to my_callback(), see the
4437 * documentation of the function that calls the callback.
4438 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4439 * my_callback should return 1 if it wants BN_is_prime_ex()
4440 * to continue, or 0 to stop.
4445 *) Change the ZLIB compression method to be stateful, and make it
4446 available to TLS with the number defined in
4447 draft-ietf-tls-compression-04.txt.
4450 *) Add the ASN.1 structures and functions for CertificatePair, which
4451 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4453 CertificatePair ::= SEQUENCE {
4454 forward [0] Certificate OPTIONAL,
4455 reverse [1] Certificate OPTIONAL,
4456 -- at least one of the pair shall be present -- }
4458 Also implement the PEM functions to read and write certificate
4459 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4461 This needed to be defined, mostly for the sake of the LDAP
4462 attribute crossCertificatePair, but may prove useful elsewhere as
4466 *) Make it possible to inhibit symlinking of shared libraries in
4467 Makefile.shared, for Cygwin's sake.
4470 *) Extend the BIGNUM API by creating a function
4471 void BN_set_negative(BIGNUM *a, int neg);
4472 and a macro that behave like
4473 int BN_is_negative(const BIGNUM *a);
4475 to avoid the need to access 'a->neg' directly in applications.
4478 *) Implement fast modular reduction for pseudo-Mersenne primes
4479 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4480 EC_GROUP_new_curve_GFp() will now automatically use this
4482 [Nils Larsch <nla@trustcenter.de>]
4484 *) Add new lock type (CRYPTO_LOCK_BN).
4487 *) Change the ENGINE framework to automatically load engines
4488 dynamically from specific directories unless they could be
4489 found to already be built in or loaded. Move all the
4490 current engines except for the cryptodev one to a new
4492 The engines in engines/ are built as shared libraries if
4493 the "shared" options was given to ./Configure or ./config.
4494 Otherwise, they are inserted in libcrypto.a.
4495 /usr/local/ssl/engines is the default directory for dynamic
4496 engines, but that can be overriden at configure time through
4497 the usual use of --prefix and/or --openssldir, and at run
4498 time with the environment variable OPENSSL_ENGINES.
4499 [Geoff Thorpe and Richard Levitte]
4501 *) Add Makefile.shared, a helper makefile to build shared
4502 libraries. Addapt Makefile.org.
4505 *) Add version info to Win32 DLLs.
4506 [Peter 'Luna' Runestig" <peter@runestig.com>]
4508 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4509 can be added using this API to created arbitrary PKCS#12
4510 files while avoiding the low level API.
4512 New options to PKCS12_create(), key or cert can be NULL and
4513 will then be omitted from the output file. The encryption
4514 algorithm NIDs can be set to -1 for no encryption, the mac
4515 iteration count can be set to 0 to omit the mac.
4517 Enhance pkcs12 utility by making the -nokeys and -nocerts
4518 options work when creating a PKCS#12 file. New option -nomac
4519 to omit the mac, NONE can be set for an encryption algorithm.
4520 New code is modified to use the enhanced PKCS12_create()
4521 instead of the low level API.
4524 *) Extend ASN1 encoder to support indefinite length constructed
4525 encoding. This can output sequences tags and octet strings in
4526 this form. Modify pk7_asn1.c to support indefinite length
4527 encoding. This is experimental and needs additional code to
4528 be useful, such as an ASN1 bio and some enhanced streaming
4531 Extend template encode functionality so that tagging is passed
4532 down to the template encoder.
4535 *) Let 'openssl req' fail if an argument to '-newkey' is not
4536 recognized instead of using RSA as a default.
4539 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4540 As these are not official, they are not included in "ALL";
4541 the "ECCdraft" ciphersuite group alias can be used to select them.
4542 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4544 *) Add ECDH engine support.
4545 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4547 *) Add ECDH in new directory crypto/ecdh/.
4548 [Douglas Stebila (Sun Microsystems Laboratories)]
4550 *) Let BN_rand_range() abort with an error after 100 iterations
4551 without success (which indicates a broken PRNG).
4554 *) Change BN_mod_sqrt() so that it verifies that the input value
4555 is really the square of the return value. (Previously,
4556 BN_mod_sqrt would show GIGO behaviour.)
4559 *) Add named elliptic curves over binary fields from X9.62, SECG,
4560 and WAP/WTLS; add OIDs that were still missing.
4562 [Sheueling Chang Shantz and Douglas Stebila
4563 (Sun Microsystems Laboratories)]
4565 *) Extend the EC library for elliptic curves over binary fields
4566 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4569 EC_GF2m_simple_method
4573 EC_GROUP_new_curve_GF2m
4574 EC_GROUP_set_curve_GF2m
4575 EC_GROUP_get_curve_GF2m
4576 EC_POINT_set_affine_coordinates_GF2m
4577 EC_POINT_get_affine_coordinates_GF2m
4578 EC_POINT_set_compressed_coordinates_GF2m
4580 Point compression for binary fields is disabled by default for
4581 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4584 As binary polynomials are represented as BIGNUMs, various members
4585 of the EC_GROUP and EC_POINT data structures can be shared
4586 between the implementations for prime fields and binary fields;
4587 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4588 are essentially identical to their ..._GFp counterparts.
4589 (For simplicity, the '..._GFp' prefix has been dropped from
4590 various internal method names.)
4592 An internal 'field_div' method (similar to 'field_mul' and
4593 'field_sqr') has been added; this is used only for binary fields.
4595 [Sheueling Chang Shantz and Douglas Stebila
4596 (Sun Microsystems Laboratories)]
4598 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4599 through methods ('mul', 'precompute_mult').
4601 The generic implementations (now internally called 'ec_wNAF_mul'
4602 and 'ec_wNAF_precomputed_mult') remain the default if these
4603 methods are undefined.
4605 [Sheueling Chang Shantz and Douglas Stebila
4606 (Sun Microsystems Laboratories)]
4608 *) New function EC_GROUP_get_degree, which is defined through
4609 EC_METHOD. For curves over prime fields, this returns the bit
4610 length of the modulus.
4612 [Sheueling Chang Shantz and Douglas Stebila
4613 (Sun Microsystems Laboratories)]
4615 *) New functions EC_GROUP_dup, EC_POINT_dup.
4616 (These simply call ..._new and ..._copy).
4618 [Sheueling Chang Shantz and Douglas Stebila
4619 (Sun Microsystems Laboratories)]
4621 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4622 Polynomials are represented as BIGNUMs (where the sign bit is not
4623 used) in the following functions [macros]:
4626 BN_GF2m_sub [= BN_GF2m_add]
4627 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4628 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4629 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4631 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4632 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4633 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4634 BN_GF2m_cmp [= BN_ucmp]
4636 (Note that only the 'mod' functions are actually for fields GF(2^m).
4637 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4639 For some functions, an the irreducible polynomial defining a
4640 field can be given as an 'unsigned int[]' with strictly
4641 decreasing elements giving the indices of those bits that are set;
4642 i.e., p[] represents the polynomial
4643 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4645 p[0] > p[1] > ... > p[k] = 0.
4646 This applies to the following functions:
4651 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4652 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4654 BN_GF2m_mod_sqrt_arr
4655 BN_GF2m_mod_solve_quad_arr
4659 Conversion can be performed by the following functions:
4664 bntest.c has additional tests for binary polynomial arithmetic.
4666 Two implementations for BN_GF2m_mod_div() are available.
4667 The default algorithm simply uses BN_GF2m_mod_inv() and
4668 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4669 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4670 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4672 [Sheueling Chang Shantz and Douglas Stebila
4673 (Sun Microsystems Laboratories)]
4675 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4676 functionality is disabled at compile-time.
4677 [Douglas Stebila <douglas.stebila@sun.com>]
4679 *) Change default behaviour of 'openssl asn1parse' so that more
4680 information is visible when viewing, e.g., a certificate:
4682 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4683 mode the content of non-printable OCTET STRINGs is output in a
4684 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4685 avoid the appearance of a printable string.
4686 [Nils Larsch <nla@trustcenter.de>]
4688 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4690 EC_GROUP_set_asn1_flag()
4691 EC_GROUP_get_asn1_flag()
4692 EC_GROUP_set_point_conversion_form()
4693 EC_GROUP_get_point_conversion_form()
4694 These control ASN1 encoding details:
4695 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4696 has been set to OPENSSL_EC_NAMED_CURVE.
4697 - Points are encoded in uncompressed form by default; options for
4698 asn1_for are as for point2oct, namely
4699 POINT_CONVERSION_COMPRESSED
4700 POINT_CONVERSION_UNCOMPRESSED
4701 POINT_CONVERSION_HYBRID
4703 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4706 EC_GROUP_get0_seed()
4707 EC_GROUP_get_seed_len()
4708 This is used only for ASN1 purposes (so far).
4709 [Nils Larsch <nla@trustcenter.de>]
4711 *) Add 'field_type' member to EC_METHOD, which holds the NID
4712 of the appropriate field type OID. The new function
4713 EC_METHOD_get_field_type() returns this value.
4714 [Nils Larsch <nla@trustcenter.de>]
4719 EC_POINT_point2hex()
4720 EC_POINT_hex2point()
4721 providing useful interfaces to EC_POINT_point2oct() and
4722 EC_POINT_oct2point().
4723 [Nils Larsch <nla@trustcenter.de>]
4725 *) Change internals of the EC library so that the functions
4726 EC_GROUP_set_generator()
4727 EC_GROUP_get_generator()
4728 EC_GROUP_get_order()
4729 EC_GROUP_get_cofactor()
4730 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4731 to methods, which would lead to unnecessary code duplication when
4732 adding different types of curves.
4733 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4735 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4736 arithmetic, and such that modified wNAFs are generated
4737 (which avoid length expansion in many cases).
4740 *) Add a function EC_GROUP_check_discriminant() (defined via
4741 EC_METHOD) that verifies that the curve discriminant is non-zero.
4743 Add a function EC_GROUP_check() that makes some sanity tests
4744 on a EC_GROUP, its generator and order. This includes
4745 EC_GROUP_check_discriminant().
4746 [Nils Larsch <nla@trustcenter.de>]
4748 *) Add ECDSA in new directory crypto/ecdsa/.
4750 Add applications 'openssl ecparam' and 'openssl ecdsa'
4751 (these are based on 'openssl dsaparam' and 'openssl dsa').
4753 ECDSA support is also included in various other files across the
4754 library. Most notably,
4755 - 'openssl req' now has a '-newkey ecdsa:file' option;
4756 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4757 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4758 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4759 them suitable for ECDSA where domain parameters must be
4760 extracted before the specific public key;
4761 - ECDSA engine support has been added.
4762 [Nils Larsch <nla@trustcenter.de>]
4764 *) Include some named elliptic curves, and add OIDs from X9.62,
4765 SECG, and WAP/WTLS. Each curve can be obtained from the new
4767 EC_GROUP_new_by_curve_name(),
4768 and the list of available named curves can be obtained with
4769 EC_get_builtin_curves().
4770 Also add a 'curve_name' member to EC_GROUP objects, which can be
4772 EC_GROUP_set_curve_name()
4773 EC_GROUP_get_curve_name()
4774 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4776 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4777 was actually never needed) and in BN_mul(). The removal in BN_mul()
4778 required a small change in bn_mul_part_recursive() and the addition
4779 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4780 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4781 bn_sub_words() and bn_add_words() except they take arrays with
4785 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4787 *) Cleanse PEM buffers before freeing them since they may contain
4789 [Benjamin Bennett <ben@psc.edu>]
4791 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4792 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4793 authentication-only ciphersuites.
4796 *) Since AES128 and AES256 share a single mask bit in the logic of
4797 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4798 kludge to work properly if AES128 is available and AES256 isn't.
4801 *) Expand security boundary to match 1.1.1 module.
4804 *) Remove redundant features: hash file source, editing of test vectors
4805 modify fipsld to use external fips_premain.c signature.
4808 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4809 run algorithm test programs.
4812 *) Make algorithm test programs more tolerant of whitespace.
4815 *) Have SSL/TLS server implementation tolerate "mismatched" record
4816 protocol version while receiving ClientHello even if the
4817 ClientHello is fragmented. (The server can't insist on the
4818 particular protocol version it has chosen before the ServerHello
4819 message has informed the client about his choice.)
4822 *) Load error codes if they are not already present instead of using a
4823 static variable. This allows them to be cleanly unloaded and reloaded.
4826 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4828 *) Introduce limits to prevent malicious keys being able to
4829 cause a denial of service. (CVE-2006-2940)
4830 [Steve Henson, Bodo Moeller]
4832 *) Fix ASN.1 parsing of certain invalid structures that can result
4833 in a denial of service. (CVE-2006-2937) [Steve Henson]
4835 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4836 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4838 *) Fix SSL client code which could crash if connecting to a
4839 malicious SSLv2 server. (CVE-2006-4343)
4840 [Tavis Ormandy and Will Drewry, Google Security Team]
4842 *) Change ciphersuite string processing so that an explicit
4843 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4844 will no longer include "AES128-SHA"), and any other similar
4845 ciphersuite (same bitmap) from *other* protocol versions (so that
4846 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4847 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4848 changes from 0.9.8b and 0.9.8d.
4851 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4853 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4854 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4856 *) Change the Unix randomness entropy gathering to use poll() when
4857 possible instead of select(), since the latter has some
4858 undesirable limitations.
4859 [Darryl Miles via Richard Levitte and Bodo Moeller]
4861 *) Disable rogue ciphersuites:
4863 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4864 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4865 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4867 The latter two were purportedly from
4868 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4871 Also deactive the remaining ciphersuites from
4872 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4873 unofficial, and the ID has long expired.
4876 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4877 dual-core machines) and other potential thread-safety issues.
4880 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4882 *) Adapt fipsld and the build system to link against the validated FIPS
4883 module in FIPS mode.
4886 *) Fixes for VC++ 2005 build under Windows.
4889 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4890 from a Windows bash shell such as MSYS. It is autodetected from the
4891 "config" script when run from a VC++ environment. Modify standard VC++
4892 build to use fipscanister.o from the GNU make build.
4895 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4897 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4898 The value now differs depending on if you build for FIPS or not.
4899 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4900 safely run with a non-FIPSed libcrypto, as it may crash because of
4901 the difference induced by this change.
4904 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4906 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4907 (part of SSL_OP_ALL). This option used to disable the
4908 countermeasure against man-in-the-middle protocol-version
4909 rollback in the SSL 2.0 server implementation, which is a bad
4910 idea. (CVE-2005-2969)
4912 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4913 for Information Security, National Institute of Advanced Industrial
4914 Science and Technology [AIST], Japan)]
4916 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4917 mainly for FIPS compliance and not fully integrated at this stage.
4920 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4921 the exponentiation using a fixed-length exponent. (Otherwise,
4922 the information leaked through timing could expose the secret key
4923 after many signatures; cf. Bleichenbacher's attack on DSA with
4927 *) Make a new fixed-window mod_exp implementation the default for
4928 RSA, DSA, and DH private-key operations so that the sequence of
4929 squares and multiplies and the memory access pattern are
4930 independent of the particular secret key. This will mitigate
4931 cache-timing and potential related attacks.
4933 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4934 and this is automatically used by BN_mod_exp_mont() if the new flag
4935 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4936 will use this BN flag for private exponents unless the flag
4937 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4938 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4940 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4942 *) Change the client implementation for SSLv23_method() and
4943 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4944 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4945 (Previously, the SSL 2.0 backwards compatible Client Hello
4946 message format would be used even with SSL_OP_NO_SSLv2.)
4949 *) Add support for smime-type MIME parameter in S/MIME messages which some
4953 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4954 a threadsafe manner. Modify rsa code to use new function and add calls
4955 to dsa and dh code (which had race conditions before).
4958 *) Include the fixed error library code in the C error file definitions
4959 instead of fixing them up at runtime. This keeps the error code
4960 structures constant.
4963 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4965 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4968 *) Fixes for newer kerberos headers. NB: the casts are needed because
4969 the 'length' field is signed on one version and unsigned on another
4970 with no (?) obvious way to tell the difference, without these VC++
4971 complains. Also the "definition" of FAR (blank) is no longer included
4972 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4973 some needed definitions.
4976 *) Undo Cygwin change.
4979 *) Added support for proxy certificates according to RFC 3820.
4980 Because they may be a security thread to unaware applications,
4981 they must be explicitely allowed in run-time. See
4982 docs/HOWTO/proxy_certificates.txt for further information.
4985 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4987 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4988 server and client random values. Previously
4989 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4990 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4992 This change has negligible security impact because:
4994 1. Server and client random values still have 24 bytes of pseudo random
4997 2. Server and client random values are sent in the clear in the initial
5000 3. The master secret is derived using the premaster secret (48 bytes in
5001 size for static RSA ciphersuites) as well as client server and random
5004 The OpenSSL team would like to thank the UK NISCC for bringing this issue
5007 [Stephen Henson, reported by UK NISCC]
5009 *) Use Windows randomness collection on Cygwin.
5012 *) Fix hang in EGD/PRNGD query when communication socket is closed
5013 prematurely by EGD/PRNGD.
5014 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
5016 *) Prompt for pass phrases when appropriate for PKCS12 input format.
5019 *) Back-port of selected performance improvements from development
5020 branch, as well as improved support for PowerPC platforms.
5023 *) Add lots of checks for memory allocation failure, error codes to indicate
5024 failure and freeing up memory if a failure occurs.
5025 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
5027 *) Add new -passin argument to dgst.
5030 *) Perform some character comparisons of different types in X509_NAME_cmp:
5031 this is needed for some certificates that reencode DNs into UTF8Strings
5032 (in violation of RFC3280) and can't or wont issue name rollover
5036 *) Make an explicit check during certificate validation to see that
5037 the CA setting in each certificate on the chain is correct. As a
5038 side effect always do the following basic checks on extensions,
5039 not just when there's an associated purpose to the check:
5041 - if there is an unhandled critical extension (unless the user
5042 has chosen to ignore this fault)
5043 - if the path length has been exceeded (if one is set at all)
5044 - that certain extensions fit the associated purpose (if one has
5048 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5050 *) Avoid a race condition when CRLs are checked in a multi threaded
5051 environment. This would happen due to the reordering of the revoked
5052 entries during signature checking and serial number lookup. Now the
5053 encoding is cached and the serial number sort performed under a lock.
5054 Add new STACK function sk_is_sorted().
5057 *) Add Delta CRL to the extension code.
5060 *) Various fixes to s3_pkt.c so alerts are sent properly.
5061 [David Holmes <d.holmes@f5.com>]
5063 *) Reduce the chances of duplicate issuer name and serial numbers (in
5064 violation of RFC3280) using the OpenSSL certificate creation utilities.
5065 This is done by creating a random 64 bit value for the initial serial
5066 number when a serial number file is created or when a self signed
5067 certificate is created using 'openssl req -x509'. The initial serial
5068 number file is created using 'openssl x509 -next_serial' in CA.pl
5069 rather than being initialized to 1.
5072 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5074 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5075 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5076 [Joe Orton, Steve Henson]
5078 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5080 [Joe Orton, Steve Henson]
5082 *) Make it possible to have multiple active certificates with the same
5083 subject in the CA index file. This is done only if the keyword
5084 'unique_subject' is set to 'no' in the main CA section (default
5085 if 'CA_default') of the configuration file. The value is saved
5086 with the database itself in a separate index attribute file,
5087 named like the index file with '.attr' appended to the name.
5090 *) X509 verify fixes. Disable broken certificate workarounds when
5091 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5092 keyUsage extension present. Don't accept CRLs with unhandled critical
5093 extensions: since verify currently doesn't process CRL extensions this
5094 rejects a CRL with *any* critical extensions. Add new verify error codes
5098 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5099 A clarification of RFC2560 will require the use of OCTET STRINGs and
5100 some implementations cannot handle the current raw format. Since OpenSSL
5101 copies and compares OCSP nonces as opaque blobs without any attempt at
5102 parsing them this should not create any compatibility issues.
5105 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5106 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5107 this HMAC (and other) operations are several times slower than OpenSSL
5111 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5112 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5114 *) Use the correct content when signing type "other".
5117 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5119 *) Fix various bugs revealed by running the NISCC test suite:
5121 Stop out of bounds reads in the ASN1 code when presented with
5122 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5124 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5126 If verify callback ignores invalid public key errors don't try to check
5127 certificate signature with the NULL public key.
5131 *) New -ignore_err option in ocsp application to stop the server
5132 exiting on the first error in a request.
5135 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5136 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5140 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5141 extra data after the compression methods not only for TLS 1.0
5142 but also for SSL 3.0 (as required by the specification).
5143 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5145 *) Change X509_certificate_type() to mark the key as exported/exportable
5146 when it's 512 *bits* long, not 512 bytes.
5149 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5150 blocks during encryption.
5153 *) Various fixes to base64 BIO and non blocking I/O. On write
5154 flushes were not handled properly if the BIO retried. On read
5155 data was not being buffered properly and had various logic bugs.
5156 This also affects blocking I/O when the data being decoded is a
5160 *) Various S/MIME bugfixes and compatibility changes:
5161 output correct application/pkcs7 MIME type if
5162 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5163 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5164 of files as .eml work). Correctly handle very long lines in MIME
5168 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5170 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5171 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5172 a protocol version number mismatch like a decryption error
5173 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5176 *) Turn on RSA blinding by default in the default implementation
5177 to avoid a timing attack. Applications that don't want it can call
5178 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5179 They would be ill-advised to do so in most cases.
5180 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5182 *) Change RSA blinding code so that it works when the PRNG is not
5183 seeded (in this case, the secret RSA exponent is abused as
5184 an unpredictable seed -- if it is not unpredictable, there
5185 is no point in blinding anyway). Make RSA blinding thread-safe
5186 by remembering the creator's thread ID in rsa->blinding and
5187 having all other threads use local one-time blinding factors
5188 (this requires more computation than sharing rsa->blinding, but
5189 avoids excessive locking; and if an RSA object is not shared
5190 between threads, blinding will still be very fast).
5193 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5194 ENGINE as defaults for all supported algorithms irrespective of
5195 the 'flags' parameter. 'flags' is now honoured, so applications
5196 should make sure they are passing it correctly.
5199 *) Target "mingw" now allows native Windows code to be generated in
5200 the Cygwin environment as well as with the MinGW compiler.
5203 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5205 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5206 via timing by performing a MAC computation even if incorrrect
5207 block cipher padding has been found. This is a countermeasure
5208 against active attacks where the attacker has to distinguish
5209 between bad padding and a MAC verification error. (CVE-2003-0078)
5211 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5212 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5213 Martin Vuagnoux (EPFL, Ilion)]
5215 *) Make the no-err option work as intended. The intention with no-err
5216 is not to have the whole error stack handling routines removed from
5217 libcrypto, it's only intended to remove all the function name and
5218 reason texts, thereby removing some of the footprint that may not
5219 be interesting if those errors aren't displayed anyway.
5221 NOTE: it's still possible for any application or module to have it's
5222 own set of error texts inserted. The routines are there, just not
5223 used by default when no-err is given.
5226 *) Add support for FreeBSD on IA64.
5227 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5229 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5230 Kerberos function mit_des_cbc_cksum(). Before this change,
5231 the value returned by DES_cbc_cksum() was like the one from
5232 mit_des_cbc_cksum(), except the bytes were swapped.
5233 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5235 *) Allow an application to disable the automatic SSL chain building.
5236 Before this a rather primitive chain build was always performed in
5237 ssl3_output_cert_chain(): an application had no way to send the
5238 correct chain if the automatic operation produced an incorrect result.
5240 Now the chain builder is disabled if either:
5242 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5244 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5246 The reasoning behind this is that an application would not want the
5247 auto chain building to take place if extra chain certificates are
5248 present and it might also want a means of sending no additional
5249 certificates (for example the chain has two certificates and the
5253 *) Add the possibility to build without the ENGINE framework.
5254 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5256 *) Under Win32 gmtime() can return NULL: check return value in
5257 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5260 *) DSA routines: under certain error conditions uninitialized BN objects
5261 could be freed. Solution: make sure initialization is performed early
5262 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5263 Nils Larsch <nla@trustcenter.de> via PR#459)
5266 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5267 checked on reconnect on the client side, therefore session resumption
5268 could still fail with a "ssl session id is different" error. This
5269 behaviour is masked when SSL_OP_ALL is used due to
5270 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5271 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5272 followup to PR #377.
5275 *) IA-32 assembler support enhancements: unified ELF targets, support
5276 for SCO/Caldera platforms, fix for Cygwin shared build.
5279 *) Add support for FreeBSD on sparc64. As a consequence, support for
5280 FreeBSD on non-x86 processors is separate from x86 processors on
5281 the config script, much like the NetBSD support.
5282 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5284 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5286 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5289 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5290 code (06) was taken as the first octet of the session ID and the last
5291 octet was ignored consequently. As a result SSLv2 client side session
5292 caching could not have worked due to the session ID mismatch between
5294 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5298 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5299 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5303 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5304 seems that in spite of existing for more than a year, many application
5305 author have done nothing to provide the necessary callbacks, which
5306 means that this particular engine will not work properly anywhere.
5307 This is a very unfortunate situation which forces us, in the name
5308 of usability, to give the hw_ncipher.c a static lock, which is part
5310 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5311 appear in 0.9.8 or later. We EXPECT application authors to have
5312 dealt properly with this when 0.9.8 is released (unless we actually
5313 make such changes in the libcrypto locking code that changes will
5314 have to be made anyway).
5317 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5318 octets have been read, EOF or an error occurs. Without this change
5319 some truncated ASN1 structures will not produce an error.
5322 *) Disable Heimdal support, since it hasn't been fully implemented.
5323 Still give the possibility to force the use of Heimdal, but with
5324 warnings and a request that patches get sent to openssl-dev.
5327 *) Add the VC-CE target, introduce the WINCE sysname, and add
5328 INSTALL.WCE and appropriate conditionals to make it build.
5329 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5331 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5332 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5333 edit numbers of the version.
5334 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5336 *) Introduce safe string copy and catenation functions
5337 (BUF_strlcpy() and BUF_strlcat()).
5338 [Ben Laurie (CHATS) and Richard Levitte]
5340 *) Avoid using fixed-size buffers for one-line DNs.
5341 [Ben Laurie (CHATS)]
5343 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5344 resizing buffers containing secrets, and use where appropriate.
5345 [Ben Laurie (CHATS)]
5347 *) Avoid using fixed size buffers for configuration file location.
5348 [Ben Laurie (CHATS)]
5350 *) Avoid filename truncation for various CA files.
5351 [Ben Laurie (CHATS)]
5353 *) Use sizeof in preference to magic numbers.
5354 [Ben Laurie (CHATS)]
5356 *) Avoid filename truncation in cert requests.
5357 [Ben Laurie (CHATS)]
5359 *) Add assertions to check for (supposedly impossible) buffer
5361 [Ben Laurie (CHATS)]
5363 *) Don't cache truncated DNS entries in the local cache (this could
5364 potentially lead to a spoofing attack).
5365 [Ben Laurie (CHATS)]
5367 *) Fix various buffers to be large enough for hex/decimal
5368 representations in a platform independent manner.
5369 [Ben Laurie (CHATS)]
5371 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5372 resizing buffers containing secrets, and use where appropriate.
5373 [Ben Laurie (CHATS)]
5375 *) Add BIO_indent() to avoid much slightly worrying code to do
5377 [Ben Laurie (CHATS)]
5379 *) Convert sprintf()/BIO_puts() to BIO_printf().
5380 [Ben Laurie (CHATS)]
5382 *) buffer_gets() could terminate with the buffer only half
5384 [Ben Laurie (CHATS)]
5386 *) Add assertions to prevent user-supplied crypto functions from
5387 overflowing internal buffers by having large block sizes, etc.
5388 [Ben Laurie (CHATS)]
5390 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5392 [Ben Laurie (CHATS)]
5394 *) Eliminate unused copy of key in RC4.
5395 [Ben Laurie (CHATS)]
5397 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5398 [Ben Laurie (CHATS)]
5400 *) Fix off-by-one error in EGD path.
5401 [Ben Laurie (CHATS)]
5403 *) If RANDFILE path is too long, ignore instead of truncating.
5404 [Ben Laurie (CHATS)]
5406 *) Eliminate unused and incorrectly sized X.509 structure
5408 [Ben Laurie (CHATS)]
5410 *) Eliminate unused and dangerous function knumber().
5411 [Ben Laurie (CHATS)]
5413 *) Eliminate unused and dangerous structure, KSSL_ERR.
5414 [Ben Laurie (CHATS)]
5416 *) Protect against overlong session ID context length in an encoded
5417 session object. Since these are local, this does not appear to be
5419 [Ben Laurie (CHATS)]
5421 *) Change from security patch (see 0.9.6e below) that did not affect
5422 the 0.9.6 release series:
5424 Remote buffer overflow in SSL3 protocol - an attacker could
5425 supply an oversized master key in Kerberos-enabled versions.
5427 [Ben Laurie (CHATS)]
5429 *) Change the SSL kerb5 codes to match RFC 2712.
5432 *) Make -nameopt work fully for req and add -reqopt switch.
5433 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5435 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5436 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5438 *) Make sure tests can be performed even if the corresponding algorithms
5439 have been removed entirely. This was also the last step to make
5440 OpenSSL compilable with DJGPP under all reasonable conditions.
5441 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5443 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5444 to allow version independent disabling of normally unselected ciphers,
5445 which may be activated as a side-effect of selecting a single cipher.
5447 (E.g., cipher list string "RSA" enables ciphersuites that are left
5448 out of "ALL" because they do not provide symmetric encryption.
5449 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5450 [Lutz Jaenicke, Bodo Moeller]
5452 *) Add appropriate support for separate platform-dependent build
5453 directories. The recommended way to make a platform-dependent
5454 build directory is the following (tested on Linux), maybe with
5457 # Place yourself outside of the OpenSSL source tree. In
5458 # this example, the environment variable OPENSSL_SOURCE
5459 # is assumed to contain the absolute OpenSSL source directory.
5460 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5461 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5462 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5463 mkdir -p `dirname $F`
5464 ln -s $OPENSSL_SOURCE/$F $F
5467 To be absolutely sure not to disturb the source tree, a "make clean"
5468 is a good thing. If it isn't successfull, don't worry about it,
5469 it probably means the source directory is very clean.
5472 *) Make sure any ENGINE control commands make local copies of string
5473 pointers passed to them whenever necessary. Otherwise it is possible
5474 the caller may have overwritten (or deallocated) the original string
5475 data when a later ENGINE operation tries to use the stored values.
5476 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5478 *) Improve diagnostics in file reading and command-line digests.
5479 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5481 *) Add AES modes CFB and OFB to the object database. Correct an
5482 error in AES-CFB decryption.
5485 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5486 allows existing EVP_CIPHER_CTX structures to be reused after
5487 calling EVP_*Final(). This behaviour is used by encryption
5488 BIOs and some applications. This has the side effect that
5489 applications must explicitly clean up cipher contexts with
5490 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5493 *) Check the values of dna and dnb in bn_mul_recursive before calling
5494 bn_mul_comba (a non zero value means the a or b arrays do not contain
5495 n2 elements) and fallback to bn_mul_normal if either is not zero.
5498 *) Fix escaping of non-ASCII characters when using the -subj option
5499 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5502 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5503 form for "surname", serialNumber has no short form.
5504 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5505 therefore remove "mail" short name for "internet 7".
5506 The OID for unique identifiers in X509 certificates is
5507 x500UniqueIdentifier, not uniqueIdentifier.
5508 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5511 *) Add an "init" command to the ENGINE config module and auto initialize
5512 ENGINEs. Without any "init" command the ENGINE will be initialized
5513 after all ctrl commands have been executed on it. If init=1 the
5514 ENGINE is initailized at that point (ctrls before that point are run
5515 on the uninitialized ENGINE and after on the initialized one). If
5516 init=0 then the ENGINE will not be iniatialized at all.
5519 *) Fix the 'app_verify_callback' interface so that the user-defined
5520 argument is actually passed to the callback: In the
5521 SSL_CTX_set_cert_verify_callback() prototype, the callback
5522 declaration has been changed from
5525 int (*cb)(X509_STORE_CTX *,void *);
5526 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5527 i=s->ctx->app_verify_callback(&ctx)
5528 has been changed into
5529 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5531 To update applications using SSL_CTX_set_cert_verify_callback(),
5532 a dummy argument can be added to their callback functions.
5533 [D. K. Smetters <smetters@parc.xerox.com>]
5535 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5536 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5538 *) Add and OPENSSL_LOAD_CONF define which will cause
5539 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5540 This allows older applications to transparently support certain
5541 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5542 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5543 load the config file and OPENSSL_add_all_algorithms_conf() which will
5544 always load it have also been added.
5547 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5548 Adjust NIDs and EVP layer.
5549 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5551 *) Config modules support in openssl utility.
5553 Most commands now load modules from the config file,
5554 though in a few (such as version) this isn't done
5555 because it couldn't be used for anything.
5557 In the case of ca and req the config file used is
5558 the same as the utility itself: that is the -config
5559 command line option can be used to specify an
5563 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5564 use "openssl_conf" if filename is NULL use default openssl config file.
5567 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5568 config section name. Add a new flag to tolerate a missing config file
5569 and move code to CONF_modules_load_file().
5572 *) Support for crypto accelerator cards from Accelerated Encryption
5573 Processing, www.aep.ie. (Use engine 'aep')
5574 The support was copied from 0.9.6c [engine] and adapted/corrected
5575 to work with the new engine framework.
5576 [AEP Inc. and Richard Levitte]
5578 *) Support for SureWare crypto accelerator cards from Baltimore
5579 Technologies. (Use engine 'sureware')
5580 The support was copied from 0.9.6c [engine] and adapted
5581 to work with the new engine framework.
5584 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5585 make the newer ENGINE framework commands for the CHIL engine work.
5586 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5588 *) Make it possible to produce shared libraries on ReliantUNIX.
5589 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5591 *) Add the configuration target debug-linux-ppro.
5592 Make 'openssl rsa' use the general key loading routines
5593 implemented in apps.c, and make those routines able to
5594 handle the key format FORMAT_NETSCAPE and the variant
5596 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5598 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5599 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5601 *) Add -keyform to rsautl, and document -engine.
5602 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5604 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5605 BIO_R_NO_SUCH_FILE error code rather than the generic
5606 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5609 *) Add new functions
5611 ERR_peek_last_error_line
5612 ERR_peek_last_error_line_data.
5613 These are similar to
5616 ERR_peek_error_line_data,
5617 but report on the latest error recorded rather than the first one
5618 still in the error queue.
5619 [Ben Laurie, Bodo Moeller]
5621 *) default_algorithms option in ENGINE config module. This allows things
5623 default_algorithms = ALL
5624 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5627 *) Prelminary ENGINE config module.
5630 *) New experimental application configuration code.
5633 *) Change the AES code to follow the same name structure as all other
5634 symmetric ciphers, and behave the same way. Move everything to
5635 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5636 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5638 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5639 [Ben Laurie and Theo de Raadt]
5641 *) Add option to output public keys in req command.
5642 [Massimiliano Pala madwolf@openca.org]
5644 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5645 (up to about 10% better than before for P-192 and P-224).
5648 *) New functions/macros
5650 SSL_CTX_set_msg_callback(ctx, cb)
5651 SSL_CTX_set_msg_callback_arg(ctx, arg)
5652 SSL_set_msg_callback(ssl, cb)
5653 SSL_set_msg_callback_arg(ssl, arg)
5655 to request calling a callback function
5657 void cb(int write_p, int version, int content_type,
5658 const void *buf, size_t len, SSL *ssl, void *arg)
5660 whenever a protocol message has been completely received
5661 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5662 protocol version according to which the SSL library interprets
5663 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5664 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5665 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5666 specification (change_cipher_spec(20), alert(21), handshake(22)).
5667 'buf' and 'len' point to the actual message, 'ssl' to the
5668 SSL object, and 'arg' is the application-defined value set by
5669 SSL[_CTX]_set_msg_callback_arg().
5671 'openssl s_client' and 'openssl s_server' have new '-msg' options
5672 to enable a callback that displays all protocol messages.
5675 *) Change the shared library support so shared libraries are built as
5676 soon as the corresponding static library is finished, and thereby get
5677 openssl and the test programs linked against the shared library.
5678 This still only happens when the keyword "shard" has been given to
5679 the configuration scripts.
5681 NOTE: shared library support is still an experimental thing, and
5682 backward binary compatibility is still not guaranteed.
5683 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5685 *) Add support for Subject Information Access extension.
5686 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5688 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5689 additional bytes when new memory had to be allocated, not just
5690 when reusing an existing buffer.
5693 *) New command line and configuration option 'utf8' for the req command.
5694 This allows field values to be specified as UTF8 strings.
5697 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5698 runs for the former and machine-readable output for the latter.
5701 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5702 of the e-mail address in the DN (i.e., it will go into a certificate
5703 extension only). The new configuration file option 'email_in_dn = no'
5704 has the same effect.
5705 [Massimiliano Pala madwolf@openca.org]
5707 *) Change all functions with names starting with des_ to be starting
5708 with DES_ instead. Add wrappers that are compatible with libdes,
5709 but are named _ossl_old_des_*. Finally, add macros that map the
5710 des_* symbols to the corresponding _ossl_old_des_* if libdes
5711 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5712 desired, the des_* symbols will be mapped to DES_*, with one
5715 Since we provide two compatibility mappings, the user needs to
5716 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5717 compatibility is desired. The default (i.e., when that macro
5718 isn't defined) is OpenSSL 0.9.6c compatibility.
5720 There are also macros that enable and disable the support of old
5721 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5722 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5723 are defined, the default will apply: to support the old des routines.
5725 In either case, one must include openssl/des.h to get the correct
5726 definitions. Do not try to just include openssl/des_old.h, that
5729 NOTE: This is a major break of an old API into a new one. Software
5730 authors are encouraged to switch to the DES_ style functions. Some
5731 time in the future, des_old.h and the libdes compatibility functions
5732 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5733 default), and then completely removed.
5736 *) Test for certificates which contain unsupported critical extensions.
5737 If such a certificate is found during a verify operation it is
5738 rejected by default: this behaviour can be overridden by either
5739 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5740 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5741 X509_supported_extension() has also been added which returns 1 if a
5742 particular extension is supported.
5745 *) Modify the behaviour of EVP cipher functions in similar way to digests
5746 to retain compatibility with existing code.
5749 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5750 compatibility with existing code. In particular the 'ctx' parameter does
5751 not have to be to be initialized before the call to EVP_DigestInit() and
5752 it is tidied up after a call to EVP_DigestFinal(). New function
5753 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5754 EVP_MD_CTX_copy() changed to not require the destination to be
5755 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5756 requires the destination to be valid.
5758 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5759 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5762 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5763 so that complete 'Handshake' protocol structures are kept in memory
5764 instead of overwriting 'msg_type' and 'length' with 'body' data.
5767 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5768 [Massimo Santin via Richard Levitte]
5770 *) Major restructuring to the underlying ENGINE code. This includes
5771 reduction of linker bloat, separation of pure "ENGINE" manipulation
5772 (initialisation, etc) from functionality dealing with implementations
5773 of specific crypto iterfaces. This change also introduces integrated
5774 support for symmetric ciphers and digest implementations - so ENGINEs
5775 can now accelerate these by providing EVP_CIPHER and EVP_MD
5776 implementations of their own. This is detailed in crypto/engine/README
5777 as it couldn't be adequately described here. However, there are a few
5778 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5779 were changed in the original introduction of ENGINE code have now
5780 reverted back - the hooking from this code to ENGINE is now a good
5781 deal more passive and at run-time, operations deal directly with
5782 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5783 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5784 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5785 they were not being used by the framework as there is no concept of a
5786 BIGNUM_METHOD and they could not be generalised to the new
5787 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5788 ENGINE_cpy() has been removed as it cannot be consistently defined in
5792 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5795 *) Change mkdef.pl to sort symbols that get the same entry number,
5796 and make sure the automatically generated functions ERR_load_*
5797 become part of libeay.num as well.
5800 *) New function SSL_renegotiate_pending(). This returns true once
5801 renegotiation has been requested (either SSL_renegotiate() call
5802 or HelloRequest/ClientHello receveived from the peer) and becomes
5803 false once a handshake has been completed.
5804 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5805 sends a HelloRequest, but does not ensure that a handshake takes
5806 place. SSL_renegotiate_pending() is useful for checking if the
5807 client has followed the request.)
5810 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5811 By default, clients may request session resumption even during
5812 renegotiation (if session ID contexts permit); with this option,
5813 session resumption is possible only in the first handshake.
5815 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5816 more bits available for options that should not be part of
5817 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5820 *) Add some demos for certificate and certificate request creation.
5823 *) Make maximum certificate chain size accepted from the peer application
5824 settable (SSL*_get/set_max_cert_list()), as proposed by
5825 "Douglas E. Engert" <deengert@anl.gov>.
5828 *) Add support for shared libraries for Unixware-7
5829 (Boyd Lynn Gerber <gerberb@zenez.com>).
5832 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5833 be done prior to destruction. Use this to unload error strings from
5834 ENGINEs that load their own error strings. NB: This adds two new API
5835 functions to "get" and "set" this destroy handler in an ENGINE.
5838 *) Alter all existing ENGINE implementations (except "openssl" and
5839 "openbsd") to dynamically instantiate their own error strings. This
5840 makes them more flexible to be built both as statically-linked ENGINEs
5841 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5842 Also, add stub code to each that makes building them as self-contained
5843 shared-libraries easier (see README.ENGINE).
5846 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5847 implementations into applications that are completely implemented in
5848 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5849 commands that can be used to configure what shared-library to load and
5850 to control aspects of the way it is handled. Also, made an update to
5851 the README.ENGINE file that brings its information up-to-date and
5852 provides some information and instructions on the "dynamic" ENGINE
5853 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5856 *) Make it possible to unload ranges of ERR strings with a new
5857 "ERR_unload_strings" function.
5860 *) Add a copy() function to EVP_MD.
5863 *) Make EVP_MD routines take a context pointer instead of just the
5864 md_data void pointer.
5867 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5868 that the digest can only process a single chunk of data
5869 (typically because it is provided by a piece of
5870 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5871 is only going to provide a single chunk of data, and hence the
5872 framework needn't accumulate the data for oneshot drivers.
5875 *) As with "ERR", make it possible to replace the underlying "ex_data"
5876 functions. This change also alters the storage and management of global
5877 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5878 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5879 index counters. The API functions that use this state have been changed
5880 to take a "class_index" rather than pointers to the class's local STACK
5881 and counter, and there is now an API function to dynamically create new
5882 classes. This centralisation allows us to (a) plug a lot of the
5883 thread-safety problems that existed, and (b) makes it possible to clean
5884 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5885 such data would previously have always leaked in application code and
5886 workarounds were in place to make the memory debugging turn a blind eye
5887 to it. Application code that doesn't use this new function will still
5888 leak as before, but their memory debugging output will announce it now
5889 rather than letting it slide.
5891 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5892 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5893 has a return value to indicate success or failure.
5896 *) Make it possible to replace the underlying "ERR" functions such that the
5897 global state (2 LHASH tables and 2 locks) is only used by the "default"
5898 implementation. This change also adds two functions to "get" and "set"
5899 the implementation prior to it being automatically set the first time
5900 any other ERR function takes place. Ie. an application can call "get",
5901 pass the return value to a module it has just loaded, and that module
5902 can call its own "set" function using that value. This means the
5903 module's "ERR" operations will use (and modify) the error state in the
5904 application and not in its own statically linked copy of OpenSSL code.
5907 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5908 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5909 the operation, and provides a more encapsulated way for external code
5910 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5911 to use these functions rather than manually incrementing the counts.
5913 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5916 *) Add EVP test program.
5919 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5922 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5923 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5924 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5925 These allow a CRL to be built without having to access X509_CRL fields
5926 directly. Modify 'ca' application to use new functions.
5929 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5930 bug workarounds. Rollback attack detection is a security feature.
5931 The problem will only arise on OpenSSL servers when TLSv1 is not
5932 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5933 Software authors not wanting to support TLSv1 will have special reasons
5934 for their choice and can explicitly enable this option.
5935 [Bodo Moeller, Lutz Jaenicke]
5937 *) Rationalise EVP so it can be extended: don't include a union of
5938 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5939 (similar to those existing for EVP_CIPHER_CTX).
5944 EVP_MD_CTX_init(&md); /* new function call */
5945 EVP_DigestInit(&md, EVP_sha1());
5946 EVP_DigestUpdate(&md, in, len);
5947 EVP_DigestFinal(&md, out, NULL);
5948 EVP_MD_CTX_cleanup(&md); /* new function call */
5952 *) Make DES key schedule conform to the usual scheme, as well as
5953 correcting its structure. This means that calls to DES functions
5954 now have to pass a pointer to a des_key_schedule instead of a
5955 plain des_key_schedule (which was actually always a pointer
5958 des_key_schedule ks;
5960 des_set_key_checked(..., &ks);
5961 des_ncbc_encrypt(..., &ks, ...);
5963 (Note that a later change renames 'des_...' into 'DES_...'.)
5966 *) Initial reduction of linker bloat: the use of some functions, such as
5967 PEM causes large amounts of unused functions to be linked in due to
5968 poor organisation. For example pem_all.c contains every PEM function
5969 which has a knock on effect of linking in large amounts of (unused)
5970 ASN1 code. Grouping together similar functions and splitting unrelated
5971 functions prevents this.
5974 *) Cleanup of EVP macros.
5977 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5978 correct _ecb suffix.
5981 *) Add initial OCSP responder support to ocsp application. The
5982 revocation information is handled using the text based index
5983 use by the ca application. The responder can either handle
5984 requests generated internally, supplied in files (for example
5985 via a CGI script) or using an internal minimal server.
5988 *) Add configuration choices to get zlib compression for TLS.
5991 *) Changes to Kerberos SSL for RFC 2712 compliance:
5992 1. Implemented real KerberosWrapper, instead of just using
5993 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5994 2. Implemented optional authenticator field of KerberosWrapper.
5996 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5997 and authenticator structs; see crypto/krb5/.
5999 Generalized Kerberos calls to support multiple Kerberos libraries.
6000 [Vern Staats <staatsvr@asc.hpc.mil>,
6001 Jeffrey Altman <jaltman@columbia.edu>
6002 via Richard Levitte]
6004 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
6005 already does with RSA. testdsa.h now has 'priv_key/pub_key'
6006 values for each of the key sizes rather than having just
6007 parameters (and 'speed' generating keys each time).
6010 *) Speed up EVP routines.
6013 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
6014 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
6015 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
6016 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
6018 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
6019 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
6020 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
6023 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
6025 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
6028 *) Added the OS2-EMX target.
6029 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
6031 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
6032 to support NCONF routines in extension code. New function CONF_set_nconf()
6033 to allow functions which take an NCONF to also handle the old LHASH
6034 structure: this means that the old CONF compatible routines can be
6035 retained (in particular wrt extensions) without having to duplicate the
6036 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
6039 *) Enhance the general user interface with mechanisms for inner control
6040 and with possibilities to have yes/no kind of prompts.
6043 *) Change all calls to low level digest routines in the library and
6044 applications to use EVP. Add missing calls to HMAC_cleanup() and
6045 don't assume HMAC_CTX can be copied using memcpy().
6046 [Verdon Walker <VWalker@novell.com>, Steve Henson]
6048 *) Add the possibility to control engines through control names but with
6049 arbitrary arguments instead of just a string.
6050 Change the key loaders to take a UI_METHOD instead of a callback
6051 function pointer. NOTE: this breaks binary compatibility with earlier
6052 versions of OpenSSL [engine].
6053 Adapt the nCipher code for these new conditions and add a card insertion
6057 *) Enhance the general user interface with mechanisms to better support
6058 dialog box interfaces, application-defined prompts, the possibility
6059 to use defaults (for example default passwords from somewhere else)
6060 and interrupts/cancellations.
6063 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6064 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6067 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6068 tidy up some unnecessarily weird code in 'sk_new()').
6069 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6071 *) Change the key loading routines for ENGINEs to use the same kind
6072 callback (pem_password_cb) as all other routines that need this
6076 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6077 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6078 than this minimum value is recommended.
6081 *) New random seeder for OpenVMS, using the system process statistics
6082 that are easily reachable.
6085 *) Windows apparently can't transparently handle global
6086 variables defined in DLLs. Initialisations such as:
6088 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6090 wont compile. This is used by the any applications that need to
6091 declare their own ASN1 modules. This was fixed by adding the option
6092 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6093 needed for static libraries under Win32.
6096 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6097 setting of purpose and trust fields. New X509_STORE trust and
6098 purpose functions and tidy up setting in other SSL functions.
6101 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6102 structure. These are inherited by X509_STORE_CTX when it is
6103 initialised. This allows various defaults to be set in the
6104 X509_STORE structure (such as flags for CRL checking and custom
6105 purpose or trust settings) for functions which only use X509_STORE_CTX
6106 internally such as S/MIME.
6108 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6109 trust settings if they are not set in X509_STORE. This allows X509_STORE
6110 purposes and trust (in S/MIME for example) to override any set by default.
6112 Add command line options for CRL checking to smime, s_client and s_server
6116 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6117 are set then the CRL is looked up in the X509_STORE structure and
6118 its validity and signature checked, then if the certificate is found
6119 in the CRL the verify fails with a revoked error.
6121 Various new CRL related callbacks added to X509_STORE_CTX structure.
6123 Command line options added to 'verify' application to support this.
6125 This needs some additional work, such as being able to handle multiple
6126 CRLs with different times, extension based lookup (rather than just
6127 by subject name) and ultimately more complete V2 CRL extension
6131 *) Add a general user interface API (crypto/ui/). This is designed
6132 to replace things like des_read_password and friends (backward
6133 compatibility functions using this new API are provided).
6134 The purpose is to remove prompting functions from the DES code
6135 section as well as provide for prompting through dialog boxes in
6136 a window system and the like.
6139 *) Add "ex_data" support to ENGINE so implementations can add state at a
6140 per-structure level rather than having to store it globally.
6143 *) Make it possible for ENGINE structures to be copied when retrieved by
6144 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6145 This causes the "original" ENGINE structure to act like a template,
6146 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6147 operational state can be localised to each ENGINE structure, despite the
6148 fact they all share the same "methods". New ENGINE structures returned in
6149 this case have no functional references and the return value is the single
6150 structural reference. This matches the single structural reference returned
6151 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6155 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6156 needs to match any other type at all we need to manually clear the
6160 *) Changes to the "openssl engine" utility to include;
6161 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6162 about an ENGINE's available control commands.
6163 - executing control commands from command line arguments using the
6164 '-pre' and '-post' switches. '-post' is only used if '-t' is
6165 specified and the ENGINE is successfully initialised. The syntax for
6166 the individual commands are colon-separated, for example;
6167 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6170 *) New dynamic control command support for ENGINEs. ENGINEs can now
6171 declare their own commands (numbers), names (strings), descriptions,
6172 and input types for run-time discovery by calling applications. A
6173 subset of these commands are implicitly classed as "executable"
6174 depending on their input type, and only these can be invoked through
6175 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6176 can be based on user input, config files, etc). The distinction is
6177 that "executable" commands cannot return anything other than a boolean
6178 result and can only support numeric or string input, whereas some
6179 discoverable commands may only be for direct use through
6180 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6181 pointers, or other custom uses. The "executable" commands are to
6182 support parameterisations of ENGINE behaviour that can be
6183 unambiguously defined by ENGINEs and used consistently across any
6184 OpenSSL-based application. Commands have been added to all the
6185 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6186 control over shared-library paths without source code alterations.
6189 *) Changed all ENGINE implementations to dynamically allocate their
6190 ENGINEs rather than declaring them statically. Apart from this being
6191 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6192 this also allows the implementations to compile without using the
6193 internal engine_int.h header.
6196 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6197 'const' value. Any code that should be able to modify a RAND_METHOD
6198 should already have non-const pointers to it (ie. they should only
6199 modify their own ones).
6202 *) Made a variety of little tweaks to the ENGINE code.
6203 - "atalla" and "ubsec" string definitions were moved from header files
6204 to C code. "nuron" string definitions were placed in variables
6205 rather than hard-coded - allowing parameterisation of these values
6206 later on via ctrl() commands.
6207 - Removed unused "#if 0"'d code.
6208 - Fixed engine list iteration code so it uses ENGINE_free() to release
6209 structural references.
6210 - Constified the RAND_METHOD element of ENGINE structures.
6211 - Constified various get/set functions as appropriate and added
6212 missing functions (including a catch-all ENGINE_cpy that duplicates
6213 all ENGINE values onto a new ENGINE except reference counts/state).
6214 - Removed NULL parameter checks in get/set functions. Setting a method
6215 or function to NULL is a way of cancelling out a previously set
6216 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6217 and doesn't justify the extra error symbols and code.
6218 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6219 flags from engine_int.h to engine.h.
6220 - Changed prototypes for ENGINE handler functions (init(), finish(),
6221 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6224 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6225 to the algorithm using long division. The binary algorithm can be
6226 used only if the modulus is odd. On 32-bit systems, it is faster
6227 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6228 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6229 up to 450 bits. In 64-bit environments, the binary algorithm
6230 appears to be advantageous for much longer moduli; here we use it
6231 for moduli up to 2048 bits.
6234 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6235 could not support the combine flag in choice fields.
6238 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6239 extensions from a certificate request to the certificate.
6242 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6243 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6244 file: this allows the display of the certificate about to be
6245 signed to be customised, to allow certain fields to be included
6246 or excluded and extension details. The old system didn't display
6247 multicharacter strings properly, omitted fields not in the policy
6248 and couldn't display additional details such as extensions.
6251 *) Function EC_POINTs_mul for multiple scalar multiplication
6252 of an arbitrary number of elliptic curve points
6253 \sum scalars[i]*points[i],
6254 optionally including the generator defined for the EC_GROUP:
6255 scalar*generator + \sum scalars[i]*points[i].
6257 EC_POINT_mul is a simple wrapper function for the typical case
6258 that the point list has just one item (besides the optional
6262 *) First EC_METHODs for curves over GF(p):
6264 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6265 operations and provides various method functions that can also
6266 operate with faster implementations of modular arithmetic.
6268 EC_GFp_mont_method() reuses most functions that are part of
6269 EC_GFp_simple_method, but uses Montgomery arithmetic.
6271 [Bodo Moeller; point addition and point doubling
6272 implementation directly derived from source code provided by
6273 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6275 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6276 crypto/ec/ec_lib.c):
6278 Curves are EC_GROUP objects (with an optional group generator)
6279 based on EC_METHODs that are built into the library.
6281 Points are EC_POINT objects based on EC_GROUP objects.
6283 Most of the framework would be able to handle curves over arbitrary
6284 finite fields, but as there are no obvious types for fields other
6285 than GF(p), some functions are limited to that for now.
6288 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6289 that the file contains a complete HTTP response.
6292 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6293 change the def and num file printf format specifier from "%-40sXXX"
6294 to "%-39s XXX". The latter will always guarantee a space after the
6295 field while the former will cause them to run together if the field
6296 is 40 of more characters long.
6299 *) Constify the cipher and digest 'method' functions and structures
6300 and modify related functions to take constant EVP_MD and EVP_CIPHER
6304 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6305 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6308 *) Modify EVP_Digest*() routines so they now return values. Although the
6309 internal software routines can never fail additional hardware versions
6313 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6315 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6316 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6319 ERR_R_NESTED_ASN1_ERROR
6321 ERR_R_MISSING_ASN1_EOS
6322 were 4 .. 9, conflicting with
6323 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6325 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6326 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6328 Add new error code 'ERR_R_INTERNAL_ERROR'.
6331 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6335 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6336 sets the subject name for a new request or supersedes the
6337 subject name in a given request. Formats that can be parsed are
6338 'CN=Some Name, OU=myOU, C=IT'
6340 'CN=Some Name/OU=myOU/C=IT'.
6342 Add options '-batch' and '-verbose' to 'openssl req'.
6343 [Massimiliano Pala <madwolf@hackmasters.net>]
6345 *) Introduce the possibility to access global variables through
6346 functions on platform were that's the best way to handle exporting
6347 global variables in shared libraries. To enable this functionality,
6348 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6349 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6350 is normally done by Configure or something similar).
6352 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6353 in the source file (foo.c) like this:
6355 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6356 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6358 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6359 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6361 OPENSSL_DECLARE_GLOBAL(int,foo);
6362 #define foo OPENSSL_GLOBAL_REF(foo)
6363 OPENSSL_DECLARE_GLOBAL(double,bar);
6364 #define bar OPENSSL_GLOBAL_REF(bar)
6366 The #defines are very important, and therefore so is including the
6367 header file everywhere where the defined globals are used.
6369 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6370 of ASN.1 items, but that structure is a bit different.
6372 The largest change is in util/mkdef.pl which has been enhanced with
6373 better and easier to understand logic to choose which symbols should
6374 go into the Windows .def files as well as a number of fixes and code
6375 cleanup (among others, algorithm keywords are now sorted
6376 lexicographically to avoid constant rewrites).
6379 *) In BN_div() keep a copy of the sign of 'num' before writing the
6380 result to 'rm' because if rm==num the value will be overwritten
6381 and produce the wrong result if 'num' is negative: this caused
6382 problems with BN_mod() and BN_nnmod().
6385 *) Function OCSP_request_verify(). This checks the signature on an
6386 OCSP request and verifies the signer certificate. The signer
6387 certificate is just checked for a generic purpose and OCSP request
6391 *) Add OCSP_check_validity() function to check the validity of OCSP
6392 responses. OCSP responses are prepared in real time and may only
6393 be a few seconds old. Simply checking that the current time lies
6394 between thisUpdate and nextUpdate max reject otherwise valid responses
6395 caused by either OCSP responder or client clock inaccuracy. Instead
6396 we allow thisUpdate and nextUpdate to fall within a certain period of
6397 the current time. The age of the response can also optionally be
6398 checked. Two new options -validity_period and -status_age added to
6402 *) If signature or public key algorithm is unrecognized print out its
6403 OID rather that just UNKNOWN.
6406 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6407 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6408 ID to be generated from the issuer certificate alone which can then be
6409 passed to OCSP_id_issuer_cmp().
6412 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6413 ASN1 modules to export functions returning ASN1_ITEM pointers
6414 instead of the ASN1_ITEM structures themselves. This adds several
6415 new macros which allow the underlying ASN1 function/structure to
6416 be accessed transparently. As a result code should not use ASN1_ITEM
6417 references directly (such as &X509_it) but instead use the relevant
6418 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6419 use of the new ASN1 code on platforms where exporting structures
6420 is problematical (for example in shared libraries) but exporting
6421 functions returning pointers to structures is not.
6424 *) Add support for overriding the generation of SSL/TLS session IDs.
6425 These callbacks can be registered either in an SSL_CTX or per SSL.
6426 The purpose of this is to allow applications to control, if they wish,
6427 the arbitrary values chosen for use as session IDs, particularly as it
6428 can be useful for session caching in multiple-server environments. A
6429 command-line switch for testing this (and any client code that wishes
6430 to use such a feature) has been added to "s_server".
6431 [Geoff Thorpe, Lutz Jaenicke]
6433 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6434 of the form '#if defined(...) || defined(...) || ...' and
6435 '#if !defined(...) && !defined(...) && ...'. This also avoids
6436 the growing number of special cases it was previously handling.
6439 *) Make all configuration macros available for application by making
6440 sure they are available in opensslconf.h, by giving them names starting
6441 with "OPENSSL_" to avoid conflicts with other packages and by making
6442 sure e_os2.h will cover all platform-specific cases together with
6444 Additionally, it is now possible to define configuration/platform-
6445 specific names (called "system identities"). In the C code, these
6446 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6447 macro with the name beginning with "OPENSSL_SYS_", which is determined
6448 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6452 *) New option -set_serial to 'req' and 'x509' this allows the serial
6453 number to use to be specified on the command line. Previously self
6454 signed certificates were hard coded with serial number 0 and the
6455 CA options of 'x509' had to use a serial number in a file which was
6459 *) New options to 'ca' utility to support V2 CRL entry extensions.
6460 Currently CRL reason, invalidity date and hold instruction are
6461 supported. Add new CRL extensions to V3 code and some new objects.
6464 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6465 disable standard block padding (aka PKCS#5 padding) in the EVP
6466 API, which was previously mandatory. This means that the data is
6467 not padded in any way and so the total length much be a multiple
6468 of the block size, otherwise an error occurs.
6471 *) Initial (incomplete) OCSP SSL support.
6474 *) New function OCSP_parse_url(). This splits up a URL into its host,
6475 port and path components: primarily to parse OCSP URLs. New -url
6476 option to ocsp utility.
6479 *) New nonce behavior. The return value of OCSP_check_nonce() now
6480 reflects the various checks performed. Applications can decide
6481 whether to tolerate certain situations such as an absent nonce
6482 in a response when one was present in a request: the ocsp application
6483 just prints out a warning. New function OCSP_add1_basic_nonce()
6484 this is to allow responders to include a nonce in a response even if
6485 the request is nonce-less.
6488 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6489 skipped when using openssl x509 multiple times on a single input file,
6490 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6493 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6494 set string type: to handle setting ASN1_TIME structures. Fix ca
6495 utility to correctly initialize revocation date of CRLs.
6498 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6499 the clients preferred ciphersuites and rather use its own preferences.
6500 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6501 Internet Explorer by ensuring unchanged hash method during stepup.
6502 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6505 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6506 to aes and add a new 'exist' option to print out symbols that don't
6510 *) Additional options to ocsp utility to allow flags to be set and
6511 additional certificates supplied.
6514 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6515 OCSP client a number of certificate to only verify the response
6519 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6520 handle the new API. Currently only ECB, CBC modes supported. Add new
6523 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6524 Encryption Standard (AES) Ciphersuites for Transport Layer
6525 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6526 not enabled by default and were not part of the "ALL" ciphersuite
6527 alias because they were not yet official; they could be
6528 explicitly requested by specifying the "AESdraft" ciphersuite
6529 group alias. In the final release of OpenSSL 0.9.7, the group
6530 alias is called "AES" and is part of "ALL".)
6531 [Ben Laurie, Steve Henson, Bodo Moeller]
6533 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6534 request to response.
6537 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6538 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6539 extract information from a certificate request. OCSP_response_create()
6540 creates a response and optionally adds a basic response structure.
6541 OCSP_basic_add1_status() adds a complete single response to a basic
6542 response and returns the OCSP_SINGLERESP structure just added (to allow
6543 extensions to be included for example). OCSP_basic_add1_cert() adds a
6544 certificate to a basic response and OCSP_basic_sign() signs a basic
6545 response with various flags. New helper functions ASN1_TIME_check()
6546 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6547 (converts ASN1_TIME to GeneralizedTime).
6550 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6551 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6552 structure from a certificate. X509_pubkey_digest() digests the public_key
6553 contents: this is used in various key identifiers.
6556 *) Make sk_sort() tolerate a NULL argument.
6557 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6559 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6560 passed by the function are trusted implicitly. If any of them signed the
6561 response then it is assumed to be valid and is not verified.
6564 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6565 to data. This was previously part of the PKCS7 ASN1 code. This
6566 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6567 [Steve Henson, reported by Kenneth R. Robinette
6568 <support@securenetterm.com>]
6570 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6571 routines: without these tracing memory leaks is very painful.
6572 Fix leaks in PKCS12 and PKCS7 routines.
6575 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6576 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6577 effectively meant GeneralizedTime would never be used. Now it
6578 is initialised to -1 but X509_time_adj() now has to check the value
6579 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6580 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6581 [Steve Henson, reported by Kenneth R. Robinette
6582 <support@securenetterm.com>]
6584 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6585 result in a zero length in the ASN1_INTEGER structure which was
6586 not consistent with the structure when d2i_ASN1_INTEGER() was used
6587 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6588 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6589 where it did not print out a minus for negative ASN1_INTEGER.
6592 *) Add summary printout to ocsp utility. The various functions which
6593 convert status values to strings have been renamed to:
6594 OCSP_response_status_str(), OCSP_cert_status_str() and
6595 OCSP_crl_reason_str() and are no longer static. New options
6596 to verify nonce values and to disable verification. OCSP response
6597 printout format cleaned up.
6600 *) Add additional OCSP certificate checks. These are those specified
6601 in RFC2560. This consists of two separate checks: the CA of the
6602 certificate being checked must either be the OCSP signer certificate
6603 or the issuer of the OCSP signer certificate. In the latter case the
6604 OCSP signer certificate must contain the OCSP signing extended key
6605 usage. This check is performed by attempting to match the OCSP
6606 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6607 in the OCSP_CERTID structures of the response.
6610 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6611 and related routines. This uses the standard OpenSSL certificate
6612 verify routines to perform initial checks (just CA validity) and
6613 to obtain the certificate chain. Then additional checks will be
6614 performed on the chain. Currently the root CA is checked to see
6615 if it is explicitly trusted for OCSP signing. This is used to set
6616 a root CA as a global signing root: that is any certificate that
6617 chains to that CA is an acceptable OCSP signing certificate.
6620 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6621 extensions from a separate configuration file.
6622 As when reading extensions from the main configuration file,
6623 the '-extensions ...' option may be used for specifying the
6625 [Massimiliano Pala <madwolf@comune.modena.it>]
6627 *) New OCSP utility. Allows OCSP requests to be generated or
6628 read. The request can be sent to a responder and the output
6629 parsed, outputed or printed in text form. Not complete yet:
6630 still needs to check the OCSP response validity.
6633 *) New subcommands for 'openssl ca':
6634 'openssl ca -status <serial>' prints the status of the cert with
6635 the given serial number (according to the index file).
6636 'openssl ca -updatedb' updates the expiry status of certificates
6638 [Massimiliano Pala <madwolf@comune.modena.it>]
6640 *) New '-newreq-nodes' command option to CA.pl. This is like
6641 '-newreq', but calls 'openssl req' with the '-nodes' option
6642 so that the resulting key is not encrypted.
6643 [Damien Miller <djm@mindrot.org>]
6645 *) New configuration for the GNU Hurd.
6646 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6648 *) Initial code to implement OCSP basic response verify. This
6649 is currently incomplete. Currently just finds the signer's
6650 certificate and verifies the signature on the response.
6653 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6654 value of OPENSSLDIR. This is available via the new '-d' option
6655 to 'openssl version', and is also included in 'openssl version -a'.
6658 *) Allowing defining memory allocation callbacks that will be given
6659 file name and line number information in additional arguments
6660 (a const char* and an int). The basic functionality remains, as
6661 well as the original possibility to just replace malloc(),
6662 realloc() and free() by functions that do not know about these
6663 additional arguments. To register and find out the current
6664 settings for extended allocation functions, the following
6665 functions are provided:
6667 CRYPTO_set_mem_ex_functions
6668 CRYPTO_set_locked_mem_ex_functions
6669 CRYPTO_get_mem_ex_functions
6670 CRYPTO_get_locked_mem_ex_functions
6672 These work the same way as CRYPTO_set_mem_functions and friends.
6673 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6674 extended allocation function is enabled.
6675 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6676 a conventional allocation function is enabled.
6677 [Richard Levitte, Bodo Moeller]
6679 *) Finish off removing the remaining LHASH function pointer casts.
6680 There should no longer be any prototype-casting required when using
6681 the LHASH abstraction, and any casts that remain are "bugs". See
6682 the callback types and macros at the head of lhash.h for details
6683 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6686 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6687 If /dev/[u]random devices are not available or do not return enough
6688 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6690 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6691 /etc/entropy will be queried once each in this sequence, quering stops
6692 when enough entropy was collected without querying more sockets.
6695 *) Change the Unix RAND_poll() variant to be able to poll several
6696 random devices, as specified by DEVRANDOM, until a sufficient amount
6697 of data has been collected. We spend at most 10 ms on each file
6698 (select timeout) and read in non-blocking mode. DEVRANDOM now
6699 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6700 (previously it was just the string "/dev/urandom"), so on typical
6701 platforms the 10 ms delay will never occur.
6702 Also separate out the Unix variant to its own file, rand_unix.c.
6703 For VMS, there's a currently-empty rand_vms.c.
6706 *) Move OCSP client related routines to ocsp_cl.c. These
6707 provide utility functions which an application needing
6708 to issue a request to an OCSP responder and analyse the
6709 response will typically need: as opposed to those which an
6710 OCSP responder itself would need which will be added later.
6712 OCSP_request_sign() signs an OCSP request with an API similar
6713 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6714 response. OCSP_response_get1_basic() extracts basic response
6715 from response. OCSP_resp_find_status(): finds and extracts status
6716 information from an OCSP_CERTID structure (which will be created
6717 when the request structure is built). These are built from lower
6718 level functions which work on OCSP_SINGLERESP structures but
6719 wont normally be used unless the application wishes to examine
6720 extensions in the OCSP response for example.
6722 Replace nonce routines with a pair of functions.
6723 OCSP_request_add1_nonce() adds a nonce value and optionally
6724 generates a random value. OCSP_check_nonce() checks the
6725 validity of the nonce in an OCSP response.
6728 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6729 This doesn't copy the supplied OCSP_CERTID and avoids the
6730 need to free up the newly created id. Change return type
6731 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6732 This can then be used to add extensions to the request.
6733 Deleted OCSP_request_new(), since most of its functionality
6734 is now in OCSP_REQUEST_new() (and the case insensitive name
6735 clash) apart from the ability to set the request name which
6736 will be added elsewhere.
6739 *) Update OCSP API. Remove obsolete extensions argument from
6740 various functions. Extensions are now handled using the new
6741 OCSP extension code. New simple OCSP HTTP function which
6742 can be used to send requests and parse the response.
6745 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6746 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6747 uses the special reorder version of SET OF to sort the attributes
6748 and reorder them to match the encoded order. This resolves a long
6749 standing problem: a verify on a PKCS7 structure just after signing
6750 it used to fail because the attribute order did not match the
6751 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6752 it uses the received order. This is necessary to tolerate some broken
6753 software that does not order SET OF. This is handled by encoding
6754 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6755 to produce the required SET OF.
6758 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6759 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6760 files to get correct declarations of the ASN.1 item variables.
6763 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6764 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6765 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6766 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6767 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6768 ASN1_ITEM and no wrapper functions.
6771 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6772 replace the old function pointer based I/O routines. Change most of
6773 the *_d2i_bio() and *_d2i_fp() functions to use these.
6776 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6777 lines, recognice more "algorithms" that can be deselected, and make
6778 it complain about algorithm deselection that isn't recognised.
6781 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6782 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6783 to use new functions. Add NO_ASN1_OLD which can be set to remove
6784 some old style ASN1 functions: this can be used to determine if old
6785 code will still work when these eventually go away.
6788 *) New extension functions for OCSP structures, these follow the
6789 same conventions as certificates and CRLs.
6792 *) New function X509V3_add1_i2d(). This automatically encodes and
6793 adds an extension. Its behaviour can be customised with various
6794 flags to append, replace or delete. Various wrappers added for
6795 certifcates and CRLs.
6798 *) Fix to avoid calling the underlying ASN1 print routine when
6799 an extension cannot be parsed. Correct a typo in the
6800 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6803 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
6804 entries for variables.
6807 *) Add functionality to apps/openssl.c for detecting locking
6808 problems: As the program is single-threaded, all we have
6809 to do is register a locking callback using an array for
6810 storing which locks are currently held by the program.
6813 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6814 SSL_get_ex_data_X509_STORE_idx(), which is used in
6815 ssl_verify_cert_chain() and thus can be called at any time
6816 during TLS/SSL handshakes so that thread-safety is essential.
6817 Unfortunately, the ex_data design is not at all suited
6818 for multi-threaded use, so it probably should be abolished.
6821 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6822 [Broadcom, tweaked and integrated by Geoff Thorpe]
6824 *) Move common extension printing code to new function
6825 X509V3_print_extensions(). Reorganise OCSP print routines and
6826 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6829 *) New function X509_signature_print() to remove duplication in some
6833 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6834 set (this was treated exactly the same as SET OF previously). This
6835 is used to reorder the STACK representing the structure to match the
6836 encoding. This will be used to get round a problem where a PKCS7
6837 structure which was signed could not be verified because the STACK
6838 order did not reflect the encoded order.
6841 *) Reimplement the OCSP ASN1 module using the new code.
6844 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6845 for its ASN1 operations. The old style function pointers still exist
6846 for now but they will eventually go away.
6849 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6850 completely replaces the old ASN1 functionality with a table driven
6851 encoder and decoder which interprets an ASN1_ITEM structure describing
6852 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6853 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6854 has also been converted to the new form.
6857 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6858 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6859 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6860 for negative moduli.
6863 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6864 of not touching the result's sign bit.
6867 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6871 *) Changed the LHASH code to use prototypes for callbacks, and created
6872 macros to declare and implement thin (optionally static) functions
6873 that provide type-safety and avoid function pointer casting for the
6874 type-specific callbacks.
6877 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6879 [Veers Staats <staatsvr@asc.hpc.mil>,
6880 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6882 *) Reformat the FAQ so the different questions and answers can be divided
6883 in sections depending on the subject.
6886 *) Have the zlib compression code load ZLIB.DLL dynamically under
6890 *) New function BN_mod_sqrt for computing square roots modulo a prime
6891 (using the probabilistic Tonelli-Shanks algorithm unless
6892 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6893 be handled deterministically).
6894 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6896 *) Make BN_mod_inverse faster by explicitly handling small quotients
6897 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6898 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6901 *) New function BN_kronecker.
6904 *) Fix BN_gcd so that it works on negative inputs; the result is
6905 positive unless both parameters are zero.
6906 Previously something reasonably close to an infinite loop was
6907 possible because numbers could be growing instead of shrinking
6908 in the implementation of Euclid's algorithm.
6911 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6912 sign of the number in question.
6914 Fix BN_is_word(a,w) to work correctly for w == 0.
6916 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6917 because its test if the absolute value of 'a' equals 'w'.
6918 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6919 it exists mostly for use in the implementations of BN_is_zero(),
6920 BN_is_one(), and BN_is_word().
6923 *) New function BN_swap.
6926 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6927 the exponentiation functions are more likely to produce reasonable
6928 results on negative inputs.
6931 *) Change BN_mod_mul so that the result is always non-negative.
6932 Previously, it could be negative if one of the factors was negative;
6933 I don't think anyone really wanted that behaviour.
6936 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6937 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6938 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6939 and add new functions:
6948 BN_mod_lshift1_quick
6952 These functions always generate non-negative results.
6954 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6955 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6957 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6958 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6959 be reduced modulo m.
6960 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6963 The following entry accidentily appeared in the CHANGES file
6964 distributed with OpenSSL 0.9.7. The modifications described in
6965 it do *not* apply to OpenSSL 0.9.7.
6967 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6968 was actually never needed) and in BN_mul(). The removal in BN_mul()
6969 required a small change in bn_mul_part_recursive() and the addition
6970 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6971 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6972 bn_sub_words() and bn_add_words() except they take arrays with
6977 *) In 'openssl passwd', verify passwords read from the terminal
6978 unless the '-salt' option is used (which usually means that
6979 verification would just waste user's time since the resulting
6980 hash is going to be compared with some given password hash)
6981 or the new '-noverify' option is used.
6983 This is an incompatible change, but it does not affect
6984 non-interactive use of 'openssl passwd' (passwords on the command
6985 line, '-stdin' option, '-in ...' option) and thus should not
6989 *) Remove all references to RSAref, since there's no more need for it.
6992 *) Make DSO load along a path given through an environment variable
6993 (SHLIB_PATH) with shl_load().
6996 *) Constify the ENGINE code as a result of BIGNUM constification.
6997 Also constify the RSA code and most things related to it. In a
6998 few places, most notable in the depth of the ASN.1 code, ugly
6999 casts back to non-const were required (to be solved at a later
7003 *) Make it so the openssl application has all engines loaded by default.
7006 *) Constify the BIGNUM routines a little more.
7009 *) Add the following functions:
7011 ENGINE_load_cswift()
7013 ENGINE_load_atalla()
7015 ENGINE_load_builtin_engines()
7017 That way, an application can itself choose if external engines that
7018 are built-in in OpenSSL shall ever be used or not. The benefit is
7019 that applications won't have to be linked with libdl or other dso
7020 libraries unless it's really needed.
7022 Changed 'openssl engine' to load all engines on demand.
7023 Changed the engine header files to avoid the duplication of some
7024 declarations (they differed!).
7027 *) 'openssl engine' can now list capabilities.
7030 *) Better error reporting in 'openssl engine'.
7033 *) Never call load_dh_param(NULL) in s_server.
7036 *) Add engine application. It can currently list engines by name and
7037 identity, and test if they are actually available.
7040 *) Improve RPM specification file by forcing symbolic linking and making
7041 sure the installed documentation is also owned by root.root.
7042 [Damien Miller <djm@mindrot.org>]
7044 *) Give the OpenSSL applications more possibilities to make use of
7045 keys (public as well as private) handled by engines.
7048 *) Add OCSP code that comes from CertCo.
7051 *) Add VMS support for the Rijndael code.
7054 *) Added untested support for Nuron crypto accelerator.
7057 *) Add support for external cryptographic devices. This code was
7058 previously distributed separately as the "engine" branch.
7059 [Geoff Thorpe, Richard Levitte]
7061 *) Rework the filename-translation in the DSO code. It is now possible to
7062 have far greater control over how a "name" is turned into a filename
7063 depending on the operating environment and any oddities about the
7064 different shared library filenames on each system.
7067 *) Support threads on FreeBSD-elf in Configure.
7070 *) Fix for SHA1 assembly problem with MASM: it produces
7071 warnings about corrupt line number information when assembling
7072 with debugging information. This is caused by the overlapping
7074 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7077 NCONF_get_number() has no error checking at all. As a replacement,
7078 NCONF_get_number_e() is defined (_e for "error checking") and is
7079 promoted strongly. The old NCONF_get_number is kept around for
7080 binary backward compatibility.
7081 Make it possible for methods to load from something other than a BIO,
7082 by providing a function pointer that is given a name instead of a BIO.
7083 For example, this could be used to load configuration data from an
7087 *) Fix for non blocking accept BIOs. Added new I/O special reason
7088 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7089 with non blocking I/O was not possible because no retry code was
7090 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7094 *) Added the beginnings of Rijndael support.
7097 *) Fix for bug in DirectoryString mask setting. Add support for
7098 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7099 to allow certificate printing to more controllable, additional
7100 'certopt' option to 'x509' to allow new printing options to be
7104 *) Clean old EAY MD5 hack from e_os.h.
7107 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7109 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7110 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7111 [Joe Orton, Steve Henson]
7113 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7115 *) Fix additional bug revealed by the NISCC test suite:
7117 Stop bug triggering large recursion when presented with
7118 certain ASN.1 tags (CVE-2003-0851)
7121 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7123 *) Fix various bugs revealed by running the NISCC test suite:
7125 Stop out of bounds reads in the ASN1 code when presented with
7126 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7128 If verify callback ignores invalid public key errors don't try to check
7129 certificate signature with the NULL public key.
7133 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7134 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7138 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7139 extra data after the compression methods not only for TLS 1.0
7140 but also for SSL 3.0 (as required by the specification).
7141 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7143 *) Change X509_certificate_type() to mark the key as exported/exportable
7144 when it's 512 *bits* long, not 512 bytes.
7147 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7149 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7150 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7151 a protocol version number mismatch like a decryption error
7152 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7155 *) Turn on RSA blinding by default in the default implementation
7156 to avoid a timing attack. Applications that don't want it can call
7157 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7158 They would be ill-advised to do so in most cases.
7159 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7161 *) Change RSA blinding code so that it works when the PRNG is not
7162 seeded (in this case, the secret RSA exponent is abused as
7163 an unpredictable seed -- if it is not unpredictable, there
7164 is no point in blinding anyway). Make RSA blinding thread-safe
7165 by remembering the creator's thread ID in rsa->blinding and
7166 having all other threads use local one-time blinding factors
7167 (this requires more computation than sharing rsa->blinding, but
7168 avoids excessive locking; and if an RSA object is not shared
7169 between threads, blinding will still be very fast).
7172 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7174 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7175 via timing by performing a MAC computation even if incorrrect
7176 block cipher padding has been found. This is a countermeasure
7177 against active attacks where the attacker has to distinguish
7178 between bad padding and a MAC verification error. (CVE-2003-0078)
7180 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7181 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7182 Martin Vuagnoux (EPFL, Ilion)]
7184 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7186 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7187 memory from it's contents. This is done with a counter that will
7188 place alternating values in each byte. This can be used to solve
7189 two issues: 1) the removal of calls to memset() by highly optimizing
7190 compilers, and 2) cleansing with other values than 0, since those can
7191 be read through on certain media, for example a swap space on disk.
7194 *) Bugfix: client side session caching did not work with external caching,
7195 because the session->cipher setting was not restored when reloading
7196 from the external cache. This problem was masked, when
7197 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7198 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7201 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7202 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7203 [Zeev Lieber <zeev-l@yahoo.com>]
7205 *) Undo an undocumented change introduced in 0.9.6e which caused
7206 repeated calls to OpenSSL_add_all_ciphers() and
7207 OpenSSL_add_all_digests() to be ignored, even after calling
7211 *) Change the default configuration reader to deal with last line not
7212 being properly terminated.
7215 *) Change X509_NAME_cmp() so it applies the special rules on handling
7216 DN values that are of type PrintableString, as well as RDNs of type
7217 emailAddress where the value has the type ia5String.
7218 [stefank@valicert.com via Richard Levitte]
7220 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7221 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7222 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7223 the bitwise-OR of the two for use by the majority of applications
7224 wanting this behaviour, and update the docs. The documented
7225 behaviour and actual behaviour were inconsistent and had been
7226 changing anyway, so this is more a bug-fix than a behavioural
7228 [Geoff Thorpe, diagnosed by Nadav Har'El]
7230 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7231 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7234 *) Fix initialization code race conditions in
7235 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7236 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7237 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7238 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7239 ssl2_get_cipher_by_char(),
7240 ssl3_get_cipher_by_char().
7241 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7243 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7244 the cached sessions are flushed, as the remove_cb() might use ex_data
7245 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7246 (see [openssl.org #212]).
7247 [Geoff Thorpe, Lutz Jaenicke]
7249 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7250 length, instead of the encoding length to d2i_ASN1_OBJECT.
7253 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7255 *) [In 0.9.6g-engine release:]
7256 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7257 [Lynn Gazis <lgazis@rainbow.com>]
7259 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7261 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7262 and get fix the header length calculation.
7263 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7264 Alon Kantor <alonk@checkpoint.com> (and others),
7267 *) Use proper error handling instead of 'assertions' in buffer
7268 overflow checks added in 0.9.6e. This prevents DoS (the
7269 assertions could call abort()).
7270 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7272 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7274 *) Add various sanity checks to asn1_get_length() to reject
7275 the ASN1 length bytes if they exceed sizeof(long), will appear
7276 negative or the content length exceeds the length of the
7278 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7280 *) Fix cipher selection routines: ciphers without encryption had no flags
7281 for the cipher strength set and where therefore not handled correctly
7282 by the selection routines (PR #130).
7285 *) Fix EVP_dsa_sha macro.
7289 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7290 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7291 that was added in OpenSSL 0.9.6d.
7293 As the countermeasure turned out to be incompatible with some
7294 broken SSL implementations, the new option is part of SSL_OP_ALL.
7295 SSL_OP_ALL is usually employed when compatibility with weird SSL
7296 implementations is desired (e.g. '-bugs' option to 's_client' and
7297 's_server'), so the new option is automatically set in many
7301 *) Changes in security patch:
7303 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7304 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7305 Air Force Materiel Command, USAF, under agreement number
7308 *) Add various sanity checks to asn1_get_length() to reject
7309 the ASN1 length bytes if they exceed sizeof(long), will appear
7310 negative or the content length exceeds the length of the
7311 supplied buffer. (CVE-2002-0659)
7312 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7314 *) Assertions for various potential buffer overflows, not known to
7316 [Ben Laurie (CHATS)]
7318 *) Various temporary buffers to hold ASCII versions of integers were
7319 too small for 64 bit platforms. (CVE-2002-0655)
7320 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7322 *) Remote buffer overflow in SSL3 protocol - an attacker could
7323 supply an oversized session ID to a client. (CVE-2002-0656)
7324 [Ben Laurie (CHATS)]
7326 *) Remote buffer overflow in SSL2 protocol - an attacker could
7327 supply an oversized client master key. (CVE-2002-0656)
7328 [Ben Laurie (CHATS)]
7330 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7332 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7333 encoded as NULL) with id-dsa-with-sha1.
7334 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7336 *) Check various X509_...() return values in apps/req.c.
7337 [Nils Larsch <nla@trustcenter.de>]
7339 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7340 an end-of-file condition would erronously be flagged, when the CRLF
7341 was just at the end of a processed block. The bug was discovered when
7342 processing data through a buffering memory BIO handing the data to a
7343 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7344 <ptsekov@syntrex.com> and Nedelcho Stanev.
7347 *) Implement a countermeasure against a vulnerability recently found
7348 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7349 before application data chunks to avoid the use of known IVs
7350 with data potentially chosen by the attacker.
7353 *) Fix length checks in ssl3_get_client_hello().
7356 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7357 to prevent ssl3_read_internal() from incorrectly assuming that
7358 ssl3_read_bytes() found application data while handshake
7359 processing was enabled when in fact s->s3->in_read_app_data was
7360 merely automatically cleared during the initial handshake.
7361 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7363 *) Fix object definitions for Private and Enterprise: they were not
7364 recognized in their shortname (=lowercase) representation. Extend
7365 obj_dat.pl to issue an error when using undefined keywords instead
7366 of silently ignoring the problem (Svenning Sorensen
7367 <sss@sss.dnsalias.net>).
7370 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7371 generators, i.e. generators other than 2 and 5. (Previously, the
7372 code did not properly initialise the 'add' and 'rem' values to
7373 BN_generate_prime().)
7375 In the new general case, we do not insist that 'generator' is
7376 actually a primitive root: This requirement is rather pointless;
7377 a generator of the order-q subgroup is just as good, if not
7381 *) Map new X509 verification errors to alerts. Discovered and submitted by
7382 Tom Wu <tom@arcot.com>.
7385 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7386 returning non-zero before the data has been completely received
7387 when using non-blocking I/O.
7388 [Bodo Moeller; problem pointed out by John Hughes]
7390 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7391 [Ben Laurie, Lutz Jaenicke]
7393 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7394 Yoram Zahavi <YoramZ@gilian.com>).
7397 *) Add information about CygWin 1.3 and on, and preserve proper
7398 configuration for the versions before that.
7399 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7401 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7402 check whether we deal with a copy of a session and do not delete from
7403 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7404 <izhar@checkpoint.com>.
7407 *) Do not store session data into the internal session cache, if it
7408 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7409 flag is set). Proposed by Aslam <aslam@funk.com>.
7412 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7416 *) [In 0.9.6d-engine release:]
7417 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7418 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7420 *) Add the configuration target linux-s390x.
7421 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7423 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7424 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7425 variable as an indication that a ClientHello message has been
7426 received. As the flag value will be lost between multiple
7427 invocations of ssl3_accept when using non-blocking I/O, the
7428 function may not be aware that a handshake has actually taken
7429 place, thus preventing a new session from being added to the
7432 To avoid this problem, we now set s->new_session to 2 instead of
7433 using a local variable.
7434 [Lutz Jaenicke, Bodo Moeller]
7436 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7437 if the SSL_R_LENGTH_MISMATCH error is detected.
7438 [Geoff Thorpe, Bodo Moeller]
7440 *) New 'shared_ldflag' column in Configure platform table.
7443 *) Fix EVP_CIPHER_mode macro.
7444 ["Dan S. Camper" <dan@bti.net>]
7446 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7447 type, we must throw them away by setting rr->length to 0.
7448 [D P Chang <dpc@qualys.com>]
7450 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7452 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7453 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7454 worked incorrectly for those cases where range = 10..._2 and
7455 3*range is two bits longer than range.)
7458 *) Only add signing time to PKCS7 structures if it is not already
7462 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7463 OBJ_ld_ce should be OBJ_id_ce.
7464 Also some ip-pda OIDs in crypto/objects/objects.txt were
7465 incorrect (cf. RFC 3039).
7466 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7468 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7469 returns early because it has nothing to do.
7470 [Andy Schneider <andy.schneider@bjss.co.uk>]
7472 *) [In 0.9.6c-engine release:]
7473 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7474 [Andy Schneider <andy.schneider@bjss.co.uk>]
7476 *) [In 0.9.6c-engine release:]
7477 Add support for Cryptographic Appliance's keyserver technology.
7478 (Use engine 'keyclient')
7479 [Cryptographic Appliances and Geoff Thorpe]
7481 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7482 is called via tools/c89.sh because arguments have to be
7483 rearranged (all '-L' options must appear before the first object
7485 [Richard Shapiro <rshapiro@abinitio.com>]
7487 *) [In 0.9.6c-engine release:]
7488 Add support for Broadcom crypto accelerator cards, backported
7490 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7492 *) [In 0.9.6c-engine release:]
7493 Add support for SureWare crypto accelerator cards from
7494 Baltimore Technologies. (Use engine 'sureware')
7495 [Baltimore Technologies and Mark Cox]
7497 *) [In 0.9.6c-engine release:]
7498 Add support for crypto accelerator cards from Accelerated
7499 Encryption Processing, www.aep.ie. (Use engine 'aep')
7500 [AEP Inc. and Mark Cox]
7502 *) Add a configuration entry for gcc on UnixWare.
7503 [Gary Benson <gbenson@redhat.com>]
7505 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7506 messages are stored in a single piece (fixed-length part and
7507 variable-length part combined) and fix various bugs found on the way.
7510 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7511 instead. BIO_gethostbyname() does not know what timeouts are
7512 appropriate, so entries would stay in cache even when they have
7514 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7516 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7517 faced with a pathologically small ClientHello fragment that does
7518 not contain client_version: Instead of aborting with an error,
7519 simply choose the highest available protocol version (i.e.,
7520 TLS 1.0 unless it is disabled). In practice, ClientHello
7521 messages are never sent like this, but this change gives us
7522 strictly correct behaviour at least for TLS.
7525 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7526 never resets s->method to s->ctx->method when called from within
7527 one of the SSL handshake functions.
7528 [Bodo Moeller; problem pointed out by Niko Baric]
7530 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7531 (sent using the client's version number) if client_version is
7532 smaller than the protocol version in use. Also change
7533 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7534 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7535 the client will at least see that alert.
7538 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7542 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7543 client receives HelloRequest while in a handshake.
7544 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7546 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7547 should end in 'break', not 'goto end' which circuments various
7548 cleanups done in state SSL_ST_OK. But session related stuff
7549 must be disabled for SSL_ST_OK in the case that we just sent a
7552 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7553 before just sending a HelloRequest.
7554 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7556 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7557 reveal whether illegal block cipher padding was found or a MAC
7558 verification error occured. (Neither SSLerr() codes nor alerts
7559 are directly visible to potential attackers, but the information
7560 may leak via logfiles.)
7562 Similar changes are not required for the SSL 2.0 implementation
7563 because the number of padding bytes is sent in clear for SSL 2.0,
7564 and the extra bytes are just ignored. However ssl/s2_pkt.c
7565 failed to verify that the purported number of padding bytes is in
7569 *) Add OpenUNIX-8 support including shared libraries
7570 (Boyd Lynn Gerber <gerberb@zenez.com>).
7573 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7574 'wristwatch attack' using huge encoding parameters (cf.
7575 James H. Manger's CRYPTO 2001 paper). Note that the
7576 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7577 encoding parameters and hence was not vulnerable.
7580 *) BN_sqr() bug fix.
7581 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7583 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7584 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7585 followed by modular reduction.
7586 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7588 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7589 equivalent based on BN_pseudo_rand() instead of BN_rand().
7592 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7593 This function was broken, as the check for a new client hello message
7594 to handle SGC did not allow these large messages.
7595 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7598 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7601 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7602 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7605 *) Rework the configuration and shared library support for Tru64 Unix.
7606 The configuration part makes use of modern compiler features and
7607 still retains old compiler behavior for those that run older versions
7608 of the OS. The shared library support part includes a variant that
7609 uses the RPATH feature, and is available through the special
7610 configuration target "alpha-cc-rpath", which will never be selected
7612 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7614 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7615 with the same message size as in ssl3_get_certificate_request().
7616 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7617 messages might inadvertently be reject as too long.
7618 [Petr Lampa <lampa@fee.vutbr.cz>]
7620 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7623 *) Modified SSL library such that the verify_callback that has been set
7624 specificly for an SSL object with SSL_set_verify() is actually being
7625 used. Before the change, a verify_callback set with this function was
7626 ignored and the verify_callback() set in the SSL_CTX at the time of
7627 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7628 to allow the necessary settings.
7631 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7632 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7633 done automatically (in contradiction to the requirements of the C
7634 standard). This made problems when used from OpenSSH.
7637 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7638 dh->length and always used
7640 BN_rand_range(priv_key, dh->p).
7642 BN_rand_range() is not necessary for Diffie-Hellman, and this
7643 specific range makes Diffie-Hellman unnecessarily inefficient if
7644 dh->length (recommended exponent length) is much smaller than the
7645 length of dh->p. We could use BN_rand_range() if the order of
7646 the subgroup was stored in the DH structure, but we only have
7651 BN_rand(priv_key, l, ...)
7653 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7659 RSA_eay_public_encrypt
7660 RSA_eay_private_decrypt
7661 RSA_eay_private_encrypt (signing)
7662 RSA_eay_public_decrypt (signature verification)
7664 (default implementations for RSA_public_encrypt,
7665 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7666 always reject numbers >= n.
7669 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7670 to synchronize access to 'locking_thread'. This is necessary on
7671 systems where access to 'locking_thread' (an 'unsigned long'
7672 variable) is not atomic.
7675 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7676 *before* setting the 'crypto_lock_rand' flag. The previous code had
7677 a race condition if 0 is a valid thread ID.
7678 [Travis Vitek <vitek@roguewave.com>]
7680 *) Add support for shared libraries under Irix.
7681 [Albert Chin-A-Young <china@thewrittenword.com>]
7683 *) Add configuration option to build on Linux on both big-endian and
7685 [Ralf Baechle <ralf@uni-koblenz.de>]
7687 *) Add the possibility to create shared libraries on HP-UX.
7690 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7692 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7693 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7694 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7695 PRNG state recovery was possible based on the output of
7696 one PRNG request appropriately sized to gain knowledge on
7697 'md' followed by enough consecutive 1-byte PRNG requests
7698 to traverse all of 'state'.
7700 1. When updating 'md_local' (the current thread's copy of 'md')
7701 during PRNG output generation, hash all of the previous
7702 'md_local' value, not just the half used for PRNG output.
7704 2. Make the number of bytes from 'state' included into the hash
7705 independent from the number of PRNG bytes requested.
7707 The first measure alone would be sufficient to avoid
7708 Markku-Juhani's attack. (Actually it had never occurred
7709 to me that the half of 'md_local' used for chaining was the
7710 half from which PRNG output bytes were taken -- I had always
7711 assumed that the secret half would be used.) The second
7712 measure makes sure that additional data from 'state' is never
7713 mixed into 'md_local' in small portions; this heuristically
7714 further strengthens the PRNG.
7717 *) Fix crypto/bn/asm/mips3.s.
7720 *) When only the key is given to "enc", the IV is undefined. Print out
7721 an error message in this case.
7724 *) Handle special case when X509_NAME is empty in X509 printing routines.
7727 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7728 positive and less than q.
7731 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7732 used: it isn't thread safe and the add_lock_callback should handle
7734 [Paul Rose <Paul.Rose@bridge.com>]
7736 *) Verify that incoming data obeys the block size in
7737 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7741 [Ulf Möller, Bodo Möller]
7743 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7744 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7745 when fixing the server behaviour for backwards-compatible 'client
7746 hello' messages. (Note that the attack is impractical against
7747 SSL 3.0 and TLS 1.0 anyway because length and version checking
7748 means that the probability of guessing a valid ciphertext is
7749 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7752 Before 0.9.5, the countermeasure (hide the error by generating a
7753 random 'decryption result') did not work properly because
7754 ERR_clear_error() was missing, meaning that SSL_get_error() would
7755 detect the supposedly ignored error.
7757 Both problems are now fixed.
7760 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7761 (previously it was 1024).
7764 *) Fix for compatibility mode trust settings: ignore trust settings
7765 unless some valid trust or reject settings are present.
7768 *) Fix for blowfish EVP: its a variable length cipher.
7771 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7772 parameters in DSA public key structures and return an error in the
7773 DSA routines if parameters are absent.
7776 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7777 in the current directory if neither $RANDFILE nor $HOME was set.
7778 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7779 caused some confusion to Windows users who haven't defined $HOME.
7780 Thus RAND_file_name() is changed again: e_os.h can define a
7781 DEFAULT_HOME, which will be used if $HOME is not set.
7782 For Windows, we use "C:"; on other platforms, we still require
7783 environment variables.
7785 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7786 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7787 having multiple threads call RAND_poll() concurrently.
7790 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7791 combination of a flag and a thread ID variable.
7792 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7793 flag), *other* threads can enter ssleay_add_bytes without obeying
7794 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7795 that they do not hold after the first thread unsets add_do_not_lock).
7798 *) Change bctest again: '-x' expressions are not available in all
7802 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7804 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7805 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7807 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7808 the default extension for executables, if any. Also, make the perl
7809 scripts that use symlink() to test if it really exists and use "cp"
7810 if it doesn't. All this made OpenSSL compilable and installable in
7814 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7815 If SEQUENCE is length is indefinite just set c->slen to the total
7816 amount of data available.
7817 [Steve Henson, reported by shige@FreeBSD.org]
7818 [This change does not apply to 0.9.7.]
7820 *) Change bctest to avoid here-documents inside command substitution
7821 (workaround for FreeBSD /bin/sh bug).
7822 For compatibility with Ultrix, avoid shell functions (introduced
7823 in the bctest version that searches along $PATH).
7826 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7827 with des_encrypt() defined on some operating systems, like Solaris
7831 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7832 On the Importance of Eliminating Errors in Cryptographic
7833 Computations, J. Cryptology 14 (2001) 2, 101-119,
7834 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7837 *) MIPS assembler BIGNUM division bug fix.
7840 *) Disabled incorrect Alpha assembler code.
7843 *) Fix PKCS#7 decode routines so they correctly update the length
7844 after reading an EOC for the EXPLICIT tag.
7846 [This change does not apply to 0.9.7.]
7848 *) Fix bug in PKCS#12 key generation routines. This was triggered
7849 if a 3DES key was generated with a 0 initial byte. Include
7850 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7851 (but broken) behaviour.
7854 *) Enhance bctest to search for a working bc along $PATH and print
7856 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7858 *) Fix memory leaks in err.c: free err_data string if necessary;
7859 don't write to the wrong index in ERR_set_error_data.
7862 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7866 *) Replace rdtsc with _emit statements for VC++ version 5.
7867 [Jeremy Cooper <jeremy@baymoo.org>]
7869 *) Make it possible to reuse SSLv2 sessions.
7872 *) In copy_email() check for >= 0 as a return value for
7873 X509_NAME_get_index_by_NID() since 0 is a valid index.
7874 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7876 *) Avoid coredump with unsupported or invalid public keys by checking if
7877 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7878 PKCS7_verify() fails with non detached data.
7881 *) Don't use getenv in library functions when run as setuid/setgid.
7882 New function OPENSSL_issetugid().
7885 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7886 due to incorrect handling of multi-threading:
7888 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7890 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7892 3. Count how many times MemCheck_off() has been called so that
7893 nested use can be treated correctly. This also avoids
7894 inband-signalling in the previous code (which relied on the
7895 assumption that thread ID 0 is impossible).
7898 *) Add "-rand" option also to s_client and s_server.
7901 *) Fix CPU detection on Irix 6.x.
7902 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7903 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7905 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7908 [This change does not apply to 0.9.7.]
7910 *) Use the cached encoding of an X509_NAME structure rather than
7911 copying it. This is apparently the reason for the libsafe "errors"
7912 but the code is actually correct.
7915 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7916 Bleichenbacher's DSA attack.
7917 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7918 to be set and top=0 forces the highest bit to be set; top=-1 is new
7919 and leaves the highest bit random.
7920 [Ulf Moeller, Bodo Moeller]
7922 *) In the NCONF_...-based implementations for CONF_... queries
7923 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7924 a temporary CONF structure with the data component set to NULL
7925 (which gives segmentation faults in lh_retrieve).
7926 Instead, use NULL for the CONF pointer in CONF_get_string and
7927 CONF_get_number (which may use environment variables) and directly
7928 return NULL from CONF_get_section.
7931 *) Fix potential buffer overrun for EBCDIC.
7934 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7935 keyUsage if basicConstraints absent for a CA.
7938 *) Make SMIME_write_PKCS7() write mail header values with a format that
7939 is more generally accepted (no spaces before the semicolon), since
7940 some programs can't parse those values properly otherwise. Also make
7941 sure BIO's that break lines after each write do not create invalid
7945 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7946 macros previously used would not encode an empty SEQUENCE OF
7947 and break the signature.
7949 [This change does not apply to 0.9.7.]
7951 *) Zero the premaster secret after deriving the master secret in
7955 *) Add some EVP_add_digest_alias registrations (as found in
7956 OpenSSL_add_all_digests()) to SSL_library_init()
7957 aka OpenSSL_add_ssl_algorithms(). This provides improved
7958 compatibility with peers using X.509 certificates
7959 with unconventional AlgorithmIdentifier OIDs.
7962 *) Fix for Irix with NO_ASM.
7963 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7965 *) ./config script fixes.
7966 [Ulf Moeller, Richard Levitte]
7968 *) Fix 'openssl passwd -1'.
7971 *) Change PKCS12_key_gen_asc() so it can cope with non null
7972 terminated strings whose length is passed in the passlen
7973 parameter, for example from PEM callbacks. This was done
7974 by adding an extra length parameter to asc2uni().
7975 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7977 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7978 call failed, free the DSA structure.
7981 *) Fix to uni2asc() to cope with zero length Unicode strings.
7982 These are present in some PKCS#12 files.
7985 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7986 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7987 when writing a 32767 byte record.
7988 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7990 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7991 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7993 (RSA objects have a reference count access to which is protected
7994 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7995 so they are meant to be shared between threads.)
7996 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7997 "Reddie, Steven" <Steven.Reddie@ca.com>]
7999 *) Fix a deadlock in CRYPTO_mem_leaks().
8002 *) Use better test patterns in bntest.
8005 *) rand_win.c fix for Borland C.
8008 *) BN_rshift bugfix for n == 0.
8011 *) Add a 'bctest' script that checks for some known 'bc' bugs
8012 so that 'make test' does not abort just because 'bc' is broken.
8015 *) Store verify_result within SSL_SESSION also for client side to
8016 avoid potential security hole. (Re-used sessions on the client side
8017 always resulted in verify_result==X509_V_OK, not using the original
8018 result of the server certificate verification.)
8021 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
8022 SSL3_RT_APPLICATION_DATA, return 0.
8023 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
8027 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
8028 releases, have been re-implemented by renaming the previous
8029 implementations of ssl2_read and ssl3_read to ssl2_read_internal
8030 and ssl3_read_internal, respectively, and adding 'peek' parameters
8031 to them. The new ssl[23]_{read,peek} functions are calls to
8032 ssl[23]_read_internal with the 'peek' flag set appropriately.
8033 A 'peek' parameter has also been added to ssl3_read_bytes, which
8034 does the actual work for ssl3_read_internal.
8037 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
8038 the method-specific "init()" handler. Also clean up ex_data after
8039 calling the method-specific "finish()" handler. Previously, this was
8040 happening the other way round.
8043 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
8044 The previous value, 12, was not always sufficient for BN_mod_exp().
8047 *) Make sure that shared libraries get the internal name engine with
8048 the full version number and not just 0. This should mark the
8049 shared libraries as not backward compatible. Of course, this should
8050 be changed again when we can guarantee backward binary compatibility.
8053 *) Fix typo in get_cert_by_subject() in by_dir.c
8054 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8056 *) Rework the system to generate shared libraries:
8058 - Make note of the expected extension for the shared libraries and
8059 if there is a need for symbolic links from for example libcrypto.so.0
8060 to libcrypto.so.0.9.7. There is extended info in Configure for
8063 - Make as few rebuilds of the shared libraries as possible.
8065 - Still avoid linking the OpenSSL programs with the shared libraries.
8067 - When installing, install the shared libraries separately from the
8071 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8073 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8074 and not in SSL_clear because the latter is also used by the
8075 accept/connect functions; previously, the settings made by
8076 SSL_set_read_ahead would be lost during the handshake.
8077 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8079 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8080 Previously, it would create entries for disableed algorithms no
8084 *) Added several new manual pages for SSL_* function.
8087 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8089 *) In ssl23_get_client_hello, generate an error message when faced
8090 with an initial SSL 3.0/TLS record that is too small to contain the
8091 first two bytes of the ClientHello message, i.e. client_version.
8092 (Note that this is a pathologic case that probably has never happened
8093 in real life.) The previous approach was to use the version number
8094 from the record header as a substitute; but our protocol choice
8095 should not depend on that one because it is not authenticated
8096 by the Finished messages.
8099 *) More robust randomness gathering functions for Windows.
8100 [Jeffrey Altman <jaltman@columbia.edu>]
8102 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8103 not set then we don't setup the error code for issuer check errors
8104 to avoid possibly overwriting other errors which the callback does
8105 handle. If an application does set the flag then we assume it knows
8106 what it is doing and can handle the new informational codes
8110 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8111 a general "ANY" type, as such it should be able to decode anything
8112 including tagged types. However it didn't check the class so it would
8113 wrongly interpret tagged types in the same way as their universal
8114 counterpart and unknown types were just rejected. Changed so that the
8115 tagged and unknown types are handled in the same way as a SEQUENCE:
8116 that is the encoding is stored intact. There is also a new type
8117 "V_ASN1_OTHER" which is used when the class is not universal, in this
8118 case we have no idea what the actual type is so we just lump them all
8122 *) On VMS, stdout may very well lead to a file that is written to
8123 in a record-oriented fashion. That means that every write() will
8124 write a separate record, which will be read separately by the
8125 programs trying to read from it. This can be very confusing.
8127 The solution is to put a BIO filter in the way that will buffer
8128 text until a linefeed is reached, and then write everything a
8129 line at a time, so every record written will be an actual line,
8130 not chunks of lines and not (usually doesn't happen, but I've
8131 seen it once) several lines in one record. BIO_f_linebuffer() is
8134 Currently, it's a VMS-only method, because that's where it has
8135 been tested well enough.
8138 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8139 it can return incorrect results.
8140 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8141 but it was in 0.9.6-beta[12].)
8144 *) Disable the check for content being present when verifying detached
8145 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8146 include zero length content when signing messages.
8149 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8150 BIO_ctrl (for BIO pairs).
8153 *) Add DSO method for VMS.
8156 *) Bug fix: Montgomery multiplication could produce results with the
8160 *) Add RPM specification openssl.spec and modify it to build three
8161 packages. The default package contains applications, application
8162 documentation and run-time libraries. The devel package contains
8163 include files, static libraries and function documentation. The
8164 doc package contains the contents of the doc directory. The original
8165 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8168 *) Add a large number of documentation files for many SSL routines.
8169 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8171 *) Add a configuration entry for Sony News 4.
8172 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8174 *) Don't set the two most significant bits to one when generating a
8175 random number < q in the DSA library.
8178 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8179 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8180 the underlying transport is blocking) if a handshake took place.
8181 (The default behaviour is needed by applications such as s_client
8182 and s_server that use select() to determine when to use SSL_read;
8183 but for applications that know in advance when to expect data, it
8184 just makes things more complicated.)
8187 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8191 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8192 work better on such systems.
8193 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8195 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8196 Update PKCS12_parse() so it copies the friendlyName and the
8197 keyid to the certificates aux info.
8200 *) Fix bug in PKCS7_verify() which caused an infinite loop
8201 if there was more than one signature.
8202 [Sven Uszpelkat <su@celocom.de>]
8204 *) Major change in util/mkdef.pl to include extra information
8205 about each symbol, as well as presentig variables as well
8206 as functions. This change means that there's n more need
8207 to rebuild the .num files when some algorithms are excluded.
8210 *) Allow the verify time to be set by an application,
8211 rather than always using the current time.
8214 *) Phase 2 verify code reorganisation. The certificate
8215 verify code now looks up an issuer certificate by a
8216 number of criteria: subject name, authority key id
8217 and key usage. It also verifies self signed certificates
8218 by the same criteria. The main comparison function is
8219 X509_check_issued() which performs these checks.
8221 Lot of changes were necessary in order to support this
8222 without completely rewriting the lookup code.
8224 Authority and subject key identifier are now cached.
8226 The LHASH 'certs' is X509_STORE has now been replaced
8227 by a STACK_OF(X509_OBJECT). This is mainly because an
8228 LHASH can't store or retrieve multiple objects with
8229 the same hash value.
8231 As a result various functions (which were all internal
8232 use only) have changed to handle the new X509_STORE
8233 structure. This will break anything that messed round
8234 with X509_STORE internally.
8236 The functions X509_STORE_add_cert() now checks for an
8237 exact match, rather than just subject name.
8239 The X509_STORE API doesn't directly support the retrieval
8240 of multiple certificates matching a given criteria, however
8241 this can be worked round by performing a lookup first
8242 (which will fill the cache with candidate certificates)
8243 and then examining the cache for matches. This is probably
8244 the best we can do without throwing out X509_LOOKUP
8245 entirely (maybe later...).
8247 The X509_VERIFY_CTX structure has been enhanced considerably.
8249 All certificate lookup operations now go via a get_issuer()
8250 callback. Although this currently uses an X509_STORE it
8251 can be replaced by custom lookups. This is a simple way
8252 to bypass the X509_STORE hackery necessary to make this
8253 work and makes it possible to use more efficient techniques
8254 in future. A very simple version which uses a simple
8255 STACK for its trusted certificate store is also provided
8256 using X509_STORE_CTX_trusted_stack().
8258 The verify_cb() and verify() callbacks now have equivalents
8259 in the X509_STORE_CTX structure.
8261 X509_STORE_CTX also has a 'flags' field which can be used
8262 to customise the verify behaviour.
8265 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8266 excludes S/MIME capabilities.
8269 *) When a certificate request is read in keep a copy of the
8270 original encoding of the signed data and use it when outputing
8271 again. Signatures then use the original encoding rather than
8272 a decoded, encoded version which may cause problems if the
8273 request is improperly encoded.
8276 *) For consistency with other BIO_puts implementations, call
8277 buffer_write(b, ...) directly in buffer_puts instead of calling
8280 In BIO_puts, increment b->num_write as in BIO_write.
8281 [Peter.Sylvester@EdelWeb.fr]
8283 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8284 BN_zero, we may not return a BIGNUM with an array consisting of
8288 *) Avoid calling abort() from within the library when problems are
8289 detected, except if preprocessor symbols have been defined
8290 (such as REF_CHECK, BN_DEBUG etc.).
8293 *) New openssl application 'rsautl'. This utility can be
8294 used for low level RSA operations. DER public key
8295 BIO/fp routines also added.
8298 *) New Configure entry and patches for compiling on QNX 4.
8299 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8301 *) A demo state-machine implementation was sponsored by
8302 Nuron (http://www.nuron.com/) and is now available in
8303 demos/state_machine.
8306 *) New options added to the 'dgst' utility for signature
8307 generation and verification.
8310 *) Unrecognized PKCS#7 content types are now handled via a
8311 catch all ASN1_TYPE structure. This allows unsupported
8312 types to be stored as a "blob" and an application can
8313 encode and decode it manually.
8316 *) Fix various signed/unsigned issues to make a_strex.c
8318 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8320 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8321 length if passed a buffer. ASN1_INTEGER_to_BN failed
8322 if passed a NULL BN and its argument was negative.
8323 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8325 *) Modification to PKCS#7 encoding routines to output definite
8326 length encoding. Since currently the whole structures are in
8327 memory there's not real point in using indefinite length
8328 constructed encoding. However if OpenSSL is compiled with
8329 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8332 *) Added BIO_vprintf() and BIO_vsnprintf().
8335 *) Added more prefixes to parse for in the the strings written
8336 through a logging bio, to cover all the levels that are available
8337 through syslog. The prefixes are now:
8339 PANIC, EMERG, EMR => LOG_EMERG
8340 ALERT, ALR => LOG_ALERT
8341 CRIT, CRI => LOG_CRIT
8342 ERROR, ERR => LOG_ERR
8343 WARNING, WARN, WAR => LOG_WARNING
8344 NOTICE, NOTE, NOT => LOG_NOTICE
8345 INFO, INF => LOG_INFO
8346 DEBUG, DBG => LOG_DEBUG
8348 and as before, if none of those prefixes are present at the
8349 beginning of the string, LOG_ERR is chosen.
8351 On Win32, the LOG_* levels are mapped according to this:
8353 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8354 LOG_WARNING => EVENTLOG_WARNING_TYPE
8355 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8359 *) Made it possible to reconfigure with just the configuration
8360 argument "reconf" or "reconfigure". The command line arguments
8361 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8362 and are retrieved from there when reconfiguring.
8366 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8368 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8371 *) The obj_dat.pl script was messing up the sorting of object
8372 names. The reason was that it compared the quoted version
8373 of strings as a result "OCSP" > "OCSP Signing" because
8374 " > SPACE. Changed script to store unquoted versions of
8375 names and add quotes on output. It was also omitting some
8376 names from the lookup table if they were given a default
8377 value (that is if SN is missing it is given the same
8378 value as LN and vice versa), these are now added on the
8379 grounds that if an object has a name we should be able to
8380 look it up. Finally added warning output when duplicate
8381 short or long names are found.
8384 *) Changes needed for Tandem NSK.
8385 [Scott Uroff <scott@xypro.com>]
8387 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8388 RSA_padding_check_SSLv23(), special padding was never detected
8389 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8390 version rollback attacks was not effective.
8392 In s23_clnt.c, don't use special rollback-attack detection padding
8393 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8394 client; similarly, in s23_srvr.c, don't do the rollback check if
8395 SSL 2.0 is the only protocol enabled in the server.
8398 *) Make it possible to get hexdumps of unprintable data with 'openssl
8399 asn1parse'. By implication, the functions ASN1_parse_dump() and
8400 BIO_dump_indent() are added.
8403 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8404 these print out strings and name structures based on various
8405 flags including RFC2253 support and proper handling of
8406 multibyte characters. Added options to the 'x509' utility
8407 to allow the various flags to be set.
8410 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8411 Also change the functions X509_cmp_current_time() and
8412 X509_gmtime_adj() work with an ASN1_TIME structure,
8413 this will enable certificates using GeneralizedTime in validity
8414 dates to be checked.
8417 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8418 negative public key encodings) on by default,
8419 NO_NEG_PUBKEY_BUG can be set to disable it.
8422 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8423 content octets. An i2c_ASN1_OBJECT is unnecessary because
8424 the encoding can be trivially obtained from the structure.
8427 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8428 not read locks (CRYPTO_r_[un]lock).
8431 *) A first attempt at creating official support for shared
8432 libraries through configuration. I've kept it so the
8433 default is static libraries only, and the OpenSSL programs
8434 are always statically linked for now, but there are
8435 preparations for dynamic linking in place.
8436 This has been tested on Linux and Tru64.
8439 *) Randomness polling function for Win9x, as described in:
8440 Peter Gutmann, Software Generation of Practically Strong
8444 *) Fix so PRNG is seeded in req if using an already existing
8448 *) New options to smime application. -inform and -outform
8449 allow alternative formats for the S/MIME message including
8450 PEM and DER. The -content option allows the content to be
8451 specified separately. This should allow things like Netscape
8452 form signing output easier to verify.
8455 *) Fix the ASN1 encoding of tags using the 'long form'.
8458 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8459 STRING types. These convert content octets to and from the
8460 underlying type. The actual tag and length octets are
8461 already assumed to have been read in and checked. These
8462 are needed because all other string types have virtually
8463 identical handling apart from the tag. By having versions
8464 of the ASN1 functions that just operate on content octets
8465 IMPLICIT tagging can be handled properly. It also allows
8466 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8467 and ASN1_INTEGER are identical apart from the tag.
8470 *) Change the handling of OID objects as follows:
8472 - New object identifiers are inserted in objects.txt, following
8473 the syntax given in objects.README.
8474 - objects.pl is used to process obj_mac.num and create a new
8476 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8479 This is currently kind of a hack, and the perl code in objects.pl
8480 isn't very elegant, but it works as I intended. The simplest way
8481 to check that it worked correctly is to look in obj_dat.h and
8482 check the array nid_objs and make sure the objects haven't moved
8483 around (this is important!). Additions are OK, as well as
8484 consistent name changes.
8487 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8490 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8491 The given file adds to whatever has already been seeded into the
8492 random pool through the RANDFILE configuration file option or
8493 environment variable, or the default random state file.
8496 *) mkstack.pl now sorts each macro group into lexical order.
8497 Previously the output order depended on the order the files
8498 appeared in the directory, resulting in needless rewriting
8502 *) Patches to make OpenSSL compile under Win32 again. Mostly
8503 work arounds for the VC++ problem that it treats func() as
8504 func(void). Also stripped out the parts of mkdef.pl that
8505 added extra typesafe functions: these no longer exist.
8508 *) Reorganisation of the stack code. The macros are now all
8509 collected in safestack.h . Each macro is defined in terms of
8510 a "stack macro" of the form SKM_<name>(type, a, b). The
8511 DEBUG_SAFESTACK is now handled in terms of function casts,
8512 this has the advantage of retaining type safety without the
8513 use of additional functions. If DEBUG_SAFESTACK is not defined
8514 then the non typesafe macros are used instead. Also modified the
8515 mkstack.pl script to handle the new form. Needs testing to see
8516 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8517 the default if no major problems. Similar behaviour for ASN1_SET_OF
8518 and PKCS12_STACK_OF.
8521 *) When some versions of IIS use the 'NET' form of private key the
8522 key derivation algorithm is different. Normally MD5(password) is
8523 used as a 128 bit RC4 key. In the modified case
8524 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
8525 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8526 as the old Netscape_RSA functions except they have an additional
8527 'sgckey' parameter which uses the modified algorithm. Also added
8528 an -sgckey command line option to the rsa utility. Thanks to
8529 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8530 algorithm to openssl-dev.
8533 *) The evp_local.h macros were using 'c.##kname' which resulted in
8534 invalid expansion on some systems (SCO 5.0.5 for example).
8535 Corrected to 'c.kname'.
8536 [Phillip Porch <root@theporch.com>]
8538 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8539 a STACK of email addresses from a certificate or request, these look
8540 in the subject name and the subject alternative name extensions and
8541 omit any duplicate addresses.
8544 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8545 This makes DSA verification about 2 % faster.
8548 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8549 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8550 plus overhead for 1024 bit moduli).
8551 This makes exponentiations about 0.5 % faster for 1024 bit
8552 exponents (as measured by "openssl speed rsa2048").
8555 *) Rename memory handling macros to avoid conflicts with other
8557 Malloc => OPENSSL_malloc
8558 Malloc_locked => OPENSSL_malloc_locked
8559 Realloc => OPENSSL_realloc
8560 Free => OPENSSL_free
8563 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8564 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8567 *) CygWin32 support.
8568 [John Jarvie <jjarvie@newsguy.com>]
8570 *) The type-safe stack code has been rejigged. It is now only compiled
8571 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8572 by default all type-specific stack functions are "#define"d back to
8573 standard stack functions. This results in more streamlined output
8574 but retains the type-safety checking possibilities of the original
8578 *) The STACK code has been cleaned up, and certain type declarations
8579 that didn't make a lot of sense have been brought in line. This has
8580 also involved a cleanup of sorts in safestack.h to more correctly
8581 map type-safe stack functions onto their plain stack counterparts.
8582 This work has also resulted in a variety of "const"ifications of
8583 lots of the code, especially "_cmp" operations which should normally
8584 be prototyped with "const" parameters anyway.
8587 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8588 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8589 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8590 where all of 'md' is used each time the PRNG is used, but 'state'
8591 is used only indexed by a cyclic counter. As entropy may not be
8592 well distributed from the beginning, 'md' is important as a
8593 chaining variable. However, the output function chains only half
8594 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8595 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8596 in all of 'state' being rewritten, with the new values depending
8597 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8600 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8601 the handshake is continued after ssl_verify_cert_chain();
8602 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8603 can lead to 'unexplainable' connection aborts later.
8604 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8606 *) Major EVP API cipher revision.
8607 Add hooks for extra EVP features. This allows various cipher
8608 parameters to be set in the EVP interface. Support added for variable
8609 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8610 setting of RC2 and RC5 parameters.
8612 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8615 Remove lots of duplicated code from the EVP library. For example *every*
8616 cipher init() function handles the 'iv' in the same way according to the
8617 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8618 for CFB and OFB modes they zero ctx->num.
8620 New functionality allows removal of S/MIME code RC2 hack.
8622 Most of the routines have the same form and so can be declared in terms
8625 By shifting this to the top level EVP_CipherInit() it can be removed from
8626 all individual ciphers. If the cipher wants to handle IVs or keys
8627 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8630 Change lots of functions like EVP_EncryptUpdate() to now return a
8631 value: although software versions of the algorithms cannot fail
8632 any installed hardware versions can.
8635 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8636 this option is set, tolerate broken clients that send the negotiated
8637 protocol version number instead of the requested protocol version
8641 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8642 i.e. non-zero for export ciphersuites, zero otherwise.
8643 Previous versions had this flag inverted, inconsistent with
8644 rsa_tmp_cb (..._TMP_RSA_CB).
8645 [Bodo Moeller; problem reported by Amit Chopra]
8647 *) Add missing DSA library text string. Work around for some IIS
8648 key files with invalid SEQUENCE encoding.
8651 *) Add a document (doc/standards.txt) that list all kinds of standards
8652 and so on that are implemented in OpenSSL.
8655 *) Enhance c_rehash script. Old version would mishandle certificates
8656 with the same subject name hash and wouldn't handle CRLs at all.
8657 Added -fingerprint option to crl utility, to support new c_rehash
8661 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8664 *) Fix for SSL server purpose checking. Server checking was
8665 rejecting certificates which had extended key usage present
8666 but no ssl client purpose.
8667 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8669 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8670 is a little unclear about how a blank password is handled.
8671 Since the password in encoded as a BMPString with terminating
8672 double NULL a zero length password would end up as just the
8673 double NULL. However no password at all is different and is
8674 handled differently in the PKCS#12 key generation code. NS
8675 treats a blank password as zero length. MSIE treats it as no
8676 password on export: but it will try both on import. We now do
8677 the same: PKCS12_parse() tries zero length and no password if
8678 the password is set to "" or NULL (NULL is now a valid password:
8679 it wasn't before) as does the pkcs12 application.
8682 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8683 perror when PEM_read_bio_X509_REQ fails, the error message must
8684 be obtained from the error queue.
8687 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8688 it in ERR_remove_state if appropriate, and change ERR_get_state
8689 accordingly to avoid race conditions (this is necessary because
8690 thread_hash is no longer constant once set).
8693 *) Bugfix for linux-elf makefile.one.
8696 *) RSA_get_default_method() will now cause a default
8697 RSA_METHOD to be chosen if one doesn't exist already.
8698 Previously this was only set during a call to RSA_new()
8699 or RSA_new_method(NULL) meaning it was possible for
8700 RSA_get_default_method() to return NULL.
8703 *) Added native name translation to the existing DSO code
8704 that will convert (if the flag to do so is set) filenames
8705 that are sufficiently small and have no path information
8706 into a canonical native form. Eg. "blah" converted to
8707 "libblah.so" or "blah.dll" etc.
8710 *) New function ERR_error_string_n(e, buf, len) which is like
8711 ERR_error_string(e, buf), but writes at most 'len' bytes
8712 including the 0 terminator. For ERR_error_string_n, 'buf'
8714 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8716 *) CONF library reworked to become more general. A new CONF
8717 configuration file reader "class" is implemented as well as a
8718 new functions (NCONF_*, for "New CONF") to handle it. The now
8719 old CONF_* functions are still there, but are reimplemented to
8720 work in terms of the new functions. Also, a set of functions
8721 to handle the internal storage of the configuration data is
8722 provided to make it easier to write new configuration file
8723 reader "classes" (I can definitely see something reading a
8724 configuration file in XML format, for example), called _CONF_*,
8725 or "the configuration storage API"...
8727 The new configuration file reading functions are:
8729 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8730 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8732 NCONF_default, NCONF_WIN32
8734 NCONF_dump_fp, NCONF_dump_bio
8736 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8737 NCONF_new creates a new CONF object. This works in the same way
8738 as other interfaces in OpenSSL, like the BIO interface.
8739 NCONF_dump_* dump the internal storage of the configuration file,
8740 which is useful for debugging. All other functions take the same
8741 arguments as the old CONF_* functions wth the exception of the
8742 first that must be a `CONF *' instead of a `LHASH *'.
8744 To make it easer to use the new classes with the old CONF_* functions,
8745 the function CONF_set_default_method is provided.
8748 *) Add '-tls1' option to 'openssl ciphers', which was already
8749 mentioned in the documentation but had not been implemented.
8750 (This option is not yet really useful because even the additional
8751 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8754 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8755 OpenSSL-based applications) load shared libraries and bind to
8756 them in a portable way.
8757 [Geoff Thorpe, with contributions from Richard Levitte]
8759 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8761 *) Make sure _lrotl and _lrotr are only used with MSVC.
8763 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8764 (the default implementation of RAND_status).
8766 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8767 to '-clrext' (= clear extensions), as intended and documented.
8768 [Bodo Moeller; inconsistency pointed out by Michael Attili
8771 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8772 was larger than the MD block size.
8773 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8775 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8776 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8777 using the passed key: if the passed key was a private key the result
8778 of X509_print(), for example, would be to print out all the private key
8782 *) des_quad_cksum() byte order bug fix.
8783 [Ulf Möller, using the problem description in krb4-0.9.7, where
8784 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8786 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8788 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8790 *) For easily testing in shell scripts whether some command
8791 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8792 returns with exit code 0 iff no command of the given name is available.
8793 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8794 the output goes to stdout and nothing is printed to stderr.
8795 Additional arguments are always ignored.
8797 Since for each cipher there is a command of the same name,
8798 the 'no-cipher' compilation switches can be tested this way.
8800 ('openssl no-XXX' is not able to detect pseudo-commands such
8801 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8804 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8807 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8808 is set; it will be thrown away anyway because each handshake creates
8810 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8811 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8812 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
8813 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8816 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8817 'Q' and 'R' lose their special meanings (quit/renegotiate).
8818 This is part of what -quiet does; unlike -quiet, -ign_eof
8819 does not suppress any output.
8822 *) Add compatibility options to the purpose and trust code. The
8823 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8824 accepts a certificate or CA, this was the previous behaviour,
8825 with all the associated security issues.
8827 X509_TRUST_COMPAT is the old trust behaviour: only and
8828 automatically trust self signed roots in certificate store. A
8829 new trust setting X509_TRUST_DEFAULT is used to specify that
8830 a purpose has no associated trust setting and it should instead
8831 use the value in the default purpose.
8834 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8835 and fix a memory leak.
8838 *) In util/mkerr.pl (which implements 'make errors'), preserve
8839 reason strings from the previous version of the .c file, as
8840 the default to have only downcase letters (and digits) in
8841 automatically generated reasons codes is not always appropriate.
8844 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8845 using strerror. Previously, ERR_reason_error_string() returned
8846 library names as reason strings for SYSerr; but SYSerr is a special
8847 case where small numbers are errno values, not library numbers.
8850 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8851 converts DSA parameters into DH parameters. (When creating parameters,
8852 DSA_generate_parameters is used.)
8855 *) Include 'length' (recommended exponent length) in C code generated
8856 by 'openssl dhparam -C'.
8859 *) The second argument to set_label in perlasm was already being used
8860 so couldn't be used as a "file scope" flag. Moved to third argument
8864 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8865 instead of RAND_bytes for encryption IVs and salts.
8868 *) Include RAND_status() into RAND_METHOD instead of implementing
8869 it only for md_rand.c Otherwise replacing the PRNG by calling
8870 RAND_set_rand_method would be impossible.
8873 *) Don't let DSA_generate_key() enter an infinite loop if the random
8874 number generation fails.
8877 *) New 'rand' application for creating pseudo-random output.
8880 *) Added configuration support for Linux/IA64
8881 [Rolf Haberrecker <rolf@suse.de>]
8883 *) Assembler module support for Mingw32.
8886 *) Shared library support for HPUX (in shlib/).
8887 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8889 *) Shared library support for Solaris gcc.
8890 [Lutz Behnke <behnke@trustcenter.de>]
8892 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8894 *) PKCS7_encrypt() was adding text MIME headers twice because they
8895 were added manually and by SMIME_crlf_copy().
8898 *) In bntest.c don't call BN_rand with zero bits argument.
8899 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8901 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8902 case was implemented. This caused BN_div_recp() to fail occasionally.
8905 *) Add an optional second argument to the set_label() in the perl
8906 assembly language builder. If this argument exists and is set
8907 to 1 it signals that the assembler should use a symbol whose
8908 scope is the entire file, not just the current function. This
8909 is needed with MASM which uses the format label:: for this scope.
8910 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8912 *) Change the ASN1 types so they are typedefs by default. Before
8913 almost all types were #define'd to ASN1_STRING which was causing
8914 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8918 *) Change names of new functions to the new get1/get0 naming
8919 convention: After 'get1', the caller owns a reference count
8920 and has to call ..._free; 'get0' returns a pointer to some
8921 data structure without incrementing reference counters.
8922 (Some of the existing 'get' functions increment a reference
8923 counter, some don't.)
8924 Similarly, 'set1' and 'add1' functions increase reference
8925 counters or duplicate objects.
8928 *) Allow for the possibility of temp RSA key generation failure:
8929 the code used to assume it always worked and crashed on failure.
8932 *) Fix potential buffer overrun problem in BIO_printf().
8933 [Ulf Möller, using public domain code by Patrick Powell; problem
8934 pointed out by David Sacerdote <das33@cornell.edu>]
8936 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8937 RAND_egd() and RAND_status(). In the command line application,
8938 the EGD socket can be specified like a seed file using RANDFILE
8942 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8943 Some CAs (e.g. Verisign) distribute certificates in this form.
8946 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8947 list to exclude them. This means that no special compilation option
8948 is needed to use anonymous DH: it just needs to be included in the
8952 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8953 EVP_MD_type. The old functionality is available in a new macro called
8954 EVP_MD_md(). Change code that uses it and update docs.
8957 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8958 where the 'void *' argument is replaced by a function pointer argument.
8959 Previously 'void *' was abused to point to functions, which works on
8960 many platforms, but is not correct. As these functions are usually
8961 called by macros defined in OpenSSL header files, most source code
8962 should work without changes.
8965 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8966 sections with information on -D... compiler switches used for
8967 compiling the library so that applications can see them. To enable
8968 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8969 must be defined. E.g.,
8970 #define OPENSSL_ALGORITHM_DEFINES
8971 #include <openssl/opensslconf.h>
8972 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8973 [Richard Levitte, Ulf and Bodo Möller]
8975 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8979 *) Change the 'other' type in certificate aux info to a STACK_OF
8980 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8981 the required ASN1 format: arbitrary types determined by an OID.
8984 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8985 argument to 'req'. This is not because the function is newer or
8986 better than others it just uses the work 'NEW' in the certificate
8987 request header lines. Some software needs this.
8990 *) Reorganise password command line arguments: now passwords can be
8991 obtained from various sources. Delete the PEM_cb function and make
8992 it the default behaviour: i.e. if the callback is NULL and the
8993 usrdata argument is not NULL interpret it as a null terminated pass
8994 phrase. If usrdata and the callback are NULL then the pass phrase
8995 is prompted for as usual.
8998 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8999 the support is automatically enabled. The resulting binaries will
9000 autodetect the card and use it if present.
9001 [Ben Laurie and Compaq Inc.]
9003 *) Work around for Netscape hang bug. This sends certificate request
9004 and server done in one record. Since this is perfectly legal in the
9005 SSL/TLS protocol it isn't a "bug" option and is on by default. See
9006 the bugs/SSLv3 entry for more info.
9009 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
9012 *) Add -rand argument to smime and pkcs12 applications and read/write
9016 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
9019 *) Add command line password options to the remaining applications.
9022 *) Bug fix for BN_div_recp() for numerators with an even number of
9026 *) More tests in bntest.c, and changed test_bn output.
9029 *) ./config recognizes MacOS X now.
9032 *) Bug fix for BN_div() when the first words of num and divsor are
9033 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
9036 *) Add support for various broken PKCS#8 formats, and command line
9037 options to produce them.
9040 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
9041 get temporary BIGNUMs from a BN_CTX.
9044 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
9048 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
9049 include a #define from the old name to the new. The original intent
9050 was that statically linked binaries could for example just call
9051 SSLeay_add_all_ciphers() to just add ciphers to the table and not
9052 link with digests. This never worked becayse SSLeay_add_all_digests()
9053 and SSLeay_add_all_ciphers() were in the same source file so calling
9054 one would link with the other. They are now in separate source files.
9057 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9060 *) Use a less unusual form of the Miller-Rabin primality test (it used
9061 a binary algorithm for exponentiation integrated into the Miller-Rabin
9062 loop, our standard modexp algorithms are faster).
9065 *) Support for the EBCDIC character set completed.
9066 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9068 *) Source code cleanups: use const where appropriate, eliminate casts,
9069 use void * instead of char * in lhash.
9072 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9073 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9074 this the server could overwrite ephemeral keys that the client
9078 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9079 using 50 iterations of the Rabin-Miller test.
9081 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9082 iterations of the Rabin-Miller test as required by the appendix
9083 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9084 As BN_is_prime_fasttest includes trial division, DSA parameter
9085 generation becomes much faster.
9087 This implies a change for the callback functions in DSA_is_prime
9088 and DSA_generate_parameters: The callback function is called once
9089 for each positive witness in the Rabin-Miller test, not just
9090 occasionally in the inner loop; and the parameters to the
9091 callback function now provide an iteration count for the outer
9092 loop rather than for the current invocation of the inner loop.
9093 DSA_generate_parameters additionally can call the callback
9094 function with an 'iteration count' of -1, meaning that a
9095 candidate has passed the trial division test (when q is generated
9096 from an application-provided seed, trial division is skipped).
9099 *) New function BN_is_prime_fasttest that optionally does trial
9100 division before starting the Rabin-Miller test and has
9101 an additional BN_CTX * argument (whereas BN_is_prime always
9102 has to allocate at least one BN_CTX).
9103 'callback(1, -1, cb_arg)' is called when a number has passed the
9104 trial division stage.
9107 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9111 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9114 *) New function BN_pseudo_rand().
9117 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9118 bignum version of BN_from_montgomery() with the working code from
9119 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9123 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9124 made it impossible to use the same SSL_SESSION data structure in
9125 SSL2 clients in multiple threads.
9128 *) The return value of RAND_load_file() no longer counts bytes obtained
9129 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9130 to seed the PRNG (previously an explicit byte count was required).
9131 [Ulf Möller, Bodo Möller]
9133 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9134 used (char *) instead of (void *) and had casts all over the place.
9137 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9140 *) Retain source code compatibility for BN_prime_checks macro:
9141 BN_is_prime(..., BN_prime_checks, ...) now uses
9142 BN_prime_checks_for_size to determine the appropriate number of
9143 Rabin-Miller iterations.
9146 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9147 DH_CHECK_P_NOT_SAFE_PRIME.
9148 (Check if this is true? OpenPGP calls them "strong".)
9151 *) Merge the functionality of "dh" and "gendh" programs into a new program
9152 "dhparam". The old programs are retained for now but will handle DH keys
9153 (instead of parameters) in future.
9156 *) Make the ciphers, s_server and s_client programs check the return values
9157 when a new cipher list is set.
9160 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9161 ciphers. Before when the 56bit ciphers were enabled the sorting was
9164 The syntax for the cipher sorting has been extended to support sorting by
9165 cipher-strength (using the strength_bits hard coded in the tables).
9166 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9168 Fix a bug in the cipher-command parser: when supplying a cipher command
9169 string with an "undefined" symbol (neither command nor alphanumeric
9170 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9171 an error is flagged.
9173 Due to the strength-sorting extension, the code of the
9174 ssl_create_cipher_list() function was completely rearranged. I hope that
9175 the readability was also increased :-)
9176 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9178 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9179 for the first serial number and places 2 in the serial number file. This
9180 avoids problems when the root CA is created with serial number zero and
9181 the first user certificate has the same issuer name and serial number
9185 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9186 the new code. Add documentation for this stuff.
9189 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9190 X509_*() to X509at_*() on the grounds that they don't handle X509
9191 structures and behave in an analagous way to the X509v3 functions:
9192 they shouldn't be called directly but wrapper functions should be used
9195 So we also now have some wrapper functions that call the X509at functions
9196 when passed certificate requests. (TO DO: similar things can be done with
9197 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9198 things. Some of these need some d2i or i2d and print functionality
9199 because they handle more complex structures.)
9202 *) Add missing #ifndefs that caused missing symbols when building libssl
9203 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9204 NO_RSA in ssl/s2*.c.
9205 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9207 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9208 has a return value which indicates the quality of the random data
9209 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9210 error queue. New function RAND_pseudo_bytes() generates output that is
9211 guaranteed to be unique but not unpredictable. RAND_add is like
9212 RAND_seed, but takes an extra argument for an entropy estimate
9213 (RAND_seed always assumes full entropy).
9216 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9217 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9218 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9219 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9220 false-positive rate of at most 2^-80 for random input.
9223 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9226 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9227 in the 0.9.5 release), this returns the chain
9228 from an X509_CTX structure with a dup of the stack and all
9229 the X509 reference counts upped: so the stack will exist
9230 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9233 Also make SSL_SESSION_print() print out the verify return
9237 *) Add manpage for the pkcs12 command. Also change the default
9238 behaviour so MAC iteration counts are used unless the new
9239 -nomaciter option is used. This improves file security and
9240 only older versions of MSIE (4.0 for example) need it.
9243 *) Honor the no-xxx Configure options when creating .DEF files.
9246 *) Add PKCS#10 attributes to field table: challengePassword,
9247 unstructuredName and unstructuredAddress. These are taken from
9248 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9249 international characters are used.
9251 More changes to X509_ATTRIBUTE code: allow the setting of types
9252 based on strings. Remove the 'loc' parameter when adding
9253 attributes because these will be a SET OF encoding which is sorted
9257 *) Initial changes to the 'req' utility to allow request generation
9258 automation. This will allow an application to just generate a template
9259 file containing all the field values and have req construct the
9262 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9263 used all over the place including certificate requests and PKCS#7
9264 structures. They are currently handled manually where necessary with
9265 some primitive wrappers for PKCS#7. The new functions behave in a
9266 manner analogous to the X509 extension functions: they allow
9267 attributes to be looked up by NID and added.
9269 Later something similar to the X509V3 code would be desirable to
9270 automatically handle the encoding, decoding and printing of the
9271 more complex types. The string types like challengePassword can
9272 be handled by the string table functions.
9274 Also modified the multi byte string table handling. Now there is
9275 a 'global mask' which masks out certain types. The table itself
9276 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9277 is useful when for example there is only one permissible type
9278 (as in countryName) and using the mask might result in no valid
9282 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9283 SSL_get_peer_finished to allow applications to obtain the latest
9284 Finished messages sent to the peer or expected from the peer,
9285 respectively. (SSL_get_peer_finished is usually the Finished message
9286 actually received from the peer, otherwise the protocol will be aborted.)
9288 As the Finished message are message digests of the complete handshake
9289 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9290 be used for external authentication procedures when the authentication
9291 provided by SSL/TLS is not desired or is not enough.
9294 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9295 the host supports BWX extension and if Compaq C is present on the
9296 $PATH. Just exploiting of the BWX extension results in 20-30%
9297 performance kick for some algorithms, e.g. DES and RC4 to mention
9298 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9302 *) Add support for MS "fast SGC". This is arguably a violation of the
9303 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9304 weak crypto and after checking the certificate is SGC a second one
9305 with strong crypto. MS SGC stops the first handshake after receiving
9306 the server certificate message and sends a second client hello. Since
9307 a server will typically do all the time consuming operations before
9308 expecting any further messages from the client (server key exchange
9309 is the most expensive) there is little difference between the two.
9311 To get OpenSSL to support MS SGC we have to permit a second client
9312 hello message after we have sent server done. In addition we have to
9313 reset the MAC if we do get this second client hello.
9316 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9317 if a DER encoded private key is RSA or DSA traditional format. Changed
9318 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9319 format DER encoded private key. Newer code should use PKCS#8 format which
9320 has the key type encoded in the ASN1 structure. Added DER private key
9321 support to pkcs8 application.
9324 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9325 ciphersuites has been selected (as required by the SSL 3/TLS 1
9326 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9327 is set, we interpret this as a request to violate the specification
9328 (the worst that can happen is a handshake failure, and 'correct'
9329 behaviour would result in a handshake failure anyway).
9332 *) In SSL_CTX_add_session, take into account that there might be multiple
9333 SSL_SESSION structures with the same session ID (e.g. when two threads
9334 concurrently obtain them from an external cache).
9335 The internal cache can handle only one SSL_SESSION with a given ID,
9336 so if there's a conflict, we now throw out the old one to achieve
9340 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9341 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9342 some routines that use cipher OIDs: some ciphers do not have OIDs
9343 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9347 *) Simplify the trust setting structure and code. Now we just have
9348 two sequences of OIDs for trusted and rejected settings. These will
9349 typically have values the same as the extended key usage extension
9350 and any application specific purposes.
9352 The trust checking code now has a default behaviour: it will just
9353 check for an object with the same NID as the passed id. Functions can
9354 be provided to override either the default behaviour or the behaviour
9355 for a given id. SSL client, server and email already have functions
9356 in place for compatibility: they check the NID and also return "trusted"
9357 if the certificate is self signed.
9360 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9361 traditional format into an EVP_PKEY structure.
9364 *) Add a password callback function PEM_cb() which either prompts for
9365 a password if usr_data is NULL or otherwise assumes it is a null
9366 terminated password. Allow passwords to be passed on command line
9367 environment or config files in a few more utilities.
9370 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9371 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9372 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9373 Update documentation.
9376 *) Support for ASN1 "NULL" type. This could be handled before by using
9377 ASN1_TYPE but there wasn't any function that would try to read a NULL
9378 and produce an error if it couldn't. For compatibility we also have
9379 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9380 don't allocate anything because they don't need to.
9383 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9385 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9387 *) Rebuild of the memory allocation routines used by OpenSSL code and
9388 possibly others as well. The purpose is to make an interface that
9389 provide hooks so anyone can build a separate set of allocation and
9390 deallocation routines to be used by OpenSSL, for example memory
9391 pool implementations, or something else, which was previously hard
9392 since Malloc(), Realloc() and Free() were defined as macros having
9393 the values malloc, realloc and free, respectively (except for Win32
9394 compilations). The same is provided for memory debugging code.
9395 OpenSSL already comes with functionality to find memory leaks, but
9396 this gives people a chance to debug other memory problems.
9398 With these changes, a new set of functions and macros have appeared:
9400 CRYPTO_set_mem_debug_functions() [F]
9401 CRYPTO_get_mem_debug_functions() [F]
9402 CRYPTO_dbg_set_options() [F]
9403 CRYPTO_dbg_get_options() [F]
9404 CRYPTO_malloc_debug_init() [M]
9406 The memory debug functions are NULL by default, unless the library
9407 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9408 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9409 gives the standard debugging functions that come with OpenSSL) or
9410 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9411 provided by the library user) must be used. When the standard
9412 debugging functions are used, CRYPTO_dbg_set_options can be used to
9413 request additional information:
9414 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9415 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9417 Also, things like CRYPTO_set_mem_functions will always give the
9418 expected result (the new set of functions is used for allocation
9419 and deallocation) at all times, regardless of platform and compiler
9422 To finish it up, some functions that were never use in any other
9423 way than through macros have a new API and new semantic:
9426 CRYPTO_dbg_realloc()
9429 All macros of value have retained their old syntax.
9430 [Richard Levitte and Bodo Moeller]
9432 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9433 ordering of SMIMECapabilities wasn't in "strength order" and there
9434 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9438 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9439 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9440 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9442 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9443 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9444 functionality to handle multipart/signed properly) and a utility
9445 called 'smime' to call all this stuff. This is based on code I
9446 originally wrote for Celo who have kindly allowed it to be
9447 included in OpenSSL.
9450 *) Add variants des_set_key_checked and des_set_key_unchecked of
9451 des_set_key (aka des_key_sched). Global variable des_check_key
9452 decides which of these is called by des_set_key; this way
9453 des_check_key behaves as it always did, but applications and
9454 the library itself, which was buggy for des_check_key == 1,
9455 have a cleaner way to pick the version they need.
9458 *) New function PKCS12_newpass() which changes the password of a
9462 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9463 dynamic mix. In both cases the ids can be used as an index into the
9464 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9465 functions so they accept a list of the field values and the
9466 application doesn't need to directly manipulate the X509_TRUST
9470 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9474 *) Modify the way the V3 extension code looks up extensions. This now
9475 works in a similar way to the object code: we have some "standard"
9476 extensions in a static table which is searched with OBJ_bsearch()
9477 and the application can add dynamic ones if needed. The file
9478 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9479 updated whenever a new extension is added to the core code and kept
9480 in ext_nid order. There is a simple program 'tabtest.c' which checks
9481 this. New extensions are not added too often so this file can readily
9482 be maintained manually.
9484 There are two big advantages in doing things this way. The extensions
9485 can be looked up immediately and no longer need to be "added" using
9486 X509V3_add_standard_extensions(): this function now does nothing.
9487 [Side note: I get *lots* of email saying the extension code doesn't
9488 work because people forget to call this function]
9489 Also no dynamic allocation is done unless new extensions are added:
9490 so if we don't add custom extensions there is no need to call
9491 X509V3_EXT_cleanup().
9494 *) Modify enc utility's salting as follows: make salting the default. Add a
9495 magic header, so unsalted files fail gracefully instead of just decrypting
9496 to garbage. This is because not salting is a big security hole, so people
9497 should be discouraged from doing it.
9500 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9501 digest to be passed on the command line but it only used this
9502 parameter when signing a certificate. Modified so all relevant
9503 operations are affected by the digest parameter including the
9504 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9505 DSA key was used because it didn't fix the digest.
9508 *) Initial certificate chain verify code. Currently tests the untrusted
9509 certificates for consistency with the verify purpose (which is set
9510 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9512 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9513 this is because it will reject chains with invalid extensions whereas
9514 every previous version of OpenSSL and SSLeay made no checks at all.
9516 Trust code: checks the root CA for the relevant trust settings. Trust
9517 settings have an initial value consistent with the verify purpose: e.g.
9518 if the verify purpose is for SSL client use it expects the CA to be
9519 trusted for SSL client use. However the default value can be changed to
9520 permit custom trust settings: one example of this would be to only trust
9521 certificates from a specific "secure" set of CAs.
9523 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9524 which should be used for version portability: especially since the
9525 verify structure is likely to change more often now.
9527 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9528 to set them. If not set then assume SSL clients will verify SSL servers
9531 Two new options to the verify program: -untrusted allows a set of
9532 untrusted certificates to be passed in and -purpose which sets the
9533 intended purpose of the certificate. If a purpose is set then the
9534 new chain verify code is used to check extension consistency.
9537 *) Support for the authority information access extension.
9540 *) Modify RSA and DSA PEM read routines to transparently handle
9541 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9542 public keys in a format compatible with certificate
9543 SubjectPublicKeyInfo structures. Unfortunately there were already
9544 functions called *_PublicKey_* which used various odd formats so
9545 these are retained for compatibility: however the DSA variants were
9546 never in a public release so they have been deleted. Changed dsa/rsa
9547 utilities to handle the new format: note no releases ever handled public
9548 keys so we should be OK.
9550 The primary motivation for this change is to avoid the same fiasco
9551 that dogs private keys: there are several incompatible private key
9552 formats some of which are standard and some OpenSSL specific and
9553 require various evil hacks to allow partial transparent handling and
9554 even then it doesn't work with DER formats. Given the option anything
9555 other than PKCS#8 should be dumped: but the other formats have to
9556 stay in the name of compatibility.
9558 With public keys and the benefit of hindsight one standard format
9559 is used which works with EVP_PKEY, RSA or DSA structures: though
9560 it clearly returns an error if you try to read the wrong kind of key.
9562 Added a -pubkey option to the 'x509' utility to output the public key.
9563 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9564 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9565 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9566 that do the same as the EVP_PKEY_assign_*() except they up the
9567 reference count of the added key (they don't "swallow" the
9571 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9572 CRLs would fail if the file contained no certificates or no CRLs:
9573 added a new function to read in both types and return the number
9574 read: this means that if none are read it will be an error. The
9575 DER versions of the certificate and CRL reader would always fail
9576 because it isn't possible to mix certificates and CRLs in DER format
9577 without choking one or the other routine. Changed this to just read
9578 a certificate: this is the best we can do. Also modified the code
9579 in apps/verify.c to take notice of return codes: it was previously
9580 attempting to read in certificates from NULL pointers and ignoring
9581 any errors: this is one reason why the cert and CRL reader seemed
9582 to work. It doesn't check return codes from the default certificate
9583 routines: these may well fail if the certificates aren't installed.
9586 *) Code to support otherName option in GeneralName.
9589 *) First update to verify code. Change the verify utility
9590 so it warns if it is passed a self signed certificate:
9591 for consistency with the normal behaviour. X509_verify
9592 has been modified to it will now verify a self signed
9593 certificate if *exactly* the same certificate appears
9594 in the store: it was previously impossible to trust a
9595 single self signed certificate. This means that:
9596 openssl verify ss.pem
9597 now gives a warning about a self signed certificate but
9598 openssl verify -CAfile ss.pem ss.pem
9602 *) For servers, store verify_result in SSL_SESSION data structure
9603 (and add it to external session representation).
9604 This is needed when client certificate verifications fails,
9605 but an application-provided verification callback (set by
9606 SSL_CTX_set_cert_verify_callback) allows accepting the session
9607 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9608 but returns 1): When the session is reused, we have to set
9609 ssl->verify_result to the appropriate error code to avoid
9611 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9613 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9614 case in PKCS7_dataInit() where the signed PKCS7 structure
9615 didn't contain any existing data because it was being created.
9616 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9618 *) Add a salt to the key derivation routines in enc.c. This
9619 forms the first 8 bytes of the encrypted file. Also add a
9620 -S option to allow a salt to be input on the command line.
9623 *) New function X509_cmp(). Oddly enough there wasn't a function
9624 to compare two certificates. We do this by working out the SHA1
9625 hash and comparing that. X509_cmp() will be needed by the trust
9629 *) SSL_get1_session() is like SSL_get_session(), but increments
9630 the reference count in the SSL_SESSION returned.
9631 [Geoff Thorpe <geoff@eu.c2.net>]
9633 *) Fix for 'req': it was adding a null to request attributes.
9634 Also change the X509_LOOKUP and X509_INFO code to handle
9635 certificate auxiliary information.
9638 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9642 *) Add the possibility to add extra information to the memory leak
9643 detecting output, to form tracebacks, showing from where each
9644 allocation was originated: CRYPTO_push_info("constant string") adds
9645 the string plus current file name and line number to a per-thread
9646 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9647 is like calling CYRPTO_pop_info() until the stack is empty.
9648 Also updated memory leak detection code to be multi-thread-safe.
9651 *) Add options -text and -noout to pkcs7 utility and delete the
9652 encryption options which never did anything. Update docs.
9655 *) Add options to some of the utilities to allow the pass phrase
9656 to be included on either the command line (not recommended on
9657 OSes like Unix) or read from the environment. Update the
9658 manpages and fix a few bugs.
9661 *) Add a few manpages for some of the openssl commands.
9664 *) Fix the -revoke option in ca. It was freeing up memory twice,
9665 leaking and not finding already revoked certificates.
9668 *) Extensive changes to support certificate auxiliary information.
9669 This involves the use of X509_CERT_AUX structure and X509_AUX
9670 functions. An X509_AUX function such as PEM_read_X509_AUX()
9671 can still read in a certificate file in the usual way but it
9672 will also read in any additional "auxiliary information". By
9673 doing things this way a fair degree of compatibility can be
9674 retained: existing certificates can have this information added
9675 using the new 'x509' options.
9677 Current auxiliary information includes an "alias" and some trust
9678 settings. The trust settings will ultimately be used in enhanced
9679 certificate chain verification routines: currently a certificate
9680 can only be trusted if it is self signed and then it is trusted
9684 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9685 The problem was that one of the replacement routines had not been working
9686 since SSLeay releases. For now the offending routine has been replaced
9687 with non-optimised assembler. Even so, this now gives around 95%
9688 performance improvement for 1024 bit RSA signs.
9691 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9692 handling. Most clients have the effective key size in bits equal to
9693 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9694 A few however don't do this and instead use the size of the decrypted key
9695 to determine the RC2 key length and the AlgorithmIdentifier to determine
9696 the effective key length. In this case the effective key length can still
9697 be 40 bits but the key length can be 168 bits for example. This is fixed
9698 by manually forcing an RC2 key into the EVP_PKEY structure because the
9699 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9700 the key length and effective key length are equal.
9703 *) Add a bunch of functions that should simplify the creation of
9704 X509_NAME structures. Now you should be able to do:
9705 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9706 and have it automatically work out the correct field type and fill in
9707 the structures. The more adventurous can try:
9708 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9709 and it will (hopefully) work out the correct multibyte encoding.
9712 *) Change the 'req' utility to use the new field handling and multibyte
9713 copy routines. Before the DN field creation was handled in an ad hoc
9714 way in req, ca, and x509 which was rather broken and didn't support
9715 BMPStrings or UTF8Strings. Since some software doesn't implement
9716 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9717 using the dirstring_type option. See the new comment in the default
9718 openssl.cnf for more info.
9721 *) Make crypto/rand/md_rand.c more robust:
9722 - Assure unique random numbers after fork().
9723 - Make sure that concurrent threads access the global counter and
9724 md serializably so that we never lose entropy in them
9725 or use exactly the same state in multiple threads.
9726 Access to the large state is not always serializable because
9727 the additional locking could be a performance killer, and
9728 md should be large enough anyway.
9731 *) New file apps/app_rand.c with commonly needed functionality
9732 for handling the random seed file.
9734 Use the random seed file in some applications that previously did not:
9736 dsaparam -genkey (which also ignored its '-rand' option),
9739 x509 (when signing).
9740 Except on systems with /dev/urandom, it is crucial to have a random
9741 seed file at least for key creation, DSA signing, and for DH exchanges;
9742 for RSA signatures we could do without one.
9744 gendh and gendsa (unlike genrsa) used to read only the first byte
9745 of each file listed in the '-rand' option. The function as previously
9746 found in genrsa is now in app_rand.c and is used by all programs
9747 that support '-rand'.
9750 *) In RAND_write_file, use mode 0600 for creating files;
9751 don't just chmod when it may be too late.
9754 *) Report an error from X509_STORE_load_locations
9755 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9758 *) New function ASN1_mbstring_copy() this copies a string in either
9759 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9760 into an ASN1_STRING type. A mask of permissible types is passed
9761 and it chooses the "minimal" type to use or an error if not type
9765 *) Add function equivalents to the various macros in asn1.h. The old
9766 macros are retained with an M_ prefix. Code inside the library can
9767 use the M_ macros. External code (including the openssl utility)
9768 should *NOT* in order to be "shared library friendly".
9771 *) Add various functions that can check a certificate's extensions
9772 to see if it usable for various purposes such as SSL client,
9773 server or S/MIME and CAs of these types. This is currently
9774 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9775 verification. Also added a -purpose flag to x509 utility to
9776 print out all the purposes.
9779 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9783 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9784 for, obtain and decode and extension and obtain its critical flag.
9785 This allows all the necessary extension code to be handled in a
9786 single function call.
9789 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9790 platforms. See crypto/rc4/rc4_enc.c for further details.
9793 *) New -noout option to asn1parse. This causes no output to be produced
9794 its main use is when combined with -strparse and -out to extract data
9795 from a file (which may not be in ASN.1 format).
9798 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9799 when producing the local key id.
9800 [Richard Levitte <levitte@stacken.kth.se>]
9802 *) New option -dhparam in s_server. This allows a DH parameter file to be
9803 stated explicitly. If it is not stated then it tries the first server
9804 certificate file. The previous behaviour hard coded the filename
9808 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9809 a public key to be input or output. For example:
9810 openssl rsa -in key.pem -pubout -out pubkey.pem
9811 Also added necessary DSA public key functions to handle this.
9814 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9815 in the message. This was handled by allowing
9816 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9817 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9819 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9820 to the end of the strings whereas this didn't. This would cause problems
9821 if strings read with d2i_ASN1_bytes() were later modified.
9822 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9824 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9825 data and it contains EOF it will end up returning an error. This is
9826 caused by input 46 bytes long. The cause is due to the way base64
9827 BIOs find the start of base64 encoded data. They do this by trying a
9828 trial decode on each line until they find one that works. When they
9829 do a flag is set and it starts again knowing it can pass all the
9830 data directly through the decoder. Unfortunately it doesn't reset
9831 the context it uses. This means that if EOF is reached an attempt
9832 is made to pass two EOFs through the context and this causes the
9833 resulting error. This can also cause other problems as well. As is
9834 usual with these problems it takes *ages* to find and the fix is
9835 trivial: move one line.
9836 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9838 *) Ugly workaround to get s_client and s_server working under Windows. The
9839 old code wouldn't work because it needed to select() on sockets and the
9840 tty (for keypresses and to see if data could be written). Win32 only
9841 supports select() on sockets so we select() with a 1s timeout on the
9842 sockets and then see if any characters are waiting to be read, if none
9843 are present then we retry, we also assume we can always write data to
9844 the tty. This isn't nice because the code then blocks until we've
9845 received a complete line of data and it is effectively polling the
9846 keyboard at 1s intervals: however it's quite a bit better than not
9847 working at all :-) A dedicated Windows application might handle this
9848 with an event loop for example.
9851 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9852 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9853 will be called when RSA_sign() and RSA_verify() are used. This is useful
9854 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9855 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9856 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9857 This necessitated the support of an extra signature type NID_md5_sha1
9858 for SSL signatures and modifications to the SSL library to use it instead
9859 of calling RSA_public_decrypt() and RSA_private_encrypt().
9862 *) Add new -verify -CAfile and -CApath options to the crl program, these
9863 will lookup a CRL issuers certificate and verify the signature in a
9864 similar way to the verify program. Tidy up the crl program so it
9865 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9866 less strict. It will now permit CRL extensions even if it is not
9867 a V2 CRL: this will allow it to tolerate some broken CRLs.
9870 *) Initialize all non-automatic variables each time one of the openssl
9871 sub-programs is started (this is necessary as they may be started
9872 multiple times from the "OpenSSL>" prompt).
9873 [Lennart Bang, Bodo Moeller]
9875 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9876 removing all other RSA functionality (this is what NO_RSA does). This
9877 is so (for example) those in the US can disable those operations covered
9878 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9882 *) Non-copying interface to BIO pairs.
9883 (still largely untested)
9886 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9887 ASCII string. This was handled independently in various places before.
9890 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9891 UTF8 strings a character at a time.
9894 *) Use client_version from client hello to select the protocol
9895 (s23_srvr.c) and for RSA client key exchange verification
9896 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9899 *) Add various utility functions to handle SPKACs, these were previously
9900 handled by poking round in the structure internals. Added new function
9901 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9902 print, verify and generate SPKACs. Based on an original idea from
9903 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9906 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9909 *) Allow the config file extension section to be overwritten on the
9910 command line. Based on an original idea from Massimiliano Pala
9911 <madwolf@comune.modena.it>. The new option is called -extensions
9912 and can be applied to ca, req and x509. Also -reqexts to override
9913 the request extensions in req and -crlexts to override the crl extensions
9917 *) Add new feature to the SPKAC handling in ca. Now you can include
9918 the same field multiple times by preceding it by "XXXX." for example:
9921 this is the same syntax as used in the req config file.
9924 *) Allow certificate extensions to be added to certificate requests. These
9925 are specified in a 'req_extensions' option of the req section of the
9926 config file. They can be printed out with the -text option to req but
9927 are otherwise ignored at present.
9930 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9931 data read consists of only the final block it would not decrypted because
9932 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9933 A misplaced 'break' also meant the decrypted final block might not be
9934 copied until the next read.
9937 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9938 a few extra parameters to the DH structure: these will be useful if
9939 for example we want the value of 'q' or implement X9.42 DH.
9942 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9943 provides hooks that allow the default DSA functions or functions on a
9944 "per key" basis to be replaced. This allows hardware acceleration and
9945 hardware key storage to be handled without major modification to the
9946 library. Also added low level modexp hooks and CRYPTO_EX structure and
9947 associated functions.
9950 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9951 as "read only": it can't be written to and the buffer it points to will
9952 not be freed. Reading from a read only BIO is much more efficient than
9953 a normal memory BIO. This was added because there are several times when
9954 an area of memory needs to be read from a BIO. The previous method was
9955 to create a memory BIO and write the data to it, this results in two
9956 copies of the data and an O(n^2) reading algorithm. There is a new
9957 function BIO_new_mem_buf() which creates a read only memory BIO from
9958 an area of memory. Also modified the PKCS#7 routines to use read only
9962 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9963 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9964 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9965 but a retry condition occured while trying to read the rest.
9968 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9969 NID_pkcs7_encrypted by default: this was wrong since this should almost
9970 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9971 the encrypted data type: this is a more sensible place to put it and it
9972 allows the PKCS#12 code to be tidied up that duplicated this
9976 *) Changed obj_dat.pl script so it takes its input and output files on
9977 the command line. This should avoid shell escape redirection problems
9981 *) Initial support for certificate extension requests, these are included
9982 in things like Xenroll certificate requests. Included functions to allow
9983 extensions to be obtained and added.
9986 *) -crlf option to s_client and s_server for sending newlines as
9987 CRLF (as required by many protocols).
9990 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9992 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9993 [Ralf S. Engelschall]
9995 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9996 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9998 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
10002 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
10003 DH parameters/keys (q is lost during that conversion, but the resulting
10004 DH parameters contain its length).
10006 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
10007 much faster than DH_generate_parameters (which creates parameters
10008 where p = 2*q + 1), and also the smaller q makes DH computations
10009 much more efficient (160-bit exponentiation instead of 1024-bit
10010 exponentiation); so this provides a convenient way to support DHE
10011 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
10012 utter importance to use
10013 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10015 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10016 when such DH parameters are used, because otherwise small subgroup
10017 attacks may become possible!
10020 *) Avoid memory leak in i2d_DHparams.
10023 *) Allow the -k option to be used more than once in the enc program:
10024 this allows the same encrypted message to be read by multiple recipients.
10027 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
10028 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
10029 it will always use the numerical form of the OID, even if it has a short
10033 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
10034 method only got called if p,q,dmp1,dmq1,iqmp components were present,
10035 otherwise bn_mod_exp was called. In the case of hardware keys for example
10036 no private key components need be present and it might store extra data
10037 in the RSA structure, which cannot be accessed from bn_mod_exp.
10038 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
10039 private key operations.
10042 *) Added support for SPARC Linux.
10045 *) pem_password_cb function type incompatibly changed from
10046 typedef int pem_password_cb(char *buf, int size, int rwflag);
10048 ....(char *buf, int size, int rwflag, void *userdata);
10049 so that applications can pass data to their callbacks:
10050 The PEM[_ASN1]_{read,write}... functions and macros now take an
10051 additional void * argument, which is just handed through whenever
10052 the password callback is called.
10053 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10055 New function SSL_CTX_set_default_passwd_cb_userdata.
10057 Compatibility note: As many C implementations push function arguments
10058 onto the stack in reverse order, the new library version is likely to
10059 interoperate with programs that have been compiled with the old
10060 pem_password_cb definition (PEM_whatever takes some data that
10061 happens to be on the stack as its last argument, and the callback
10062 just ignores this garbage); but there is no guarantee whatsoever that
10065 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10066 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10067 problems not only on Windows, but also on some Unix platforms.
10068 To avoid problematic command lines, these definitions are now in an
10069 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10070 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10073 *) MIPS III/IV assembler module is reimplemented.
10076 *) More DES library cleanups: remove references to srand/rand and
10077 delete an unused file.
10080 *) Add support for the the free Netwide assembler (NASM) under Win32,
10081 since not many people have MASM (ml) and it can be hard to obtain.
10082 This is currently experimental but it seems to work OK and pass all
10083 the tests. Check out INSTALL.W32 for info.
10086 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10087 without temporary keys kept an extra copy of the server key,
10088 and connections with temporary keys did not free everything in case
10092 *) New function RSA_check_key and new openssl rsa option -check
10093 for verifying the consistency of RSA keys.
10094 [Ulf Moeller, Bodo Moeller]
10096 *) Various changes to make Win32 compile work:
10097 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10098 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10099 comparison" warnings.
10100 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10103 *) Add a debugging option to PKCS#5 v2 key generation function: when
10104 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10105 derived keys are printed to stderr.
10108 *) Copy the flags in ASN1_STRING_dup().
10109 [Roman E. Pavlov <pre@mo.msk.ru>]
10111 *) The x509 application mishandled signing requests containing DSA
10112 keys when the signing key was also DSA and the parameters didn't match.
10114 It was supposed to omit the parameters when they matched the signing key:
10115 the verifying software was then supposed to automatically use the CA's
10116 parameters if they were absent from the end user certificate.
10118 Omitting parameters is no longer recommended. The test was also
10119 the wrong way round! This was probably due to unusual behaviour in
10120 EVP_cmp_parameters() which returns 1 if the parameters match.
10121 This meant that parameters were omitted when they *didn't* match and
10122 the certificate was useless. Certificates signed with 'ca' didn't have
10124 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10126 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10127 The interface is as follows:
10128 Applications can use
10129 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10130 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10131 "off" is now the default.
10132 The library internally uses
10133 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10134 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10135 to disable memory-checking temporarily.
10137 Some inconsistent states that previously were possible (and were
10138 even the default) are now avoided.
10140 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10141 with each memory chunk allocated; this is occasionally more helpful
10142 than just having a counter.
10144 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10146 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10150 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10151 which largely parallels "options", but is for changing API behaviour,
10152 whereas "options" are about protocol behaviour.
10153 Initial "mode" flags are:
10155 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10156 a single record has been written.
10157 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10158 retries use the same buffer location.
10159 (But all of the contents must be
10163 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10166 *) Fix problems with no-hmac etc.
10167 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10169 *) New functions RSA_get_default_method(), RSA_set_method() and
10170 RSA_get_method(). These allows replacement of RSA_METHODs without having
10171 to mess around with the internals of an RSA structure.
10174 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10175 Also really enable memory leak checks in openssl.c and in some
10177 [Chad C. Mulligan, Bodo Moeller]
10179 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10180 up the length of negative integers. This has now been simplified to just
10181 store the length when it is first determined and use it later, rather
10182 than trying to keep track of where data is copied and updating it to
10184 [Steve Henson, reported by Brien Wheeler
10185 <bwheeler@authentica-security.com>]
10187 *) Add a new function PKCS7_signatureVerify. This allows the verification
10188 of a PKCS#7 signature but with the signing certificate passed to the
10189 function itself. This contrasts with PKCS7_dataVerify which assumes the
10190 certificate is present in the PKCS#7 structure. This isn't always the
10191 case: certificates can be omitted from a PKCS#7 structure and be
10192 distributed by "out of band" means (such as a certificate database).
10195 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10196 function prototypes in pem.h, also change util/mkdef.pl to add the
10197 necessary function names.
10200 *) mk1mf.pl (used by Windows builds) did not properly read the
10201 options set by Configure in the top level Makefile, and Configure
10202 was not even able to write more than one option correctly.
10203 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10206 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10207 file to be loaded from a BIO or FILE pointer. The BIO version will
10208 for example allow memory BIOs to contain config info.
10211 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10212 Whoever hopes to achieve shared-library compatibility across versions
10213 must use this, not the compile-time macro.
10214 (Exercise 0.9.4: Which is the minimum library version required by
10216 Note: All this applies only to multi-threaded programs, others don't
10220 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10221 through a BIO pair triggered the default case, i.e.
10222 SSLerr(...,SSL_R_UNKNOWN_STATE).
10225 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10226 can use the SSL library even if none of the specific BIOs is
10230 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10231 for the encoded length.
10232 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10234 *) Add initial documentation of the X509V3 functions.
10237 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10238 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10239 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10240 secure PKCS#8 private key format with a high iteration count.
10243 *) Fix determination of Perl interpreter: A perl or perl5
10244 _directory_ in $PATH was also accepted as the interpreter.
10245 [Ralf S. Engelschall]
10247 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10248 wrong with it but it was very old and did things like calling
10249 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10250 unusual formatting.
10253 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10254 to use the new extension code.
10257 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10258 with macros. This should make it easier to change their form, add extra
10259 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10263 *) Add to configuration table a new entry that can specify an alternative
10264 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10265 according to Mark Crispin <MRC@Panda.COM>.
10269 *) DES CBC did not update the IV. Weird.
10272 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10273 Changing the behaviour of the former might break existing programs --
10274 where IV updating is needed, des_ncbc_encrypt can be used.
10277 *) When bntest is run from "make test" it drives bc to check its
10278 calculations, as well as internally checking them. If an internal check
10279 fails, it needs to cause bc to give a non-zero result or make test carries
10280 on without noticing the failure. Fixed.
10283 *) DES library cleanups.
10286 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10287 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10288 ciphers. NOTE: although the key derivation function has been verified
10289 against some published test vectors it has not been extensively tested
10290 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10294 *) Instead of "mkdir -p", which is not fully portable, use new
10295 Perl script "util/mkdir-p.pl".
10298 *) Rewrite the way password based encryption (PBE) is handled. It used to
10299 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10300 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10301 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10302 the 'parameter' field of the AlgorithmIdentifier is passed to the
10303 underlying key generation function so it must do its own ASN1 parsing.
10304 This has also changed the EVP_PBE_CipherInit() function which now has a
10305 'parameter' argument instead of literal salt and iteration count values
10306 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10309 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10310 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10311 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10312 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10313 value was just used as a "magic string" and not used directly its
10314 value doesn't matter.
10317 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10321 *) "linux-sparc64" configuration (ultrapenguin).
10322 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10323 "linux-sparc" configuration.
10324 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10326 *) config now generates no-xxx options for missing ciphers.
10329 *) Support the EBCDIC character set (work in progress).
10330 File ebcdic.c not yet included because it has a different license.
10331 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10333 *) Support BS2000/OSD-POSIX.
10334 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10336 *) Make callbacks for key generation use void * instead of char *.
10339 *) Make S/MIME samples compile (not yet tested).
10342 *) Additional typesafe stacks.
10345 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10349 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10351 *) New configuration variant "sco5-gcc".
10353 *) Updated some demos.
10354 [Sean O Riordain, Wade Scholine]
10356 *) Add missing BIO_free at exit of pkcs12 application.
10359 *) Fix memory leak in conf.c.
10362 *) Updates for Win32 to assembler version of MD5.
10365 *) Set #! path to perl in apps/der_chop to where we found it
10366 instead of using a fixed path.
10369 *) SHA library changes for irix64-mips4-cc.
10372 *) Improvements for VMS support.
10376 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10378 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10379 This also avoids the problems with SC4.2 and unpatched SC5.
10380 [Andy Polyakov <appro@fy.chalmers.se>]
10382 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10383 These are required because of the typesafe stack would otherwise break
10384 existing code. If old code used a structure member which used to be STACK
10385 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10386 sk_num or sk_value it would produce an error because the num, data members
10387 are not present in STACK_OF. Now it just produces a warning. sk_set
10388 replaces the old method of assigning a value to sk_value
10389 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10390 that does this will no longer work (and should use sk_set instead) but
10391 this could be regarded as a "questionable" behaviour anyway.
10394 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10395 correctly handle encrypted S/MIME data.
10398 *) Change type of various DES function arguments from des_cblock
10399 (which means, in function argument declarations, pointer to char)
10400 to des_cblock * (meaning pointer to array with 8 char elements),
10401 which allows the compiler to do more typechecking; it was like
10402 that back in SSLeay, but with lots of ugly casts.
10404 Introduce new type const_des_cblock.
10407 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10408 problems: find RecipientInfo structure that matches recipient certificate
10409 and initialise the ASN1 structures properly based on passed cipher.
10412 *) Belatedly make the BN tests actually check the results.
10415 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10416 to and from BNs: it was completely broken. New compilation option
10417 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10418 key elements as negative integers.
10421 *) Reorganize and speed up MD5.
10422 [Andy Polyakov <appro@fy.chalmers.se>]
10425 [Richard Levitte <richard@levitte.org>]
10427 *) New option -out to asn1parse to allow the parsed structure to be
10428 output to a file. This is most useful when combined with the -strparse
10429 option to examine the output of things like OCTET STRINGS.
10432 *) Make SSL library a little more fool-proof by not requiring any longer
10433 that SSL_set_{accept,connect}_state be called before
10434 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10435 in many applications because usually everything *appeared* to work as
10436 intended anyway -- now it really works as intended).
10439 *) Move openssl.cnf out of lib/.
10442 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10443 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10444 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10445 [Ralf S. Engelschall]
10447 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10448 handle PKCS#7 enveloped data properly.
10449 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10451 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10452 copying pointers. The cert_st handling is changed by this in
10453 various ways (and thus what used to be known as ctx->default_cert
10454 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10455 any longer when s->cert does not give us what we need).
10456 ssl_cert_instantiate becomes obsolete by this change.
10457 As soon as we've got the new code right (possibly it already is?),
10458 we have solved a couple of bugs of the earlier code where s->cert
10459 was used as if it could not have been shared with other SSL structures.
10461 Note that using the SSL API in certain dirty ways now will result
10462 in different behaviour than observed with earlier library versions:
10463 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10464 does not influence s as it used to.
10466 In order to clean up things more thoroughly, inside SSL_SESSION
10467 we don't use CERT any longer, but a new structure SESS_CERT
10468 that holds per-session data (if available); currently, this is
10469 the peer's certificate chain and, for clients, the server's certificate
10470 and temporary key. CERT holds only those values that can have
10471 meaningful defaults in an SSL_CTX.
10474 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10475 from the internal representation. Various PKCS#7 fixes: remove some
10476 evil casts and set the enc_dig_alg field properly based on the signing
10480 *) Allow PKCS#12 password to be set from the command line or the
10481 environment. Let 'ca' get its config file name from the environment
10482 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10486 *) Allow certificate policies extension to use an IA5STRING for the
10487 organization field. This is contrary to the PKIX definition but
10488 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10492 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10493 without disallowing inline assembler and the like for non-pedantic builds.
10496 *) Support Borland C++ builder.
10497 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10499 *) Support Mingw32.
10502 *) SHA-1 cleanups and performance enhancements.
10503 [Andy Polyakov <appro@fy.chalmers.se>]
10505 *) Sparc v8plus assembler for the bignum library.
10506 [Andy Polyakov <appro@fy.chalmers.se>]
10508 *) Accept any -xxx and +xxx compiler options in Configure.
10511 *) Update HPUX configuration.
10514 *) Add missing sk_<type>_unshift() function to safestack.h
10515 [Ralf S. Engelschall]
10517 *) New function SSL_CTX_use_certificate_chain_file that sets the
10518 "extra_cert"s in addition to the certificate. (This makes sense
10519 only for "PEM" format files, as chains as a whole are not
10523 *) Support verify_depth from the SSL API.
10524 x509_vfy.c had what can be considered an off-by-one-error:
10525 Its depth (which was not part of the external interface)
10526 was actually counting the number of certificates in a chain;
10527 now it really counts the depth.
10530 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10531 instead of X509err, which often resulted in confusing error
10532 messages since the error codes are not globally unique
10533 (e.g. an alleged error in ssl3_accept when a certificate
10534 didn't match the private key).
10536 *) New function SSL_CTX_set_session_id_context that allows to set a default
10537 value (so that you don't need SSL_set_session_id_context for each
10538 connection using the SSL_CTX).
10541 *) OAEP decoding bug fix.
10544 *) Support INSTALL_PREFIX for package builders, as proposed by
10548 *) New Configure options "threads" and "no-threads". For systems
10549 where the proper compiler options are known (currently Solaris
10550 and Linux), "threads" is the default.
10553 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10556 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10557 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10558 such as /usr/local/bin.
10561 *) "make linux-shared" to build shared libraries.
10562 [Niels Poppe <niels@netbox.org>]
10564 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10567 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10568 extension adding in x509 utility.
10571 *) Remove NOPROTO sections and error code comments.
10574 *) Partial rewrite of the DEF file generator to now parse the ANSI
10578 *) New Configure options --prefix=DIR and --openssldir=DIR.
10581 *) Complete rewrite of the error code script(s). It is all now handled
10582 by one script at the top level which handles error code gathering,
10583 header rewriting and C source file generation. It should be much better
10584 than the old method: it now uses a modified version of Ulf's parser to
10585 read the ANSI prototypes in all header files (thus the old K&R definitions
10586 aren't needed for error creation any more) and do a better job of
10587 translating function codes into names. The old 'ASN1 error code imbedded
10588 in a comment' is no longer necessary and it doesn't use .err files which
10589 have now been deleted. Also the error code call doesn't have to appear all
10590 on one line (which resulted in some large lines...).
10593 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10596 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10597 0 (which usually indicates a closed connection), but continue reading.
10600 *) Fix some race conditions.
10603 *) Add support for CRL distribution points extension. Add Certificate
10604 Policies and CRL distribution points documentation.
10607 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10610 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10611 8 of keying material. Merlin has also confirmed interop with this fix
10612 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10613 [Merlin Hughes <merlin@baltimore.ie>]
10615 *) Fix lots of warnings.
10616 [Richard Levitte <levitte@stacken.kth.se>]
10618 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10619 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10620 [Richard Levitte <levitte@stacken.kth.se>]
10622 *) Fix problems with sizeof(long) == 8.
10623 [Andy Polyakov <appro@fy.chalmers.se>]
10625 *) Change functions to ANSI C.
10628 *) Fix typos in error codes.
10629 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10631 *) Remove defunct assembler files from Configure.
10634 *) SPARC v8 assembler BIGNUM implementation.
10635 [Andy Polyakov <appro@fy.chalmers.se>]
10637 *) Support for Certificate Policies extension: both print and set.
10638 Various additions to support the r2i method this uses.
10641 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10642 return a const string when you are expecting an allocated buffer.
10645 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10646 types DirectoryString and DisplayText.
10649 *) Add code to allow r2i extensions to access the configuration database,
10650 add an LHASH database driver and add several ctx helper functions.
10653 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10654 fail when they extended the size of a BIGNUM.
10657 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10658 support typesafe stack.
10661 *) Fix typo in SSL_[gs]et_options().
10662 [Nils Frostberg <nils@medcom.se>]
10664 *) Delete various functions and files that belonged to the (now obsolete)
10665 old X509V3 handling code.
10668 *) New Configure option "rsaref".
10671 *) Don't auto-generate pem.h.
10674 *) Introduce type-safe ASN.1 SETs.
10677 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10678 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10680 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10681 that links with OpenSSL (well at least cause lots of warnings), but fear
10682 not: the conversion is trivial, and it eliminates loads of evil casts. A
10683 few STACKed things have been converted already. Feel free to convert more.
10684 In the fullness of time, I'll do away with the STACK type altogether.
10687 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10688 specified in <certfile> by updating the entry in the index.txt file.
10689 This way one no longer has to edit the index.txt file manually for
10690 revoking a certificate. The -revoke option does the gory details now.
10691 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10693 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10694 `-text' option at all and this way the `-noout -text' combination was
10695 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10696 [Ralf S. Engelschall]
10698 *) Make sure a corresponding plain text error message exists for the
10699 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10700 verify callback function determined that a certificate was revoked.
10701 [Ralf S. Engelschall]
10703 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10704 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10705 all available cipers including rc5, which was forgotten until now.
10706 In order to let the testing shell script know which algorithms
10707 are available, a new (up to now undocumented) command
10708 "openssl list-cipher-commands" is used.
10711 *) Bugfix: s_client occasionally would sleep in select() when
10712 it should have checked SSL_pending() first.
10715 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10716 the raw DSA values prior to ASN.1 encoding.
10719 *) Tweaks to Configure
10720 [Niels Poppe <niels@netbox.org>]
10722 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10726 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10729 *) New config option to avoid instructions that are illegal on the 80386.
10730 The default code is faster, but requires at least a 486.
10733 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10734 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10735 same as SSL2_VERSION anyway.
10738 *) New "-showcerts" option for s_client.
10741 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10742 application. Various cleanups and fixes.
10745 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10746 modify error routines to work internally. Add error codes and PBE init
10747 to library startup routines.
10750 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10751 packing functions to asn1 and evp. Changed function names and error
10752 codes along the way.
10755 *) PKCS12 integration: and so it begins... First of several patches to
10756 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10757 objects to objects.h
10760 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10761 and display support for Thawte strong extranet extension.
10764 *) Add LinuxPPC support.
10765 [Jeff Dubrule <igor@pobox.org>]
10767 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10768 bn_div_words in alpha.s.
10769 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10771 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10772 OAEP isn't supported when OpenSSL is built with RSAref.
10773 [Ulf Moeller <ulf@fitug.de>]
10775 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10776 so they no longer are missing under -DNOPROTO.
10777 [Soren S. Jorvang <soren@t.dk>]
10780 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10782 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10783 doesn't work when the session is reused. Coming soon!
10786 *) Fix a security hole, that allows sessions to be reused in the wrong
10787 context thus bypassing client cert protection! All software that uses
10788 client certs and session caches in multiple contexts NEEDS PATCHING to
10789 allow session reuse! A fuller solution is in the works.
10790 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10792 *) Some more source tree cleanups (removed obsolete files
10793 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10794 permission on "config" script to be executable) and a fix for the INSTALL
10796 [Ulf Moeller <ulf@fitug.de>]
10798 *) Remove some legacy and erroneous uses of malloc, free instead of
10800 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10802 *) Make rsa_oaep_test return non-zero on error.
10803 [Ulf Moeller <ulf@fitug.de>]
10805 *) Add support for native Solaris shared libraries. Configure
10806 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10807 if someone would make that last step automatic.
10808 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10810 *) ctx_size was not built with the right compiler during "make links". Fixed.
10813 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10814 except NULL ciphers". This means the default cipher list will no longer
10815 enable NULL ciphers. They need to be specifically enabled e.g. with
10816 the string "DEFAULT:eNULL".
10819 *) Fix to RSA private encryption routines: if p < q then it would
10820 occasionally produce an invalid result. This will only happen with
10821 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10824 *) Be less restrictive and allow also `perl util/perlpath.pl
10825 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10826 because this way one can also use an interpreter named `perl5' (which is
10827 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10828 installed as `perl').
10829 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10831 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10832 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10834 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10835 advapi32.lib to Win32 build and change the pem test comparision
10836 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10837 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10838 and crypto/des/ede_cbcm_enc.c.
10841 *) DES quad checksum was broken on big-endian architectures. Fixed.
10844 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10845 Win32 test batch file so it (might) work again. The Win32 test batch file
10846 is horrible: I feel ill....
10849 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10850 in e_os.h. Audit of header files to check ANSI and non ANSI
10851 sections: 10 functions were absent from non ANSI section and not exported
10852 from Windows DLLs. Fixed up libeay.num for new functions.
10855 *) Make `openssl version' output lines consistent.
10856 [Ralf S. Engelschall]
10858 *) Fix Win32 symbol export lists for BIO functions: Added
10859 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10860 to ms/libeay{16,32}.def.
10861 [Ralf S. Engelschall]
10863 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10864 fine under Unix and passes some trivial tests I've now added. But the
10865 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10866 added to make sure no one expects that this stuff really works in the
10867 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10868 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10870 [Ralf S. Engelschall]
10872 *) Fix the generation of two part addresses in perl.
10873 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10875 *) Add config entry for Linux on MIPS.
10876 [John Tobey <jtobey@channel1.com>]
10878 *) Make links whenever Configure is run, unless we are on Windoze.
10881 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10882 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10886 *) Add a useful kludge to allow package maintainers to specify compiler and
10887 other platforms details on the command line without having to patch the
10888 Configure script everytime: One now can use ``perl Configure
10889 <id>:<details>'', i.e. platform ids are allowed to have details appended
10890 to them (seperated by colons). This is treated as there would be a static
10891 pre-configured entry in Configure's %table under key <id> with value
10892 <details> and ``perl Configure <id>'' is called. So, when you want to
10893 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10894 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10895 now, which overrides the FreeBSD-elf entry on-the-fly.
10896 [Ralf S. Engelschall]
10898 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10901 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10902 on the `perl Configure ...' command line. This way one can compile
10903 OpenSSL libraries with Position Independent Code (PIC) which is needed
10904 for linking it into DSOs.
10905 [Ralf S. Engelschall]
10907 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10911 *) Cleaned up the LICENSE document: The official contact for any license
10912 questions now is the OpenSSL core team under openssl-core@openssl.org.
10913 And add a paragraph about the dual-license situation to make sure people
10914 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10915 to the OpenSSL toolkit.
10916 [Ralf S. Engelschall]
10918 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10919 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10920 Additonally cleaned up the `make links' target: Remove unnecessary
10921 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10922 to speed processing and no longer clutter the display with confusing
10923 stuff. Instead only the actually done links are displayed.
10924 [Ralf S. Engelschall]
10926 *) Permit null encryption ciphersuites, used for authentication only. It used
10927 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10928 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10932 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10933 signed attributes when verifying signatures (this would break them),
10934 the detached data encoding was wrong and public keys obtained using
10935 X509_get_pubkey() weren't freed.
10938 *) Add text documentation for the BUFFER functions. Also added a work around
10939 to a Win95 console bug. This was triggered by the password read stuff: the
10940 last character typed gets carried over to the next fread(). If you were
10941 generating a new cert request using 'req' for example then the last
10942 character of the passphrase would be CR which would then enter the first
10946 *) Added the new `Includes OpenSSL Cryptography Software' button as
10947 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10948 button and can be used by applications based on OpenSSL to show the
10949 relationship to the OpenSSL project.
10950 [Ralf S. Engelschall]
10952 *) Remove confusing variables in function signatures in files
10953 ssl/ssl_lib.c and ssl/ssl.h.
10954 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10956 *) Don't install bss_file.c under PREFIX/include/
10957 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10959 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10960 functions that return function pointers and has support for NT specific
10961 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10962 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10963 unsigned to signed types: this was killing the Win32 compile.
10966 *) Add new certificate file to stack functions,
10967 SSL_add_dir_cert_subjects_to_stack() and
10968 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10969 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10970 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10971 This means that Apache-SSL and similar packages don't have to mess around
10972 to add as many CAs as they want to the preferred list.
10975 *) Experiment with doxygen documentation. Currently only partially applied to
10977 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10978 openssl.doxy as the configuration file.
10981 *) Get rid of remaining C++-style comments which strict C compilers hate.
10982 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10984 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10985 compiled in by default: it has problems with large keys.
10988 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10989 DH private keys and/or callback functions which directly correspond to
10990 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10991 is needed for applications which have to configure certificates on a
10992 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10994 For the RSA certificate situation is makes no difference, but
10995 for the DSA certificate situation this fixes the "no shared cipher"
10996 problem where the OpenSSL cipher selection procedure failed because the
10997 temporary keys were not overtaken from the context and the API provided
10998 no way to reconfigure them.
10999 The new functions now let applications reconfigure the stuff and they
11000 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
11001 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
11002 non-public-API function ssl_cert_instantiate() is used as a helper
11003 function and also to reduce code redundancy inside ssl_rsa.c.
11004 [Ralf S. Engelschall]
11006 *) Move s_server -dcert and -dkey options out of the undocumented feature
11007 area because they are useful for the DSA situation and should be
11008 recognized by the users.
11009 [Ralf S. Engelschall]
11011 *) Fix the cipher decision scheme for export ciphers: the export bits are
11012 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
11013 SSL_EXP_MASK. So, the original variable has to be used instead of the
11014 already masked variable.
11015 [Richard Levitte <levitte@stacken.kth.se>]
11017 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
11018 [Richard Levitte <levitte@stacken.kth.se>]
11020 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
11021 from `int' to `unsigned int' because it's a length and initialized by
11022 EVP_DigestFinal() which expects an `unsigned int *'.
11023 [Richard Levitte <levitte@stacken.kth.se>]
11025 *) Don't hard-code path to Perl interpreter on shebang line of Configure
11026 script. Instead use the usual Shell->Perl transition trick.
11027 [Ralf S. Engelschall]
11029 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
11030 (in addition to RSA certificates) to match the behaviour of `openssl dsa
11031 -noout -modulus' as it's already the case for `openssl rsa -noout
11032 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
11033 currently the public key is printed (a decision which was already done by
11034 `openssl dsa -modulus' in the past) which serves a similar purpose.
11035 Additionally the NO_RSA no longer completely removes the whole -modulus
11036 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
11038 [Ralf S. Engelschall]
11040 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
11041 BIO. See the source (crypto/evp/bio_ok.c) for more info.
11042 [Arne Ansper <arne@ats.cyber.ee>]
11044 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
11045 to be added. Now both 'req' and 'ca' can use new objects defined in the
11049 *) Add cool BIO that does syslog (or event log on NT).
11050 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
11052 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
11053 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11054 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11055 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11058 *) Add preliminary config info for new extension code.
11061 *) Make RSA_NO_PADDING really use no padding.
11062 [Ulf Moeller <ulf@fitug.de>]
11064 *) Generate errors when private/public key check is done.
11067 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11068 for some CRL extensions and new objects added.
11071 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11072 key usage extension and fuller support for authority key id.
11075 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11076 padding method for RSA, which is recommended for new applications in PKCS
11077 #1 v2.0 (RFC 2437, October 1998).
11078 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11079 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11080 against Bleichbacher's attack on RSA.
11081 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11084 *) Updates to the new SSL compression code
11085 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11087 *) Fix so that the version number in the master secret, when passed
11088 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11089 (because the server will not accept higher), that the version number
11090 is 0x03,0x01, not 0x03,0x00
11091 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11093 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11094 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11095 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11098 *) Support for RAW extensions where an arbitrary extension can be
11099 created by including its DER encoding. See apps/openssl.cnf for
11103 *) Make sure latest Perl versions don't interpret some generated C array
11104 code as Perl array code in the crypto/err/err_genc.pl script.
11105 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11107 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11108 not many people have the assembler. Various Win32 compilation fixes and
11109 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11110 build instructions.
11113 *) Modify configure script 'Configure' to automatically create crypto/date.h
11114 file under Win32 and also build pem.h from pem.org. New script
11115 util/mkfiles.pl to create the MINFO file on environments that can't do a
11116 'make files': perl util/mkfiles.pl >MINFO should work.
11119 *) Major rework of DES function declarations, in the pursuit of correctness
11120 and purity. As a result, many evil casts evaporated, and some weirdness,
11121 too. You may find this causes warnings in your code. Zapping your evil
11122 casts will probably fix them. Mostly.
11125 *) Fix for a typo in asn1.h. Bug fix to object creation script
11126 obj_dat.pl. It considered a zero in an object definition to mean
11127 "end of object": none of the objects in objects.h have any zeros
11128 so it wasn't spotted.
11129 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11131 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11132 Masking (CBCM). In the absence of test vectors, the best I have been able
11133 to do is check that the decrypt undoes the encrypt, so far. Send me test
11134 vectors if you have them.
11137 *) Correct calculation of key length for export ciphers (too much space was
11138 allocated for null ciphers). This has not been tested!
11141 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11142 message is now correct (it understands "crypto" and "ssl" on its
11143 command line). There is also now an "update" option. This will update
11144 the util/ssleay.num and util/libeay.num files with any new functions.
11146 perl util/mkdef.pl crypto ssl update
11147 it will update them.
11150 *) Overhauled the Perl interface (perl/*):
11151 - ported BN stuff to OpenSSL's different BN library
11152 - made the perl/ source tree CVS-aware
11153 - renamed the package from SSLeay to OpenSSL (the files still contain
11154 their history because I've copied them in the repository)
11155 - removed obsolete files (the test scripts will be replaced
11156 by better Test::Harness variants in the future)
11157 [Ralf S. Engelschall]
11159 *) First cut for a very conservative source tree cleanup:
11160 1. merge various obsolete readme texts into doc/ssleay.txt
11161 where we collect the old documents and readme texts.
11162 2. remove the first part of files where I'm already sure that we no
11163 longer need them because of three reasons: either they are just temporary
11164 files which were left by Eric or they are preserved original files where
11165 I've verified that the diff is also available in the CVS via "cvs diff
11166 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11167 the crypto/md/ stuff).
11168 [Ralf S. Engelschall]
11170 *) More extension code. Incomplete support for subject and issuer alt
11171 name, issuer and authority key id. Change the i2v function parameters
11172 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11173 what that's for :-) Fix to ASN1 macro which messed up
11174 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11177 *) Preliminary support for ENUMERATED type. This is largely copied from the
11181 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11182 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11184 *) Make sure `make rehash' target really finds the `openssl' program.
11185 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11187 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11188 like to hear about it if this slows down other processors.
11191 *) Add CygWin32 platform information to Configure script.
11192 [Alan Batie <batie@aahz.jf.intel.com>]
11194 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11195 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11197 *) New program nseq to manipulate netscape certificate sequences
11200 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11204 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11205 but the BN code had some problems that would cause failures when
11206 doing certificate verification and some other functions.
11207 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11209 *) Add ASN1 and PEM code to support netscape certificate sequences.
11212 *) Add ASN1 and PEM code to support netscape certificate sequences.
11215 *) Add several PKIX and private extended key usage OIDs.
11218 *) Modify the 'ca' program to handle the new extension code. Modify
11219 openssl.cnf for new extension format, add comments.
11222 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11223 and add a sample to openssl.cnf so req -x509 now adds appropriate
11227 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11228 error code, add initial support to X509_print() and x509 application.
11231 *) Takes a deep breath and start addding X509 V3 extension support code. Add
11232 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11233 stuff is currently isolated and isn't even compiled yet.
11236 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11237 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11238 Removed the versions check from X509 routines when loading extensions:
11239 this allows certain broken certificates that don't set the version
11240 properly to be processed.
11243 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11244 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11245 can still be regenerated with "make depend".
11248 *) Spelling mistake in C version of CAST-128.
11249 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11251 *) Changes to the error generation code. The perl script err-code.pl
11252 now reads in the old error codes and retains the old numbers, only
11253 adding new ones if necessary. It also only changes the .err files if new
11254 codes are added. The makefiles have been modified to only insert errors
11255 when needed (to avoid needlessly modifying header files). This is done
11256 by only inserting errors if the .err file is newer than the auto generated
11257 C file. To rebuild all the error codes from scratch (the old behaviour)
11258 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11259 or delete all the .err files.
11262 *) CAST-128 was incorrectly implemented for short keys. The C version has
11263 been fixed, but is untested. The assembler versions are also fixed, but
11264 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11265 to regenerate it if needed.
11266 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11267 Hagino <itojun@kame.net>]
11269 *) File was opened incorrectly in randfile.c.
11270 [Ulf Möller <ulf@fitug.de>]
11272 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11273 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11274 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11275 al: it's just almost always a UTCTime. Note this patch adds new error
11276 codes so do a "make errors" if there are problems.
11279 *) Correct Linux 1 recognition in config.
11280 [Ulf Möller <ulf@fitug.de>]
11282 *) Remove pointless MD5 hash when using DSA keys in ca.
11283 [Anonymous <nobody@replay.com>]
11285 *) Generate an error if given an empty string as a cert directory. Also
11286 generate an error if handed NULL (previously returned 0 to indicate an
11287 error, but didn't set one).
11288 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11290 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11293 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11294 parameters. This was causing a warning which killed off the Win32 compile.
11297 *) Remove C++ style comments from crypto/bn/bn_local.h.
11298 [Neil Costigan <neil.costigan@celocom.com>]
11300 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11301 based on a text string, looking up short and long names and finally
11302 "dot" format. The "dot" format stuff didn't work. Added new function
11303 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11304 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11305 OID is not part of the table.
11308 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11309 X509_LOOKUP_by_alias().
11312 *) Sort openssl functions by name.
11315 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11316 encryption from sample DSA keys (in case anyone is interested the password
11320 *) Make _all_ *_free functions accept a NULL pointer.
11321 [Frans Heymans <fheymans@isaserver.be>]
11323 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11325 [Anonymous <nobody@replay.com>]
11327 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11328 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11330 *) Don't blow it for numeric -newkey arguments to apps/req.
11331 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11333 *) Temp key "for export" tests were wrong in s3_srvr.c.
11334 [Anonymous <nobody@replay.com>]
11336 *) Add prototype for temp key callback functions
11337 SSL_CTX_set_tmp_{rsa,dh}_callback().
11340 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11341 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11344 *) X509_name_add_entry() freed the wrong thing after an error.
11345 [Arne Ansper <arne@ats.cyber.ee>]
11347 *) rsa_eay.c would attempt to free a NULL context.
11348 [Arne Ansper <arne@ats.cyber.ee>]
11350 *) BIO_s_socket() had a broken should_retry() on Windoze.
11351 [Arne Ansper <arne@ats.cyber.ee>]
11353 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11354 [Arne Ansper <arne@ats.cyber.ee>]
11356 *) Make sure the already existing X509_STORE->depth variable is initialized
11357 in X509_STORE_new(), but document the fact that this variable is still
11358 unused in the certificate verification process.
11359 [Ralf S. Engelschall]
11361 *) Fix the various library and apps files to free up pkeys obtained from
11362 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11365 *) Fix reference counting in X509_PUBKEY_get(). This makes
11366 demos/maurice/example2.c work, amongst others, probably.
11367 [Steve Henson and Ben Laurie]
11369 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11370 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11371 are no longer created. This way we have a single and consistent command
11372 line interface `openssl <command>', similar to `cvs <command>'.
11373 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11375 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11376 BIT STRING wrapper always have zero unused bits.
11379 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11382 *) Make the top-level INSTALL documentation easier to understand.
11385 *) Makefiles updated to exit if an error occurs in a sub-directory
11386 make (including if user presses ^C) [Paul Sutton]
11388 *) Make Montgomery context stuff explicit in RSA data structure.
11391 *) Fix build order of pem and err to allow for generated pem.h.
11394 *) Fix renumbering bug in X509_NAME_delete_entry().
11397 *) Enhanced the err-ins.pl script so it makes the error library number
11398 global and can add a library name. This is needed for external ASN1 and
11399 other error libraries.
11402 *) Fixed sk_insert which never worked properly.
11405 *) Fix ASN1 macros so they can handle indefinite length construted
11406 EXPLICIT tags. Some non standard certificates use these: they can now
11410 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11411 into a single doc/ssleay.txt bundle. This way the information is still
11412 preserved but no longer messes up this directory. Now it's new room for
11413 the new set of documenation files.
11414 [Ralf S. Engelschall]
11416 *) SETs were incorrectly DER encoded. This was a major pain, because they
11417 shared code with SEQUENCEs, which aren't coded the same. This means that
11418 almost everything to do with SETs or SEQUENCEs has either changed name or
11419 number of arguments.
11420 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11422 *) Fix test data to work with the above.
11425 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11426 was already fixed by Eric for 0.9.1 it seems.
11427 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11429 *) Autodetect FreeBSD3.
11432 *) Fix various bugs in Configure. This affects the following platforms:
11436 unixware-2.0-pentium
11440 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11441 before they are needed.
11444 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11448 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11450 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11451 changed SSLeay to OpenSSL in version strings.
11452 [Ralf S. Engelschall]
11454 *) Some fixups to the top-level documents.
11457 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11458 because the symlink to include/ was missing.
11459 [Ralf S. Engelschall]
11461 *) Incorporated the popular no-RSA/DSA-only patches
11462 which allow to compile a RSA-free SSLeay.
11463 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11465 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11466 when "ssleay" is still not found.
11467 [Ralf S. Engelschall]
11469 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11470 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11472 *) Updated the README file.
11473 [Ralf S. Engelschall]
11475 *) Added various .cvsignore files in the CVS repository subdirs
11476 to make a "cvs update" really silent.
11477 [Ralf S. Engelschall]
11479 *) Recompiled the error-definition header files and added
11480 missing symbols to the Win32 linker tables.
11481 [Ralf S. Engelschall]
11483 *) Cleaned up the top-level documents;
11484 o new files: CHANGES and LICENSE
11485 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11486 o merged COPYRIGHT into LICENSE
11487 o removed obsolete TODO file
11488 o renamed MICROSOFT to INSTALL.W32
11489 [Ralf S. Engelschall]
11491 *) Removed dummy files from the 0.9.1b source tree:
11492 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11493 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11494 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11495 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11496 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11497 [Ralf S. Engelschall]
11499 *) Added various platform portability fixes.
11502 *) The Genesis of the OpenSSL rpject:
11503 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11504 Young and Tim J. Hudson created while they were working for C2Net until
11506 [The OpenSSL Project]
11509 Changes between 0.9.0b and 0.9.1b [not released]
11511 *) Updated a few CA certificates under certs/
11514 *) Changed some BIGNUM api stuff.
11517 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11518 DGUX x86, Linux Alpha, etc.
11521 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11522 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11526 *) Add -strparse option to asn1pars program which parses nested
11528 [Dr Stephen Henson <shenson@bigfoot.com>]
11530 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11533 *) DSA fix for "ca" program.
11536 *) Added "-genkey" option to "dsaparam" program.
11539 *) Added RIPE MD160 (rmd160) message digest.
11542 *) Added -a (all) option to "ssleay version" command.
11545 *) Added PLATFORM define which is the id given to Configure.
11548 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11551 *) Extended the ASN.1 parser routines.
11554 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11557 *) Added a BN_CTX to the BN library.
11560 *) Fixed the weak key values in DES library
11563 *) Changed API in EVP library for cipher aliases.
11566 *) Added support for RC2/64bit cipher.
11569 *) Converted the lhash library to the crypto/mem.c functions.
11572 *) Added more recognized ASN.1 object ids.
11575 *) Added more RSA padding checks for SSL/TLS.
11578 *) Added BIO proxy/filter functionality.
11581 *) Added extra_certs to SSL_CTX which can be used
11582 send extra CA certificates to the client in the CA cert chain sending
11583 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11586 *) Now Fortezza is denied in the authentication phase because
11587 this is key exchange mechanism is not supported by SSLeay at all.
11590 *) Additional PKCS1 checks.
11593 *) Support the string "TLSv1" for all TLS v1 ciphers.
11596 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11597 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11600 *) Fixed a few memory leaks.
11603 *) Fixed various code and comment typos.
11606 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11607 bytes sent in the client random.
11608 [Edward Bishop <ebishop@spyglass.com>]