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]
14 Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
16 *) Constructed ASN.1 types with a recursive definition could exceed the stack
18 Constructed ASN.1 types with a recursive definition (such as can be found
19 in PKCS7) could eventually exceed the stack given malicious input with
20 excessive recursion. This could result in a Denial Of Service attack. There
21 are no such structures used within SSL/TLS that come from untrusted sources
22 so this is considered safe.
24 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
29 Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
31 *) Read/write after SSL object in error state
33 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
34 mechanism. The intent was that if a fatal error occurred during a handshake
35 then OpenSSL would move into the error state and would immediately fail if
36 you attempted to continue the handshake. This works as designed for the
37 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
38 SSL_connect()), however due to a bug it does not work correctly if
39 SSL_read() or SSL_write() is called directly. In that scenario, if the
40 handshake fails then a fatal error will be returned in the initial function
41 call. If SSL_read()/SSL_write() is subsequently called by the application
42 for the same SSL object then it will succeed and the data is passed without
43 being decrypted/encrypted directly from the SSL/TLS record layer.
45 In order to exploit this issue an application bug would have to be present
46 that resulted in a call to SSL_read()/SSL_write() being issued after having
47 already received a fatal error.
49 This issue was reported to OpenSSL by David Benjamin (Google).
53 *) rsaz_1024_mul_avx2 overflow bug on x86_64
55 There is an overflow bug in the AVX2 Montgomery multiplication procedure
56 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
57 Analysis suggests that attacks against RSA and DSA as a result of this
58 defect would be very difficult to perform and are not believed likely.
59 Attacks against DH1024 are considered just feasible, because most of the
60 work necessary to deduce information about a private key may be performed
61 offline. The amount of resources required for such an attack would be
62 significant. However, for an attack on TLS to be meaningful, the server
63 would have to share the DH1024 private key among multiple clients, which is
64 no longer an option since CVE-2016-0701.
66 This only affects processors that support the AVX2 but not ADX extensions
67 like Intel Haswell (4th generation).
69 This issue was reported to OpenSSL by David Benjamin (Google). The issue
70 was originally found via the OSS-Fuzz project.
74 Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
76 *) bn_sqrx8x_internal carry bug on x86_64
78 There is a carry propagating bug in the x86_64 Montgomery squaring
79 procedure. No EC algorithms are affected. Analysis suggests that attacks
80 against RSA and DSA as a result of this defect would be very difficult to
81 perform and are not believed likely. Attacks against DH are considered just
82 feasible (although very difficult) because most of the work necessary to
83 deduce information about a private key may be performed offline. The amount
84 of resources required for such an attack would be very significant and
85 likely only accessible to a limited number of attackers. An attacker would
86 additionally need online access to an unpatched system using the target
87 private key in a scenario with persistent DH parameters and a private
88 key that is shared between multiple clients.
90 This only affects processors that support the BMI1, BMI2 and ADX extensions
91 like Intel Broadwell (5th generation) and later or AMD Ryzen.
93 This issue was reported to OpenSSL by the OSS-Fuzz project.
97 *) Malformed X.509 IPAddressFamily could cause OOB read
99 If an X.509 certificate has a malformed IPAddressFamily extension,
100 OpenSSL could do a one-byte buffer overread. The most likely result
101 would be an erroneous display of the certificate in text format.
103 This issue was reported to OpenSSL by the OSS-Fuzz project.
107 Changes between 1.0.2k and 1.0.2l [25 May 2017]
109 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
110 platform rather than 'mingw'.
113 Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
115 *) Truncated packet could crash via OOB read
117 If one side of an SSL/TLS path is running on a 32-bit host and a specific
118 cipher is being used, then a truncated packet can cause that host to
119 perform an out-of-bounds read, usually resulting in a crash.
121 This issue was reported to OpenSSL by Robert Święcki of Google.
125 *) BN_mod_exp may produce incorrect results on x86_64
127 There is a carry propagating bug in the x86_64 Montgomery squaring
128 procedure. No EC algorithms are affected. Analysis suggests that attacks
129 against RSA and DSA as a result of this defect would be very difficult to
130 perform and are not believed likely. Attacks against DH are considered just
131 feasible (although very difficult) because most of the work necessary to
132 deduce information about a private key may be performed offline. The amount
133 of resources required for such an attack would be very significant and
134 likely only accessible to a limited number of attackers. An attacker would
135 additionally need online access to an unpatched system using the target
136 private key in a scenario with persistent DH parameters and a private
137 key that is shared between multiple clients. For example this can occur by
138 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
139 similar to CVE-2015-3193 but must be treated as a separate problem.
141 This issue was reported to OpenSSL by the OSS-Fuzz project.
145 *) Montgomery multiplication may produce incorrect results
147 There is a carry propagating bug in the Broadwell-specific Montgomery
148 multiplication procedure that handles input lengths divisible by, but
149 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
150 and DH private keys are impossible. This is because the subroutine in
151 question is not used in operations with the private key itself and an input
152 of the attacker's direct choice. Otherwise the bug can manifest itself as
153 transient authentication and key negotiation failures or reproducible
154 erroneous outcome of public-key operations with specially crafted input.
155 Among EC algorithms only Brainpool P-512 curves are affected and one
156 presumably can attack ECDH key negotiation. Impact was not analyzed in
157 detail, because pre-requisites for attack are considered unlikely. Namely
158 multiple clients have to choose the curve in question and the server has to
159 share the private key among them, neither of which is default behaviour.
160 Even then only clients that chose the curve will be affected.
162 This issue was publicly reported as transient failures and was not
163 initially recognized as a security issue. Thanks to Richard Morgan for
164 providing reproducible case.
168 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
169 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
170 prevent issues where no progress is being made and the peer continually
171 sends unrecognised record types, using up resources processing them.
174 Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
176 *) Missing CRL sanity check
178 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
179 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
180 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
182 This issue only affects the OpenSSL 1.0.2i
186 Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
188 *) OCSP Status Request extension unbounded memory growth
190 A malicious client can send an excessively large OCSP Status Request
191 extension. If that client continually requests renegotiation, sending a
192 large OCSP Status Request extension each time, then there will be unbounded
193 memory growth on the server. This will eventually lead to a Denial Of
194 Service attack through memory exhaustion. Servers with a default
195 configuration are vulnerable even if they do not support OCSP. Builds using
196 the "no-ocsp" build time option are not affected.
198 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
202 *) In order to mitigate the SWEET32 attack, the DES ciphers were moved from
205 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
210 *) OOB write in MDC2_Update()
212 An overflow can occur in MDC2_Update() either if called directly or
213 through the EVP_DigestUpdate() function using MDC2. If an attacker
214 is able to supply very large amounts of input data after a previous
215 call to EVP_EncryptUpdate() with a partial block then a length check
216 can overflow resulting in a heap corruption.
218 The amount of data needed is comparable to SIZE_MAX which is impractical
221 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
225 *) Malformed SHA512 ticket DoS
227 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
228 DoS attack where a malformed ticket will result in an OOB read which will
231 The use of SHA512 in TLS session tickets is comparatively rare as it requires
232 a custom server callback and ticket lookup mechanism.
234 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
238 *) OOB write in BN_bn2dec()
240 The function BN_bn2dec() does not check the return value of BN_div_word().
241 This can cause an OOB write if an application uses this function with an
242 overly large BIGNUM. This could be a problem if an overly large certificate
243 or CRL is printed out from an untrusted source. TLS is not affected because
244 record limits will reject an oversized certificate before it is parsed.
246 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
250 *) OOB read in TS_OBJ_print_bio()
252 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
253 the total length the OID text representation would use and not the amount
254 of data written. This will result in OOB reads when large OIDs are
257 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
261 *) Pointer arithmetic undefined behaviour
263 Avoid some undefined pointer arithmetic
265 A common idiom in the codebase is to check limits in the following manner:
268 Where "p" points to some malloc'd data of SIZE bytes and
271 "len" here could be from some externally supplied data (e.g. from a TLS
274 The rules of C pointer arithmetic are such that "p + len" is only well
275 defined where len <= SIZE. Therefore the above idiom is actually
278 For example this could cause problems if some malloc implementation
279 provides an address for "p" such that "p + len" actually overflows for
280 values of len that are too big and therefore p + len < limit.
282 This issue was reported to OpenSSL by Guido Vranken
286 *) Constant time flag not preserved in DSA signing
288 Operations in the DSA signing algorithm should run in constant time in
289 order to avoid side channel attacks. A flaw in the OpenSSL DSA
290 implementation means that a non-constant time codepath is followed for
291 certain operations. This has been demonstrated through a cache-timing
292 attack to be sufficient for an attacker to recover the private DSA key.
294 This issue was reported by César Pereida (Aalto University), Billy Brumley
295 (Tampere University of Technology), and Yuval Yarom (The University of
300 *) DTLS buffered message DoS
302 In a DTLS connection where handshake messages are delivered out-of-order
303 those messages that OpenSSL is not yet ready to process will be buffered
304 for later use. Under certain circumstances, a flaw in the logic means that
305 those messages do not get removed from the buffer even though the handshake
306 has been completed. An attacker could force up to approx. 15 messages to
307 remain in the buffer when they are no longer required. These messages will
308 be cleared when the DTLS connection is closed. The default maximum size for
309 a message is 100k. Therefore the attacker could force an additional 1500k
310 to be consumed per connection. By opening many simulataneous connections an
311 attacker could cause a DoS attack through memory exhaustion.
313 This issue was reported to OpenSSL by Quan Luo.
317 *) DTLS replay protection DoS
319 A flaw in the DTLS replay attack protection mechanism means that records
320 that arrive for future epochs update the replay protection "window" before
321 the MAC for the record has been validated. This could be exploited by an
322 attacker by sending a record for the next epoch (which does not have to
323 decrypt or have a valid MAC), with a very large sequence number. This means
324 that all subsequent legitimate packets are dropped causing a denial of
325 service for a specific DTLS connection.
327 This issue was reported to OpenSSL by the OCAP audit team.
331 *) Certificate message OOB reads
333 In OpenSSL 1.0.2 and earlier some missing message length checks can result
334 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
335 theoretical DoS risk but this has not been observed in practice on common
338 The messages affected are client certificate, client certificate request
339 and server certificate. As a result the attack can only be performed
340 against a client or a server which enables client authentication.
342 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
346 Changes between 1.0.2g and 1.0.2h [3 May 2016]
348 *) Prevent padding oracle in AES-NI CBC MAC check
350 A MITM attacker can use a padding oracle attack to decrypt traffic
351 when the connection uses an AES CBC cipher and the server support
354 This issue was introduced as part of the fix for Lucky 13 padding
355 attack (CVE-2013-0169). The padding check was rewritten to be in
356 constant time by making sure that always the same bytes are read and
357 compared against either the MAC or padding bytes. But it no longer
358 checked that there was enough data to have both the MAC and padding
361 This issue was reported by Juraj Somorovsky using TLS-Attacker.
365 *) Fix EVP_EncodeUpdate overflow
367 An overflow can occur in the EVP_EncodeUpdate() function which is used for
368 Base64 encoding of binary data. If an attacker is able to supply very large
369 amounts of input data then a length check can overflow resulting in a heap
372 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
373 the PEM_write_bio* family of functions. These are mainly used within the
374 OpenSSL command line applications, so any application which processes data
375 from an untrusted source and outputs it as a PEM file should be considered
376 vulnerable to this issue. User applications that call these APIs directly
377 with large amounts of untrusted data may also be vulnerable.
379 This issue was reported by Guido Vranken.
383 *) Fix EVP_EncryptUpdate overflow
385 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
386 is able to supply very large amounts of input data after a previous call to
387 EVP_EncryptUpdate() with a partial block then a length check can overflow
388 resulting in a heap corruption. Following an analysis of all OpenSSL
389 internal usage of the EVP_EncryptUpdate() function all usage is one of two
390 forms. The first form is where the EVP_EncryptUpdate() call is known to be
391 the first called function after an EVP_EncryptInit(), and therefore that
392 specific call must be safe. The second form is where the length passed to
393 EVP_EncryptUpdate() can be seen from the code to be some small value and
394 therefore there is no possibility of an overflow. Since all instances are
395 one of these two forms, it is believed that there can be no overflows in
396 internal code due to this problem. It should be noted that
397 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
398 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
399 of these calls have also been analysed too and it is believed there are no
400 instances in internal usage where an overflow could occur.
402 This issue was reported by Guido Vranken.
406 *) Prevent ASN.1 BIO excessive memory allocation
408 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
409 a short invalid encoding can casuse allocation of large amounts of memory
410 potentially consuming excessive resources or exhausting memory.
412 Any application parsing untrusted data through d2i BIO functions is
413 affected. The memory based functions such as d2i_X509() are *not* affected.
414 Since the memory based functions are used by the TLS library, TLS
415 applications are not affected.
417 This issue was reported by Brian Carpenter.
423 ASN1 Strings that are over 1024 bytes can cause an overread in applications
424 using the X509_NAME_oneline() function on EBCDIC systems. This could result
425 in arbitrary stack data being returned in the buffer.
427 This issue was reported by Guido Vranken.
431 *) Modify behavior of ALPN to invoke callback after SNI/servername
432 callback, such that updates to the SSL_CTX affect ALPN.
435 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
439 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
440 methods are enabled and ssl2 is disabled the methods return NULL.
443 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
445 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
446 Builds that are not configured with "enable-weak-ssl-ciphers" will not
447 provide any "EXPORT" or "LOW" strength ciphers.
450 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
451 is by default disabled at build-time. Builds that are not configured with
452 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
453 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
454 will need to explicitly call either of:
456 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
458 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
460 as appropriate. Even if either of those is used, or the application
461 explicitly uses the version-specific SSLv2_method() or its client and
462 server variants, SSLv2 ciphers vulnerable to exhaustive search key
463 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
464 ciphers, and SSLv2 56-bit DES are no longer available.
468 *) Fix a double-free in DSA code
470 A double free bug was discovered when OpenSSL parses malformed DSA private
471 keys and could lead to a DoS attack or memory corruption for applications
472 that receive DSA private keys from untrusted sources. This scenario is
475 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
480 *) Disable SRP fake user seed to address a server memory leak.
482 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
484 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
485 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
486 was changed to ignore the "fake user" SRP seed, even if the seed
489 Users should use SRP_VBASE_get1_by_user instead. Note that in
490 SRP_VBASE_get1_by_user, caller must free the returned value. Note
491 also that even though configuring the SRP seed attempts to hide
492 invalid usernames by continuing the handshake with fake
493 credentials, this behaviour is not constant time and no strong
494 guarantees are made that the handshake is indistinguishable from
495 that of a valid user.
499 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
501 In the BN_hex2bn function the number of hex digits is calculated using an
502 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
503 large values of |i| this can result in |bn_expand| not allocating any
504 memory because |i * 4| is negative. This can leave the internal BIGNUM data
505 field as NULL leading to a subsequent NULL ptr deref. For very large values
506 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
507 In this case memory is allocated to the internal BIGNUM data field, but it
508 is insufficiently sized leading to heap corruption. A similar issue exists
509 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
510 is ever called by user applications with very large untrusted hex/dec data.
511 This is anticipated to be a rare occurrence.
513 All OpenSSL internal usage of these functions use data that is not expected
514 to be untrusted, e.g. config file data or application command line
515 arguments. If user developed applications generate config file data based
516 on untrusted data then it is possible that this could also lead to security
517 consequences. This is also anticipated to be rare.
519 This issue was reported to OpenSSL by Guido Vranken.
523 *) Fix memory issues in BIO_*printf functions
525 The internal |fmtstr| function used in processing a "%s" format string in
526 the BIO_*printf functions could overflow while calculating the length of a
527 string and cause an OOB read when printing very long strings.
529 Additionally the internal |doapr_outch| function can attempt to write to an
530 OOB memory location (at an offset from the NULL pointer) in the event of a
531 memory allocation failure. In 1.0.2 and below this could be caused where
532 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
533 could be in processing a very long "%s" format string. Memory leaks can
536 The first issue may mask the second issue dependent on compiler behaviour.
537 These problems could enable attacks where large amounts of untrusted data
538 is passed to the BIO_*printf functions. If applications use these functions
539 in this way then they could be vulnerable. OpenSSL itself uses these
540 functions when printing out human-readable dumps of ASN.1 data. Therefore
541 applications that print this data could be vulnerable if the data is from
542 untrusted sources. OpenSSL command line applications could also be
543 vulnerable where they print out ASN.1 data, or if untrusted data is passed
544 as command line arguments.
546 Libssl is not considered directly vulnerable. Additionally certificates etc
547 received via remote connections via libssl are also unlikely to be able to
548 trigger these issues because of message size limits enforced within libssl.
550 This issue was reported to OpenSSL Guido Vranken.
554 *) Side channel attack on modular exponentiation
556 A side-channel attack was found which makes use of cache-bank conflicts on
557 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
558 of RSA keys. The ability to exploit this issue is limited as it relies on
559 an attacker who has control of code in a thread running on the same
560 hyper-threaded core as the victim thread which is performing decryptions.
562 This issue was reported to OpenSSL by Yuval Yarom, The University of
563 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
564 Nadia Heninger, University of Pennsylvania with more information at
565 http://cachebleed.info.
569 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
570 if no keysize is specified with default_bits. This fixes an
571 omission in an earlier change that changed all RSA/DSA key generation
572 apps to use 2048 bits by default.
575 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
577 *) DH small subgroups
579 Historically OpenSSL only ever generated DH parameters based on "safe"
580 primes. More recently (in version 1.0.2) support was provided for
581 generating X9.42 style parameter files such as those required for RFC 5114
582 support. The primes used in such files may not be "safe". Where an
583 application is using DH configured with parameters based on primes that are
584 not "safe" then an attacker could use this fact to find a peer's private
585 DH exponent. This attack requires that the attacker complete multiple
586 handshakes in which the peer uses the same private DH exponent. For example
587 this could be used to discover a TLS server's private DH exponent if it's
588 reusing the private DH exponent or it's using a static DH ciphersuite.
590 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
591 TLS. It is not on by default. If the option is not set then the server
592 reuses the same private DH exponent for the life of the server process and
593 would be vulnerable to this attack. It is believed that many popular
594 applications do set this option and would therefore not be at risk.
596 The fix for this issue adds an additional check where a "q" parameter is
597 available (as is the case in X9.42 based parameters). This detects the
598 only known attack, and is the only possible defense for static DH
599 ciphersuites. This could have some performance impact.
601 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
602 default and cannot be disabled. This could have some performance impact.
604 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
608 *) SSLv2 doesn't block disabled ciphers
610 A malicious client can negotiate SSLv2 ciphers that have been disabled on
611 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
612 been disabled, provided that the SSLv2 protocol was not also disabled via
615 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
616 and Sebastian Schinzel.
620 *) Reject DH handshakes with parameters shorter than 1024 bits.
623 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
625 *) BN_mod_exp may produce incorrect results on x86_64
627 There is a carry propagating bug in the x86_64 Montgomery squaring
628 procedure. No EC algorithms are affected. Analysis suggests that attacks
629 against RSA and DSA as a result of this defect would be very difficult to
630 perform and are not believed likely. Attacks against DH are considered just
631 feasible (although very difficult) because most of the work necessary to
632 deduce information about a private key may be performed offline. The amount
633 of resources required for such an attack would be very significant and
634 likely only accessible to a limited number of attackers. An attacker would
635 additionally need online access to an unpatched system using the target
636 private key in a scenario with persistent DH parameters and a private
637 key that is shared between multiple clients. For example this can occur by
638 default in OpenSSL DHE based SSL/TLS ciphersuites.
640 This issue was reported to OpenSSL by Hanno Böck.
644 *) Certificate verify crash with missing PSS parameter
646 The signature verification routines will crash with a NULL pointer
647 dereference if presented with an ASN.1 signature using the RSA PSS
648 algorithm and absent mask generation function parameter. Since these
649 routines are used to verify certificate signature algorithms this can be
650 used to crash any certificate verification operation and exploited in a
651 DoS attack. Any application which performs certificate verification is
652 vulnerable including OpenSSL clients and servers which enable client
655 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
659 *) X509_ATTRIBUTE memory leak
661 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
662 memory. This structure is used by the PKCS#7 and CMS routines so any
663 application which reads PKCS#7 or CMS data from untrusted sources is
664 affected. SSL/TLS is not affected.
666 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
671 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
672 This changes the decoding behaviour for some invalid messages,
673 though the change is mostly in the more lenient direction, and
674 legacy behaviour is preserved as much as possible.
677 *) In DSA_generate_parameters_ex, if the provided seed is too short,
678 use a random seed, as already documented.
679 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
681 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
683 *) Alternate chains certificate forgery
685 During certificate verfification, OpenSSL will attempt to find an
686 alternative certificate chain if the first attempt to build such a chain
687 fails. An error in the implementation of this logic can mean that an
688 attacker could cause certain checks on untrusted certificates to be
689 bypassed, such as the CA flag, enabling them to use a valid leaf
690 certificate to act as a CA and "issue" an invalid certificate.
692 This issue was reported to OpenSSL by Adam Langley/David Benjamin
697 *) Race condition handling PSK identify hint
699 If PSK identity hints are received by a multi-threaded client then
700 the values are wrongly updated in the parent SSL_CTX structure. This can
701 result in a race condition potentially leading to a double free of the
706 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
708 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
709 incompatibility in the handling of HMAC. The previous ABI has now been
712 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
714 *) Malformed ECParameters causes infinite loop
716 When processing an ECParameters structure OpenSSL enters an infinite loop
717 if the curve specified is over a specially malformed binary polynomial
720 This can be used to perform denial of service against any
721 system which processes public keys, certificate requests or
722 certificates. This includes TLS clients and TLS servers with
723 client authentication enabled.
725 This issue was reported to OpenSSL by Joseph Barr-Pixton.
729 *) Exploitable out-of-bounds read in X509_cmp_time
731 X509_cmp_time does not properly check the length of the ASN1_TIME
732 string and can read a few bytes out of bounds. In addition,
733 X509_cmp_time accepts an arbitrary number of fractional seconds in the
736 An attacker can use this to craft malformed certificates and CRLs of
737 various sizes and potentially cause a segmentation fault, resulting in
738 a DoS on applications that verify certificates or CRLs. TLS clients
739 that verify CRLs are affected. TLS clients and servers with client
740 authentication enabled may be affected if they use custom verification
743 This issue was reported to OpenSSL by Robert Swiecki (Google), and
744 independently by Hanno Böck.
748 *) PKCS7 crash with missing EnvelopedContent
750 The PKCS#7 parsing code does not handle missing inner EncryptedContent
751 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
752 with missing content and trigger a NULL pointer dereference on parsing.
754 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
755 structures from untrusted sources are affected. OpenSSL clients and
756 servers are not affected.
758 This issue was reported to OpenSSL by Michal Zalewski (Google).
762 *) CMS verify infinite loop with unknown hash function
764 When verifying a signedData message the CMS code can enter an infinite loop
765 if presented with an unknown hash function OID. This can be used to perform
766 denial of service against any system which verifies signedData messages using
768 This issue was reported to OpenSSL by Johannes Bauer.
772 *) Race condition handling NewSessionTicket
774 If a NewSessionTicket is received by a multi-threaded client when attempting to
775 reuse a previous ticket then a race condition can occur potentially leading to
776 a double free of the ticket data.
780 *) Removed support for the two export grade static DH ciphersuites
781 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
782 were newly added (along with a number of other static DH ciphersuites) to
783 1.0.2. However the two export ones have *never* worked since they were
784 introduced. It seems strange in any case to be adding new export
785 ciphersuites, and given "logjam" it also does not seem correct to fix them.
788 *) Only support 256-bit or stronger elliptic curves with the
789 'ecdh_auto' setting (server) or by default (client). Of supported
790 curves, prefer P-256 (both).
793 *) Reject DH handshakes with parameters shorter than 768 bits.
794 [Kurt Roeckx and Emilia Kasper]
796 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
798 *) ClientHello sigalgs DoS fix
800 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
801 invalid signature algorithms extension a NULL pointer dereference will
802 occur. This can be exploited in a DoS attack against the server.
804 This issue was was reported to OpenSSL by David Ramos of Stanford
807 [Stephen Henson and Matt Caswell]
809 *) Multiblock corrupted pointer fix
811 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
812 feature only applies on 64 bit x86 architecture platforms that support AES
813 NI instructions. A defect in the implementation of "multiblock" can cause
814 OpenSSL's internal write buffer to become incorrectly set to NULL when
815 using non-blocking IO. Typically, when the user application is using a
816 socket BIO for writing, this will only result in a failed connection.
817 However if some other BIO is used then it is likely that a segmentation
818 fault will be triggered, thus enabling a potential DoS attack.
820 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
824 *) Segmentation fault in DTLSv1_listen fix
826 The DTLSv1_listen function is intended to be stateless and processes the
827 initial ClientHello from many peers. It is common for user code to loop
828 over the call to DTLSv1_listen until a valid ClientHello is received with
829 an associated cookie. A defect in the implementation of DTLSv1_listen means
830 that state is preserved in the SSL object from one invocation to the next
831 that can lead to a segmentation fault. Errors processing the initial
832 ClientHello can trigger this scenario. An example of such an error could be
833 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
836 This issue was reported to OpenSSL by Per Allansson.
840 *) Segmentation fault in ASN1_TYPE_cmp fix
842 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
843 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
844 certificate signature algorithm consistency this can be used to crash any
845 certificate verification operation and exploited in a DoS attack. Any
846 application which performs certificate verification is vulnerable including
847 OpenSSL clients and servers which enable client authentication.
851 *) Segmentation fault for invalid PSS parameters fix
853 The signature verification routines will crash with a NULL pointer
854 dereference if presented with an ASN.1 signature using the RSA PSS
855 algorithm and invalid parameters. Since these routines are used to verify
856 certificate signature algorithms this can be used to crash any
857 certificate verification operation and exploited in a DoS attack. Any
858 application which performs certificate verification is vulnerable including
859 OpenSSL clients and servers which enable client authentication.
861 This issue was was reported to OpenSSL by Brian Carpenter.
865 *) ASN.1 structure reuse memory corruption fix
867 Reusing a structure in ASN.1 parsing may allow an attacker to cause
868 memory corruption via an invalid write. Such reuse is and has been
869 strongly discouraged and is believed to be rare.
871 Applications that parse structures containing CHOICE or ANY DEFINED BY
872 components may be affected. Certificate parsing (d2i_X509 and related
873 functions) are however not affected. OpenSSL clients and servers are
878 *) PKCS7 NULL pointer dereferences fix
880 The PKCS#7 parsing code does not handle missing outer ContentInfo
881 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
882 missing content and trigger a NULL pointer dereference on parsing.
884 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
885 otherwise parse PKCS#7 structures from untrusted sources are
886 affected. OpenSSL clients and servers are not affected.
888 This issue was reported to OpenSSL by Michal Zalewski (Google).
892 *) DoS via reachable assert in SSLv2 servers fix
894 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
895 servers that both support SSLv2 and enable export cipher suites by sending
896 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
898 This issue was discovered by Sean Burford (Google) and Emilia Käsper
899 (OpenSSL development team).
903 *) Empty CKE with client auth and DHE fix
905 If client auth is used then a server can seg fault in the event of a DHE
906 ciphersuite being selected and a zero length ClientKeyExchange message
907 being sent by the client. This could be exploited in a DoS attack.
911 *) Handshake with unseeded PRNG fix
913 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
914 with an unseeded PRNG. The conditions are:
915 - The client is on a platform where the PRNG has not been seeded
916 automatically, and the user has not seeded manually
917 - A protocol specific client method version has been used (i.e. not
918 SSL_client_methodv23)
919 - A ciphersuite is used that does not require additional random data from
920 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
922 If the handshake succeeds then the client random that has been used will
923 have been generated from a PRNG with insufficient entropy and therefore the
924 output may be predictable.
926 For example using the following command with an unseeded openssl will
927 succeed on an unpatched platform:
929 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
933 *) Use After Free following d2i_ECPrivatekey error fix
935 A malformed EC private key file consumed via the d2i_ECPrivateKey function
936 could cause a use after free condition. This, in turn, could cause a double
937 free in several private key parsing functions (such as d2i_PrivateKey
938 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
939 for applications that receive EC private keys from untrusted
940 sources. This scenario is considered rare.
942 This issue was discovered by the BoringSSL project and fixed in their
947 *) X509_to_X509_REQ NULL pointer deref fix
949 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
950 the certificate key is invalid. This function is rarely used in practice.
952 This issue was discovered by Brian Carpenter.
956 *) Removed the export ciphers from the DEFAULT ciphers
959 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
961 *) Change RSA and DH/DSA key generation apps to generate 2048-bit
965 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
966 ARMv5 through ARMv8, as opposite to "locking" it to single one.
967 So far those who have to target multiple plaforms would compromise
968 and argue that binary targeting say ARMv5 would still execute on
969 ARMv8. "Universal" build resolves this compromise by providing
970 near-optimal performance even on newer platforms.
973 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
974 (other platforms pending).
975 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
977 *) Add support for the SignedCertificateTimestampList certificate and
978 OCSP response extensions from RFC6962.
981 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
982 for corner cases. (Certain input points at infinity could lead to
983 bogus results, with non-infinity inputs mapped to infinity too.)
986 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
987 This covers AES, SHA256/512 and GHASH. "Initial" means that most
988 common cases are optimized and there still is room for further
989 improvements. Vector Permutation AES for Altivec is also added.
992 *) Add support for little-endian ppc64 Linux target.
993 [Marcelo Cerri (IBM)]
995 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
996 SHA1, SHA256 and GHASH. "Initial" means that most common cases
997 are optimized and there still is room for further improvements.
998 Both 32- and 64-bit modes are supported.
999 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1001 *) Improved ARMv7 NEON support.
1004 *) Support for SPARC Architecture 2011 crypto extensions, first
1005 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1006 SHA256/512, MD5, GHASH and modular exponentiation.
1007 [Andy Polyakov, David Miller]
1009 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1011 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1013 *) Support for new and upcoming Intel processors, including AVX2,
1014 BMI and SHA ISA extensions. This includes additional "stitched"
1015 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1018 This work was sponsored by Intel Corp.
1021 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1022 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1023 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1026 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1027 this fixes a limiation in previous versions of OpenSSL.
1030 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1031 MGF1 digest and OAEP label.
1034 *) Add EVP support for key wrapping algorithms, to avoid problems with
1035 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1036 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1037 algorithms and include tests cases.
1040 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1042 [Douglas E. Engert, Steve Henson]
1044 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1045 difference in days and seconds between two tm or ASN1_TIME structures.
1048 *) Add -rev test option to s_server to just reverse order of characters
1049 received by client and send back to server. Also prints an abbreviated
1050 summary of the connection parameters.
1053 *) New option -brief for s_client and s_server to print out a brief summary
1054 of connection parameters.
1057 *) Add callbacks for arbitrary TLS extensions.
1058 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1060 *) New option -crl_download in several openssl utilities to download CRLs
1061 from CRLDP extension in certificates.
1064 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1067 *) New function X509_CRL_diff to generate a delta CRL from the difference
1068 of two full CRLs. Add support to "crl" utility.
1071 *) New functions to set lookup_crls function and to retrieve
1072 X509_STORE from X509_STORE_CTX.
1075 *) Print out deprecated issuer and subject unique ID fields in
1079 *) Extend OCSP I/O functions so they can be used for simple general purpose
1080 HTTP as well as OCSP. New wrapper function which can be used to download
1081 CRLs using the OCSP API.
1084 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1087 *) SSL_CONF* functions. These provide a common framework for application
1088 configuration using configuration files or command lines.
1091 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1092 message callback and prints the results. Needs compile time option
1093 "enable-ssl-trace". New options to s_client and s_server to enable
1097 *) New ctrl and macro to retrieve supported points extensions.
1098 Print out extension in s_server and s_client.
1101 *) New functions to retrieve certificate signature and signature
1105 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1109 *) New Suite B modes for TLS code. These use and enforce the requirements
1110 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1111 only use Suite B curves. The Suite B modes can be set by using the
1112 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1115 *) New chain verification flags for Suite B levels of security. Check
1116 algorithms are acceptable when flags are set in X509_verify_cert.
1119 *) Make tls1_check_chain return a set of flags indicating checks passed
1120 by a certificate chain. Add additional tests to handle client
1121 certificates: checks for matching certificate type and issuer name
1125 *) If an attempt is made to use a signature algorithm not in the peer
1126 preference list abort the handshake. If client has no suitable
1127 signature algorithms in response to a certificate request do not
1128 use the certificate.
1131 *) If server EC tmp key is not in client preference list abort handshake.
1134 *) Add support for certificate stores in CERT structure. This makes it
1135 possible to have different stores per SSL structure or one store in
1136 the parent SSL_CTX. Include distint stores for certificate chain
1137 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1138 to build and store a certificate chain in CERT structure: returing
1139 an error if the chain cannot be built: this will allow applications
1140 to test if a chain is correctly configured.
1142 Note: if the CERT based stores are not set then the parent SSL_CTX
1143 store is used to retain compatibility with existing behaviour.
1147 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1148 mask based on the current session, check mask when sending client
1149 hello and checking the requested ciphersuite.
1152 *) New ctrls to retrieve and set certificate types in a certificate
1153 request message. Print out received values in s_client. If certificate
1154 types is not set with custom values set sensible values based on
1155 supported signature algorithms.
1158 *) Support for distinct client and server supported signature algorithms.
1161 *) Add certificate callback. If set this is called whenever a certificate
1162 is required by client or server. An application can decide which
1163 certificate chain to present based on arbitrary criteria: for example
1164 supported signature algorithms. Add very simple example to s_server.
1165 This fixes many of the problems and restrictions of the existing client
1166 certificate callback: for example you can now clear an existing
1167 certificate and specify the whole chain.
1170 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1171 the certificate can be used for (if anything). Set valid_flags field
1172 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1173 to have similar checks in it.
1175 Add new "cert_flags" field to CERT structure and include a "strict mode".
1176 This enforces some TLS certificate requirements (such as only permitting
1177 certificate signature algorithms contained in the supported algorithms
1178 extension) which some implementations ignore: this option should be used
1179 with caution as it could cause interoperability issues.
1182 *) Update and tidy signature algorithm extension processing. Work out
1183 shared signature algorithms based on preferences and peer algorithms
1184 and print them out in s_client and s_server. Abort handshake if no
1185 shared signature algorithms.
1188 *) Add new functions to allow customised supported signature algorithms
1189 for SSL and SSL_CTX structures. Add options to s_client and s_server
1193 *) New function SSL_certs_clear() to delete all references to certificates
1194 from an SSL structure. Before this once a certificate had been added
1195 it couldn't be removed.
1198 *) Integrate hostname, email address and IP address checking with certificate
1199 verification. New verify options supporting checking in opensl utility.
1202 *) Fixes and wildcard matching support to hostname and email checking
1203 functions. Add manual page.
1204 [Florian Weimer (Red Hat Product Security Team)]
1206 *) New functions to check a hostname email or IP address against a
1207 certificate. Add options x509 utility to print results of checks against
1211 *) Fix OCSP checking.
1212 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1214 *) Initial experimental support for explicitly trusted non-root CAs.
1215 OpenSSL still tries to build a complete chain to a root but if an
1216 intermediate CA has a trust setting included that is used. The first
1217 setting is used: whether to trust (e.g., -addtrust option to the x509
1221 *) Add -trusted_first option which attempts to find certificates in the
1222 trusted store even if an untrusted chain is also supplied.
1225 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1226 platform support for Linux and Android.
1229 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1232 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1233 When in FIPS mode the approved implementations are used as normal,
1234 when not in FIPS mode the internal unapproved versions are used instead.
1235 This means that the FIPS capable OpenSSL isn't forced to use the
1236 (often lower perfomance) FIPS implementations outside FIPS mode.
1239 *) Transparently support X9.42 DH parameters when calling
1240 PEM_read_bio_DHparameters. This means existing applications can handle
1241 the new parameter format automatically.
1244 *) Initial experimental support for X9.42 DH parameter format: mainly
1245 to support use of 'q' parameter for RFC5114 parameters.
1248 *) Add DH parameters from RFC5114 including test data to dhtest.
1251 *) Support for automatic EC temporary key parameter selection. If enabled
1252 the most preferred EC parameters are automatically used instead of
1253 hardcoded fixed parameters. Now a server just has to call:
1254 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1255 support ECDH and use the most appropriate parameters.
1258 *) Enhance and tidy EC curve and point format TLS extension code. Use
1259 static structures instead of allocation if default values are used.
1260 New ctrls to set curves we wish to support and to retrieve shared curves.
1261 Print out shared curves in s_server. New options to s_server and s_client
1262 to set list of supported curves.
1265 *) New ctrls to retrieve supported signature algorithms and
1266 supported curve values as an array of NIDs. Extend openssl utility
1267 to print out received values.
1270 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1271 between NIDs and the more common NIST names such as "P-256". Enhance
1272 ecparam utility and ECC method to recognise the NIST names for curves.
1275 *) Enhance SSL/TLS certificate chain handling to support different
1276 chains for each certificate instead of one chain in the parent SSL_CTX.
1279 *) Support for fixed DH ciphersuite client authentication: where both
1280 server and client use DH certificates with common parameters.
1283 *) Support for fixed DH ciphersuites: those requiring DH server
1287 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1289 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1290 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1291 X509_CINF_get_signature were reverted post internal team review.
1293 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1295 *) Build fixes for the Windows and OpenVMS platforms
1296 [Matt Caswell and Richard Levitte]
1298 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1300 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1301 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1302 dereference. This could lead to a Denial Of Service attack. Thanks to
1303 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1307 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1308 dtls1_buffer_record function under certain conditions. In particular this
1309 could occur if an attacker sent repeated DTLS records with the same
1310 sequence number but for the next epoch. The memory leak could be exploited
1311 by an attacker in a Denial of Service attack through memory exhaustion.
1312 Thanks to Chris Mueller for reporting this issue.
1316 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1317 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1318 method would be set to NULL which could later result in a NULL pointer
1319 dereference. Thanks to Frank Schmirler for reporting this issue.
1323 *) Abort handshake if server key exchange message is omitted for ephemeral
1326 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1327 reporting this issue.
1331 *) Remove non-export ephemeral RSA code on client and server. This code
1332 violated the TLS standard by allowing the use of temporary RSA keys in
1333 non-export ciphersuites and could be used by a server to effectively
1334 downgrade the RSA key length used to a value smaller than the server
1335 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1336 INRIA or reporting this issue.
1340 *) Fixed issue where DH client certificates are accepted without verification.
1341 An OpenSSL server will accept a DH certificate for client authentication
1342 without the certificate verify message. This effectively allows a client to
1343 authenticate without the use of a private key. This only affects servers
1344 which trust a client certificate authority which issues certificates
1345 containing DH keys: these are extremely rare and hardly ever encountered.
1346 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1351 *) Ensure that the session ID context of an SSL is updated when its
1352 SSL_CTX is updated via SSL_set_SSL_CTX.
1354 The session ID context is typically set from the parent SSL_CTX,
1355 and can vary with the CTX.
1358 *) Fix various certificate fingerprint issues.
1360 By using non-DER or invalid encodings outside the signed portion of a
1361 certificate the fingerprint can be changed without breaking the signature.
1362 Although no details of the signed portion of the certificate can be changed
1363 this can cause problems with some applications: e.g. those using the
1364 certificate fingerprint for blacklists.
1366 1. Reject signatures with non zero unused bits.
1368 If the BIT STRING containing the signature has non zero unused bits reject
1369 the signature. All current signature algorithms require zero unused bits.
1371 2. Check certificate algorithm consistency.
1373 Check the AlgorithmIdentifier inside TBS matches the one in the
1374 certificate signature. NB: this will result in signature failure
1375 errors for some broken certificates.
1377 Thanks to Konrad Kraszewski from Google for reporting this issue.
1379 3. Check DSA/ECDSA signatures use DER.
1381 Reencode DSA/ECDSA signatures and compare with the original received
1382 signature. Return an error if there is a mismatch.
1384 This will reject various cases including garbage after signature
1385 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1386 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1387 (negative or with leading zeroes).
1389 Further analysis was conducted and fixes were developed by Stephen Henson
1390 of the OpenSSL core team.
1395 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1396 results on some platforms, including x86_64. This bug occurs at random
1397 with a very low probability, and is not known to be exploitable in any
1398 way, though its exact impact is difficult to determine. Thanks to Pieter
1399 Wuille (Blockstream) who reported this issue and also suggested an initial
1400 fix. Further analysis was conducted by the OpenSSL development team and
1401 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1402 the OpenSSL core team.
1406 *) Do not resume sessions on the server if the negotiated protocol
1407 version does not match the session's version. Resuming with a different
1408 version, while not strictly forbidden by the RFC, is of questionable
1409 sanity and breaks all known clients.
1410 [David Benjamin, Emilia Käsper]
1412 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1413 early CCS messages during renegotiation. (Note that because
1414 renegotiation is encrypted, this early CCS was not exploitable.)
1417 *) Tighten client-side session ticket handling during renegotiation:
1418 ensure that the client only accepts a session ticket if the server sends
1419 the extension anew in the ServerHello. Previously, a TLS client would
1420 reuse the old extension state and thus accept a session ticket if one was
1421 announced in the initial ServerHello.
1423 Similarly, ensure that the client requires a session ticket if one
1424 was advertised in the ServerHello. Previously, a TLS client would
1425 ignore a missing NewSessionTicket message.
1428 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1430 *) SRTP Memory Leak.
1432 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1433 sends a carefully crafted handshake message, to cause OpenSSL to fail
1434 to free up to 64k of memory causing a memory leak. This could be
1435 exploited in a Denial Of Service attack. This issue affects OpenSSL
1436 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1437 whether SRTP is used or configured. Implementations of OpenSSL that
1438 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1440 The fix was developed by the OpenSSL team.
1444 *) Session Ticket Memory Leak.
1446 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1447 integrity of that ticket is first verified. In the event of a session
1448 ticket integrity check failing, OpenSSL will fail to free memory
1449 causing a memory leak. By sending a large number of invalid session
1450 tickets an attacker could exploit this issue in a Denial Of Service
1455 *) Build option no-ssl3 is incomplete.
1457 When OpenSSL is configured with "no-ssl3" as a build option, servers
1458 could accept and complete a SSL 3.0 handshake, and clients could be
1459 configured to send them.
1461 [Akamai and the OpenSSL team]
1463 *) Add support for TLS_FALLBACK_SCSV.
1464 Client applications doing fallback retries should call
1465 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1467 [Adam Langley, Bodo Moeller]
1469 *) Add additional DigestInfo checks.
1471 Reencode DigestInto in DER and check against the original when
1472 verifying RSA signature: this will reject any improperly encoded
1473 DigestInfo structures.
1475 Note: this is a precautionary measure and no attacks are currently known.
1479 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1481 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1482 SRP code can be overrun an internal buffer. Add sanity check that
1483 g, A, B < N to SRP code.
1485 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1486 Group for discovering this issue.
1490 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1491 TLS 1.0 instead of higher protocol versions when the ClientHello message
1492 is badly fragmented. This allows a man-in-the-middle attacker to force a
1493 downgrade to TLS 1.0 even if both the server and the client support a
1494 higher protocol version, by modifying the client's TLS records.
1496 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1497 researching this issue.
1501 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1502 to a denial of service attack. A malicious server can crash the client
1503 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1504 ciphersuite and sending carefully crafted handshake messages.
1506 Thanks to Felix Gröbert (Google) for discovering and researching this
1511 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1512 to leak memory. This can be exploited through a Denial of Service attack.
1513 Thanks to Adam Langley for discovering and researching this issue.
1517 *) An attacker can force openssl to consume large amounts of memory whilst
1518 processing DTLS handshake messages. This can be exploited through a
1519 Denial of Service attack.
1520 Thanks to Adam Langley for discovering and researching this issue.
1524 *) An attacker can force an error condition which causes openssl to crash
1525 whilst processing DTLS packets due to memory being freed twice. This
1526 can be exploited through a Denial of Service attack.
1527 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1532 *) If a multithreaded client connects to a malicious server using a resumed
1533 session and the server sends an ec point format extension it could write
1534 up to 255 bytes to freed memory.
1536 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1541 *) A malicious server can crash an OpenSSL client with a null pointer
1542 dereference (read) by specifying an SRP ciphersuite even though it was not
1543 properly negotiated with the client. This can be exploited through a
1544 Denial of Service attack.
1546 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1547 discovering and researching this issue.
1551 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1552 X509_name_oneline, X509_name_print_ex et al. to leak some information
1553 from the stack. Applications may be affected if they echo pretty printing
1554 output to the attacker.
1556 Thanks to Ivan Fratric (Google) for discovering this issue.
1558 [Emilia Käsper, and Steve Henson]
1560 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1561 for corner cases. (Certain input points at infinity could lead to
1562 bogus results, with non-infinity inputs mapped to infinity too.)
1565 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1567 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1568 handshake can force the use of weak keying material in OpenSSL
1569 SSL/TLS clients and servers.
1571 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1572 researching this issue. (CVE-2014-0224)
1573 [KIKUCHI Masashi, Steve Henson]
1575 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1576 OpenSSL DTLS client the code can be made to recurse eventually crashing
1579 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1581 [Imre Rad, Steve Henson]
1583 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1584 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1585 client or server. This is potentially exploitable to run arbitrary
1586 code on a vulnerable client or server.
1588 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1589 [Jüri Aedla, Steve Henson]
1591 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1592 are subject to a denial of service attack.
1594 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1595 this issue. (CVE-2014-3470)
1596 [Felix Gröbert, Ivan Fratric, Steve Henson]
1598 *) Harmonize version and its documentation. -f flag is used to display
1600 [mancha <mancha1@zoho.com>]
1602 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1603 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1604 [mancha <mancha1@zoho.com>]
1606 *) Fix some double frees. These are not thought to be exploitable.
1607 [mancha <mancha1@zoho.com>]
1609 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1611 *) A missing bounds check in the handling of the TLS heartbeat extension
1612 can be used to reveal up to 64k of memory to a connected client or
1615 Thanks for Neel Mehta of Google Security for discovering this bug and to
1616 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1617 preparing the fix (CVE-2014-0160)
1618 [Adam Langley, Bodo Moeller]
1620 *) Fix for the attack described in the paper "Recovering OpenSSL
1621 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1622 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1623 http://eprint.iacr.org/2014/140
1625 Thanks to Yuval Yarom and Naomi Benger for discovering this
1626 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1627 [Yuval Yarom and Naomi Benger]
1629 *) TLS pad extension: draft-agl-tls-padding-03
1631 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1632 TLS client Hello record length value would otherwise be > 255 and
1633 less that 512 pad with a dummy extension containing zeroes so it
1634 is at least 512 bytes long.
1636 [Adam Langley, Steve Henson]
1638 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1640 *) Fix for TLS record tampering bug. A carefully crafted invalid
1641 handshake could crash OpenSSL with a NULL pointer exception.
1642 Thanks to Anton Johansson for reporting this issues.
1645 *) Keep original DTLS digest and encryption contexts in retransmission
1646 structures so we can use the previous session parameters if they need
1647 to be resent. (CVE-2013-6450)
1650 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1651 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1652 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1653 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1654 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1655 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1656 [Rob Stradling, Adam Langley]
1658 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1660 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1661 supporting platforms or when small records were transferred.
1662 [Andy Polyakov, Steve Henson]
1664 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1666 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1668 This addresses the flaw in CBC record processing discovered by
1669 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1670 at: http://www.isg.rhul.ac.uk/tls/
1672 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1673 Security Group at Royal Holloway, University of London
1674 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1675 Emilia Käsper for the initial patch.
1677 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1679 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1680 ciphersuites which can be exploited in a denial of service attack.
1681 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1682 and detecting this bug and to Wolfgang Ettlinger
1683 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1687 *) Return an error when checking OCSP signatures when key is NULL.
1688 This fixes a DoS attack. (CVE-2013-0166)
1691 *) Make openssl verify return errors.
1692 [Chris Palmer <palmer@google.com> and Ben Laurie]
1694 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1695 the right response is stapled. Also change SSL_get_certificate()
1696 so it returns the certificate actually sent.
1697 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1698 [Rob Stradling <rob.stradling@comodo.com>]
1700 *) Fix possible deadlock when decoding public keys.
1703 *) Don't use TLS 1.0 record version number in initial client hello
1707 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1709 *) Sanity check record length before skipping explicit IV in TLS
1710 1.2, 1.1 and DTLS to fix DoS attack.
1712 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1713 fuzzing as a service testing platform.
1717 *) Initialise tkeylen properly when encrypting CMS messages.
1718 Thanks to Solar Designer of Openwall for reporting this issue.
1721 *) In FIPS mode don't try to use composite ciphers as they are not
1725 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1727 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1728 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1729 mean any application compiled against OpenSSL 1.0.0 headers setting
1730 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1731 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1732 0x10000000L Any application which was previously compiled against
1733 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1734 will need to be recompiled as a result. Letting be results in
1735 inability to disable specifically TLS 1.1 and in client context,
1736 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1739 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1740 disable just protocol X, but all protocols above X *if* there are
1741 protocols *below* X still enabled. In more practical terms it means
1742 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1743 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1744 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1748 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1750 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1751 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1752 in CRYPTO_realloc_clean.
1754 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1755 issue and to Adam Langley <agl@chromium.org> for fixing it.
1757 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1759 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1762 *) Workarounds for some broken servers that "hang" if a client hello
1763 record length exceeds 255 bytes.
1765 1. Do not use record version number > TLS 1.0 in initial client
1766 hello: some (but not all) hanging servers will now work.
1767 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1768 the number of ciphers sent in the client hello. This should be
1769 set to an even number, such as 50, for example by passing:
1770 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1771 Most broken servers should now work.
1772 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1773 TLS 1.2 client support entirely.
1776 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1779 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1781 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1782 STRING form instead of a DigestInfo.
1785 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1786 and the RSA_sign/RSA_verify functions. This was made more apparent when
1787 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1788 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1789 the correct format in RSA_verify so both forms transparently work.
1792 *) Some servers which support TLS 1.0 can choke if we initially indicate
1793 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1794 encrypted premaster secret. As a workaround use the maximum pemitted
1795 client version in client hello, this should keep such servers happy
1796 and still work with previous versions of OpenSSL.
1799 *) Add support for TLS/DTLS heartbeats.
1800 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1802 *) Add support for SCTP.
1803 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1805 *) Improved PRNG seeding for VOS.
1806 [Paul Green <Paul.Green@stratus.com>]
1808 *) Extensive assembler packs updates, most notably:
1810 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1811 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1812 - x86_64: bit-sliced AES implementation;
1813 - ARM: NEON support, contemporary platforms optimizations;
1814 - s390x: z196 support;
1815 - *: GHASH and GF(2^m) multiplication implementations;
1819 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1820 (removal of unnecessary code)
1821 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1823 *) Add TLS key material exporter from RFC 5705.
1826 *) Add DTLS-SRTP negotiation from RFC 5764.
1829 *) Add Next Protocol Negotiation,
1830 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1831 disabled with a no-npn flag to config or Configure. Code donated
1833 [Adam Langley <agl@google.com> and Ben Laurie]
1835 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1836 NIST-P256, NIST-P521, with constant-time single point multiplication on
1837 typical inputs. Compiler support for the nonstandard type __uint128_t is
1838 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1839 Code made available under Apache License version 2.0.
1841 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1842 line to include this in your build of OpenSSL, and run "make depend" (or
1843 "make update"). This enables the following EC_METHODs:
1845 EC_GFp_nistp224_method()
1846 EC_GFp_nistp256_method()
1847 EC_GFp_nistp521_method()
1849 EC_GROUP_new_by_curve_name() will automatically use these (while
1850 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1852 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1854 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1855 all platforms. Move ssize_t definition from e_os.h to the public
1856 header file e_os2.h as it now appears in public header file cms.h
1859 *) New -sigopt option to the ca, req and x509 utilities. Additional
1860 signature parameters can be passed using this option and in
1864 *) Add RSA PSS signing function. This will generate and set the
1865 appropriate AlgorithmIdentifiers for PSS based on those in the
1866 corresponding EVP_MD_CTX structure. No application support yet.
1869 *) Support for companion algorithm specific ASN1 signing routines.
1870 New function ASN1_item_sign_ctx() signs a pre-initialised
1871 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1872 the appropriate parameters.
1875 *) Add new algorithm specific ASN1 verification initialisation function
1876 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1877 handling will be the same no matter what EVP_PKEY_METHOD is used.
1878 Add a PSS handler to support verification of PSS signatures: checked
1879 against a number of sample certificates.
1882 *) Add signature printing for PSS. Add PSS OIDs.
1883 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1885 *) Add algorithm specific signature printing. An individual ASN1 method
1886 can now print out signatures instead of the standard hex dump.
1888 More complex signatures (e.g. PSS) can print out more meaningful
1889 information. Include DSA version that prints out the signature
1893 *) Password based recipient info support for CMS library: implementing
1897 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1898 neatly separates the code into cipher and PBE sections and is required
1899 for some algorithms that split PBES2 into separate pieces (such as
1900 password based CMS).
1903 *) Session-handling fixes:
1904 - Fix handling of connections that are resuming with a session ID,
1905 but also support Session Tickets.
1906 - Fix a bug that suppressed issuing of a new ticket if the client
1907 presented a ticket with an expired session.
1908 - Try to set the ticket lifetime hint to something reasonable.
1909 - Make tickets shorter by excluding irrelevant information.
1910 - On the client side, don't ignore renewed tickets.
1911 [Adam Langley, Bodo Moeller (Google)]
1913 *) Fix PSK session representation.
1916 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1918 This work was sponsored by Intel.
1921 *) Add GCM support to TLS library. Some custom code is needed to split
1922 the IV between the fixed (from PRF) and explicit (from TLS record)
1923 portions. This adds all GCM ciphersuites supported by RFC5288 and
1924 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
1925 add a special AESGCM string for GCM only.
1928 *) Expand range of ctrls for AES GCM. Permit setting invocation
1929 field on decrypt and retrieval of invocation field only on encrypt.
1932 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1933 As required by RFC5289 these ciphersuites cannot be used if for
1934 versions of TLS earlier than 1.2.
1937 *) For FIPS capable OpenSSL interpret a NULL default public key method
1938 as unset and return the appopriate default but do *not* set the default.
1939 This means we can return the appopriate method in applications that
1940 swicth between FIPS and non-FIPS modes.
1943 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1944 ENGINE is used then we cannot handle that in the FIPS module so we
1945 keep original code iff non-FIPS operations are allowed.
1948 *) Add -attime option to openssl utilities.
1949 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1951 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1954 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1955 FIPS EC methods unconditionally for now.
1958 *) New build option no-ec2m to disable characteristic 2 code.
1961 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1962 all cases can be covered as some introduce binary incompatibilities.
1965 *) Redirect RSA operations to FIPS module including keygen,
1966 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1969 *) Add similar low level API blocking to ciphers.
1972 *) Low level digest APIs are not approved in FIPS mode: any attempt
1973 to use these will cause a fatal error. Applications that *really* want
1974 to use them can use the private_* version instead.
1977 *) Redirect cipher operations to FIPS module for FIPS builds.
1980 *) Redirect digest operations to FIPS module for FIPS builds.
1983 *) Update build system to add "fips" flag which will link in fipscanister.o
1984 for static and shared library builds embedding a signature if needed.
1987 *) Output TLS supported curves in preference order instead of numerical
1988 order. This is currently hardcoded for the highest order curves first.
1989 This should be configurable so applications can judge speed vs strength.
1992 *) Add TLS v1.2 server support for client authentication.
1995 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1999 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2000 FIPS modules versions.
2003 *) Add TLS v1.2 client side support for client authentication. Keep cache
2004 of handshake records longer as we don't know the hash algorithm to use
2005 until after the certificate request message is received.
2008 *) Initial TLS v1.2 client support. Add a default signature algorithms
2009 extension including all the algorithms we support. Parse new signature
2010 format in client key exchange. Relax some ECC signing restrictions for
2011 TLS v1.2 as indicated in RFC5246.
2014 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2015 to new signature format when needed using client digest preference.
2016 All server ciphersuites should now work correctly in TLS v1.2. No client
2017 support yet and no support for client certificates.
2020 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2021 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2022 ciphersuites. At present only RSA key exchange ciphersuites work with
2023 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2024 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2025 and version checking.
2028 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2029 with this defined it will not be affected by any changes to ssl internal
2030 structures. Add several utility functions to allow openssl application
2031 to work with OPENSSL_NO_SSL_INTERN defined.
2034 *) A long standing patch to add support for SRP from EdelWeb (Peter
2035 Sylvester and Christophe Renou) was integrated.
2036 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
2037 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
2040 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2043 *) Permit abbreviated handshakes when renegotiating using the function
2044 SSL_renegotiate_abbreviated().
2045 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2047 *) Add call to ENGINE_register_all_complete() to
2048 ENGINE_load_builtin_engines(), so some implementations get used
2049 automatically instead of needing explicit application support.
2052 *) Add support for TLS key exporter as described in RFC5705.
2053 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2055 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2056 a few changes are required:
2058 Add SSL_OP_NO_TLSv1_1 flag.
2059 Add TLSv1_1 methods.
2060 Update version checking logic to handle version 1.1.
2061 Add explicit IV handling (ported from DTLS code).
2062 Add command line options to s_client/s_server.
2065 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2067 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2068 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2069 content decryption and always return the same error. Note: this attack
2070 needs on average 2^20 messages so it only affects automated senders. The
2071 old behaviour can be reenabled in the CMS code by setting the
2072 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2073 an MMA defence is not necessary.
2074 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2075 this issue. (CVE-2012-0884)
2078 *) Fix CVE-2011-4619: make sure we really are receiving a
2079 client hello before rejecting multiple SGC restarts. Thanks to
2080 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2083 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2085 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2086 Thanks to Antonio Martin, Enterprise Secure Access Research and
2087 Development, Cisco Systems, Inc. for discovering this bug and
2088 preparing a fix. (CVE-2012-0050)
2091 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2093 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2094 of the Vaudenay padding oracle attack on CBC mode encryption
2095 which enables an efficient plaintext recovery attack against
2096 the OpenSSL implementation of DTLS. Their attack exploits timing
2097 differences arising during decryption processing. A research
2098 paper describing this attack can be found at:
2099 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2100 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2101 Security Group at Royal Holloway, University of London
2102 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2103 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2104 for preparing the fix. (CVE-2011-4108)
2105 [Robin Seggelmann, Michael Tuexen]
2107 *) Clear bytes used for block padding of SSL 3.0 records.
2109 [Adam Langley (Google)]
2111 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2112 Kadianakis <desnacked@gmail.com> for discovering this issue and
2113 Adam Langley for preparing the fix. (CVE-2011-4619)
2114 [Adam Langley (Google)]
2116 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2117 [Andrey Kulikov <amdeich@gmail.com>]
2119 *) Prevent malformed RFC3779 data triggering an assertion failure.
2120 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2121 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2122 [Rob Austein <sra@hactrn.net>]
2124 *) Improved PRNG seeding for VOS.
2125 [Paul Green <Paul.Green@stratus.com>]
2127 *) Fix ssl_ciph.c set-up race.
2128 [Adam Langley (Google)]
2130 *) Fix spurious failures in ecdsatest.c.
2131 [Emilia Käsper (Google)]
2133 *) Fix the BIO_f_buffer() implementation (which was mixing different
2134 interpretations of the '..._len' fields).
2135 [Adam Langley (Google)]
2137 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2138 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2139 threads won't reuse the same blinding coefficients.
2141 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2142 lock to call BN_BLINDING_invert_ex, and avoids one use of
2143 BN_BLINDING_update for each BN_BLINDING structure (previously,
2144 the last update always remained unused).
2145 [Emilia Käsper (Google)]
2147 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2148 [Bob Buckholz (Google)]
2150 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2152 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2153 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2154 [Kaspar Brand <ossl@velox.ch>]
2156 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2157 for multi-threaded use of ECDH. (CVE-2011-3210)
2158 [Adam Langley (Google)]
2160 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2163 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2164 signature public key algorithm by using OID xref utilities instead.
2165 Before this you could only use some ECC ciphersuites with SHA1 only.
2168 *) Add protection against ECDSA timing attacks as mentioned in the paper
2169 by Billy Bob Brumley and Nicola Tuveri, see:
2171 http://eprint.iacr.org/2011/232.pdf
2173 [Billy Bob Brumley and Nicola Tuveri]
2175 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2177 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2178 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2180 *) Fix bug in string printing code: if *any* escaping is enabled we must
2181 escape the escape character (backslash) or the resulting string is
2185 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2187 *) Disable code workaround for ancient and obsolete Netscape browsers
2188 and servers: an attacker can use it in a ciphersuite downgrade attack.
2189 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2192 *) Fixed J-PAKE implementation error, originally discovered by
2193 Sebastien Martini, further info and confirmation from Stefan
2194 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2197 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2199 *) Fix extension code to avoid race conditions which can result in a buffer
2200 overrun vulnerability: resumed sessions must not be modified as they can
2201 be shared by multiple threads. CVE-2010-3864
2204 *) Fix WIN32 build system to correctly link an ENGINE directory into
2208 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2210 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2212 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2214 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2216 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2217 context. The operation can be customised via the ctrl mechanism in
2218 case ENGINEs want to include additional functionality.
2221 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2224 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2225 output hashes compatible with older versions of OpenSSL.
2226 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2228 *) Fix compression algorithm handling: if resuming a session use the
2229 compression algorithm of the resumed session instead of determining
2230 it from client hello again. Don't allow server to change algorithm.
2233 *) Add load_crls() function to apps tidying load_certs() too. Add option
2234 to verify utility to allow additional CRLs to be included.
2237 *) Update OCSP request code to permit adding custom headers to the request:
2238 some responders need this.
2241 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2243 [Julia Lawall <julia@diku.dk>]
2245 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2246 needlessly dereferenced structures, used obsolete functions and
2247 didn't handle all updated verify codes correctly.
2250 *) Disable MD2 in the default configuration.
2253 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2254 indicate the initial BIO being pushed or popped. This makes it possible
2255 to determine whether the BIO is the one explicitly called or as a result
2256 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2257 it handles reference counts correctly and doesn't zero out the I/O bio
2258 when it is not being explicitly popped. WARNING: applications which
2259 included workarounds for the old buggy behaviour will need to be modified
2260 or they could free up already freed BIOs.
2263 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2264 renaming to all platforms (within the 0.9.8 branch, this was
2265 done conditionally on Netware platforms to avoid a name clash).
2266 [Guenter <lists@gknw.net>]
2268 *) Add ECDHE and PSK support to DTLS.
2269 [Michael Tuexen <tuexen@fh-muenster.de>]
2271 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2275 *) Add "missing" function EVP_MD_flags() (without this the only way to
2276 retrieve a digest flags is by accessing the structure directly. Update
2277 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2278 or cipher is registered as in the "from" argument. Print out all
2279 registered digests in the dgst usage message instead of manually
2280 attempting to work them out.
2283 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2284 this allows the use of compression and extensions. Change default cipher
2285 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2286 by default unless an application cipher string requests it.
2289 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2290 key ids to find matching certificates and keys but some PKCS#12 files
2291 don't follow the (somewhat unwritten) rules and this strategy fails.
2292 Now just gather all certificates together and the first private key
2293 then look for the first certificate that matches the key.
2296 *) Support use of registered digest and cipher names for dgst and cipher
2297 commands instead of having to add each one as a special case. So now
2304 openssl dgst -sha256 foo
2306 and this works for ENGINE based algorithms too.
2310 *) Update Gost ENGINE to support parameter files.
2311 [Victor B. Wagner <vitus@cryptocom.ru>]
2313 *) Support GeneralizedTime in ca utility.
2314 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2316 *) Enhance the hash format used for certificate directory links. The new
2317 form uses the canonical encoding (meaning equivalent names will work
2318 even if they aren't identical) and uses SHA1 instead of MD5. This form
2319 is incompatible with the older format and as a result c_rehash should
2320 be used to rebuild symbolic links.
2323 *) Make PKCS#8 the default write format for private keys, replacing the
2324 traditional format. This form is standardised, more secure and doesn't
2325 include an implicit MD5 dependency.
2328 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2329 committed to OpenSSL should pass this lot as a minimum.
2332 *) Add session ticket override functionality for use by EAP-FAST.
2333 [Jouni Malinen <j@w1.fi>]
2335 *) Modify HMAC functions to return a value. Since these can be implemented
2336 in an ENGINE errors can occur.
2339 *) Type-checked OBJ_bsearch_ex.
2342 *) Type-checked OBJ_bsearch. Also some constification necessitated
2343 by type-checking. Still to come: TXT_DB, bsearch(?),
2344 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2348 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2349 seconds to a tm structure directly, instead of going through OS
2350 specific date routines. This avoids any issues with OS routines such
2351 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2352 and X509_time_adj_ex() to cover the extended range. The existing
2353 X509_time_adj() is still usable and will no longer have any date issues.
2356 *) Delta CRL support. New use deltas option which will attempt to locate
2357 and search any appropriate delta CRLs available.
2359 This work was sponsored by Google.
2362 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2363 code and add additional score elements. Validate alternate CRL paths
2364 as part of the CRL checking and indicate a new error "CRL path validation
2365 error" in this case. Applications wanting additional details can use
2366 the verify callback and check the new "parent" field. If this is not
2367 NULL CRL path validation is taking place. Existing applications wont
2368 see this because it requires extended CRL support which is off by
2371 This work was sponsored by Google.
2374 *) Support for freshest CRL extension.
2376 This work was sponsored by Google.
2379 *) Initial indirect CRL support. Currently only supported in the CRLs
2380 passed directly and not via lookup. Process certificate issuer
2381 CRL entry extension and lookup CRL entries by bother issuer name
2382 and serial number. Check and process CRL issuer entry in IDP extension.
2384 This work was sponsored by Google.
2387 *) Add support for distinct certificate and CRL paths. The CRL issuer
2388 certificate is validated separately in this case. Only enabled if
2389 an extended CRL support flag is set: this flag will enable additional
2390 CRL functionality in future.
2392 This work was sponsored by Google.
2395 *) Add support for policy mappings extension.
2397 This work was sponsored by Google.
2400 *) Fixes to pathlength constraint, self issued certificate handling,
2401 policy processing to align with RFC3280 and PKITS tests.
2403 This work was sponsored by Google.
2406 *) Support for name constraints certificate extension. DN, email, DNS
2407 and URI types are currently supported.
2409 This work was sponsored by Google.
2412 *) To cater for systems that provide a pointer-based thread ID rather
2413 than numeric, deprecate the current numeric thread ID mechanism and
2414 replace it with a structure and associated callback type. This
2415 mechanism allows a numeric "hash" to be extracted from a thread ID in
2416 either case, and on platforms where pointers are larger than 'long',
2417 mixing is done to help ensure the numeric 'hash' is usable even if it
2418 can't be guaranteed unique. The default mechanism is to use "&errno"
2419 as a pointer-based thread ID to distinguish between threads.
2421 Applications that want to provide their own thread IDs should now use
2422 CRYPTO_THREADID_set_callback() to register a callback that will call
2423 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2425 Note that ERR_remove_state() is now deprecated, because it is tied
2426 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2427 to free the current thread's error state should be replaced by
2428 ERR_remove_thread_state(NULL).
2430 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2431 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2432 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2433 application was previously providing a numeric thread callback that
2434 was inappropriate for distinguishing threads, then uniqueness might
2435 have been obtained with &errno that happened immediately in the
2436 intermediate development versions of OpenSSL; this is no longer the
2437 case, the numeric thread callback will now override the automatic use
2439 [Geoff Thorpe, with help from Bodo Moeller]
2441 *) Initial support for different CRL issuing certificates. This covers a
2442 simple case where the self issued certificates in the chain exist and
2443 the real CRL issuer is higher in the existing chain.
2445 This work was sponsored by Google.
2448 *) Removed effectively defunct crypto/store from the build.
2451 *) Revamp of STACK to provide stronger type-checking. Still to come:
2452 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2453 ASN1_STRING, CONF_VALUE.
2456 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2457 RAM on SSL connections. This option can save about 34k per idle SSL.
2460 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2461 STACK, TXT_DB, bsearch, qsort.
2464 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2465 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2466 support for data, signedData, compressedData, digestedData and
2467 encryptedData, envelopedData types included. Scripts to check against
2468 RFC4134 examples draft and interop and consistency checks of many
2469 content types and variants.
2472 *) Add options to enc utility to support use of zlib compression BIO.
2475 *) Extend mk1mf to support importing of options and assembly language
2476 files from Configure script, currently only included in VC-WIN32.
2477 The assembly language rules can now optionally generate the source
2478 files from the associated perl scripts.
2481 *) Implement remaining functionality needed to support GOST ciphersuites.
2482 Interop testing has been performed using CryptoPro implementations.
2483 [Victor B. Wagner <vitus@cryptocom.ru>]
2485 *) s390x assembler pack.
2488 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2492 *) Implement Opaque PRF Input TLS extension as specified in
2493 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2494 official specification yet and no extension type assignment by
2495 IANA exists, this extension (for now) will have to be explicitly
2496 enabled when building OpenSSL by providing the extension number
2497 to use. For example, specify an option
2499 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2501 to the "config" or "Configure" script to enable the extension,
2502 assuming extension number 0x9527 (which is a completely arbitrary
2503 and unofficial assignment based on the MD5 hash of the Internet
2504 Draft). Note that by doing so, you potentially lose
2505 interoperability with other TLS implementations since these might
2506 be using the same extension number for other purposes.
2508 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2509 opaque PRF input value to use in the handshake. This will create
2510 an interal copy of the length-'len' string at 'src', and will
2511 return non-zero for success.
2513 To get more control and flexibility, provide a callback function
2516 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2517 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2521 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2524 Callback function 'cb' will be called in handshakes, and is
2525 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2526 Argument 'arg' is for application purposes (the value as given to
2527 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2528 be provided to the callback function). The callback function
2529 has to return non-zero to report success: usually 1 to use opaque
2530 PRF input just if possible, or 2 to enforce use of the opaque PRF
2531 input. In the latter case, the library will abort the handshake
2532 if opaque PRF input is not successfully negotiated.
2534 Arguments 'peerinput' and 'len' given to the callback function
2535 will always be NULL and 0 in the case of a client. A server will
2536 see the client's opaque PRF input through these variables if
2537 available (NULL and 0 otherwise). Note that if the server
2538 provides an opaque PRF input, the length must be the same as the
2539 length of the client's opaque PRF input.
2541 Note that the callback function will only be called when creating
2542 a new session (session resumption can resume whatever was
2543 previously negotiated), and will not be called in SSL 2.0
2544 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2545 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2546 for applications that need to enforce opaque PRF input.
2550 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2553 [Victor B. Wagner <vitus@cryptocom.ru>]
2555 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2556 RFC4507bis. The encrypted ticket format is an encrypted encoded
2557 SSL_SESSION structure, that way new session features are automatically
2560 If a client application caches session in an SSL_SESSION structure
2561 support is transparent because tickets are now stored in the encoded
2564 The SSL_CTX structure automatically generates keys for ticket
2565 protection in servers so again support should be possible
2566 with no application modification.
2568 If a client or server wishes to disable RFC4507 support then the option
2569 SSL_OP_NO_TICKET can be set.
2571 Add a TLS extension debugging callback to allow the contents of any client
2572 or server extensions to be examined.
2574 This work was sponsored by Google.
2577 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2578 OpenSSL should now compile cleanly on gcc 4.2
2579 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2581 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2582 support including streaming MAC support: this is required for GOST
2583 ciphersuite support.
2584 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2586 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2587 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2588 to output in BER and PEM format.
2591 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2592 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2593 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2594 ENGINE support for HMAC keys which are unextractable. New -mac and
2595 -macopt options to dgst utility.
2598 *) New option -sigopt to dgst utility. Update dgst to use
2599 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2600 alternative signing paramaters such as X9.31 or PSS in the dgst
2604 *) Change ssl_cipher_apply_rule(), the internal function that does
2605 the work each time a ciphersuite string requests enabling
2606 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2607 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2608 the order of disabled ciphersuites such that those ciphersuites
2609 that most recently went from enabled to disabled not only stay
2610 in order with respect to each other, but also have higher priority
2611 than other disabled ciphersuites the next time ciphersuites are
2614 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2615 the same ciphersuites as with "HIGH" alone, but in a specific
2616 order where the PSK ciphersuites come first (since they are the
2617 most recently disabled ciphersuites when "HIGH" is parsed).
2619 Also, change ssl_create_cipher_list() (using this new
2620 funcionality) such that between otherwise identical
2621 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2625 *) Change ssl_create_cipher_list() so that it automatically
2626 arranges the ciphersuites in reasonable order before starting
2627 to process the rule string. Thus, the definition for "DEFAULT"
2628 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2629 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2630 This makes it much easier to arrive at a reasonable default order
2631 in applications for which anonymous ciphers are OK (meaning
2632 that you can't actually use DEFAULT).
2633 [Bodo Moeller; suggested by Victor Duchovni]
2635 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2636 processing) into multiple integers instead of setting
2637 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2638 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2639 (These masks as well as the individual bit definitions are hidden
2640 away into the non-exported interface ssl/ssl_locl.h, so this
2641 change to the definition of the SSL_CIPHER structure shouldn't
2642 affect applications.) This give us more bits for each of these
2643 categories, so there is no longer a need to coagulate AES128 and
2644 AES256 into a single algorithm bit, and to coagulate Camellia128
2645 and Camellia256 into a single algorithm bit, which has led to all
2648 Thus, among other things, the kludge introduced in 0.9.7m and
2649 0.9.8e for masking out AES256 independently of AES128 or masking
2650 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2652 With the change, we also introduce new ciphersuite aliases that
2653 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2657 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2658 Use the leftmost N bytes of the signature input if the input is
2659 larger than the prime q (with N being the size in bytes of q).
2662 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2663 it yet and it is largely untested.
2666 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2669 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2670 some compilers (gcc 4.2 and later) reject their use. Safestack is
2671 reimplemented. Update ASN1 to avoid use of legacy functions.
2674 *) Win32/64 targets are linked with Winsock2.
2677 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2678 to external functions. This can be used to increase CRL handling
2679 efficiency especially when CRLs are very large by (for example) storing
2680 the CRL revoked certificates in a database.
2683 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2684 new CRLs added to a directory can be used. New command line option
2685 -verify_return_error to s_client and s_server. This causes real errors
2686 to be returned by the verify callback instead of carrying on no matter
2687 what. This reflects the way a "real world" verify callback would behave.
2690 *) GOST engine, supporting several GOST algorithms and public key formats.
2691 Kindly donated by Cryptocom.
2694 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2695 partitioned by DP are handled but no indirect CRL or reason partitioning
2696 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2697 selected via a scoring technique which handles IDP and AKID in CRLs.
2700 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2701 will ultimately be used for all verify operations: this will remove the
2702 X509_STORE dependency on certificate verification and allow alternative
2703 lookup methods. X509_STORE based implementations of these two callbacks.
2706 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2707 Modify get_crl() to find a valid (unexpired) CRL if possible.
2710 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2711 this would be called X509_CRL_cmp() but that name is already used by
2712 a function that just compares CRL issuer names. Cache several CRL
2713 extensions in X509_CRL structure and cache CRLDP in X509.
2716 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2717 this maps equivalent X509_NAME structures into a consistent structure.
2718 Name comparison can then be performed rapidly using memcmp().
2721 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2725 *) Allow digests to supply their own micalg string for S/MIME type using
2726 the ctrl EVP_MD_CTRL_MICALG.
2729 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2730 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2731 ctrl. It can then customise the structure before and/or after signing
2735 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2736 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2737 to free up any added signature OIDs.
2740 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2741 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2742 digest and cipher tables. New options added to openssl utility:
2743 list-message-digest-algorithms and list-cipher-algorithms.
2746 *) Change the array representation of binary polynomials: the list
2747 of degrees of non-zero coefficients is now terminated with -1.
2748 Previously it was terminated with 0, which was also part of the
2749 value; thus, the array representation was not applicable to
2750 polynomials where t^0 has coefficient zero. This change makes
2751 the array representation useful in a more general context.
2754 *) Various modifications and fixes to SSL/TLS cipher string
2755 handling. For ECC, the code now distinguishes between fixed ECDH
2756 with RSA certificates on the one hand and with ECDSA certificates
2757 on the other hand, since these are separate ciphersuites. The
2758 unused code for Fortezza ciphersuites has been removed.
2760 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2761 (not "ECDHE"). For consistency with the code for DH
2762 certificates, use of ECDH certificates is now considered ECDH
2763 authentication, not RSA or ECDSA authentication (the latter is
2764 merely the CA's signing algorithm and not actively used in the
2767 The temporary ciphersuite alias "ECCdraft" is no longer
2768 available, and ECC ciphersuites are no longer excluded from "ALL"
2769 and "DEFAULT". The following aliases now exist for RFC 4492
2770 ciphersuites, most of these by analogy with the DH case:
2772 kECDHr - ECDH cert, signed with RSA
2773 kECDHe - ECDH cert, signed with ECDSA
2774 kECDH - ECDH cert (signed with either RSA or ECDSA)
2775 kEECDH - ephemeral ECDH
2776 ECDH - ECDH cert or ephemeral ECDH
2782 AECDH - anonymous ECDH
2783 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2787 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2788 Use correct micalg parameters depending on digest(s) in signed message.
2791 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2792 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2795 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2796 an engine to register a method. Add ENGINE lookups for methods and
2797 functional reference processing.
2800 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2801 EVP_{Sign,Verify}* which allow an application to customise the signature
2805 *) New -resign option to smime utility. This adds one or more signers
2806 to an existing PKCS#7 signedData structure. Also -md option to use an
2807 alternative message digest algorithm for signing.
2810 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2811 create PKCS7 structures containing multiple signers. Update smime
2812 application to support multiple signers.
2815 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2819 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2820 Reorganize PBE internals to lookup from a static table using NIDs,
2821 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2822 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2823 PRF which will be automatically used with PBES2.
2826 *) Replace the algorithm specific calls to generate keys in "req" with the
2830 *) Update PKCS#7 enveloped data routines to use new API. This is now
2831 supported by any public key method supporting the encrypt operation. A
2832 ctrl is added to allow the public key algorithm to examine or modify
2833 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2837 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2838 a default digest type to use. In most cases this will be SHA1 but some
2839 algorithms (such as GOST) need to specify an alternative digest. The
2840 return value indicates how strong the prefernce is 1 means optional and
2841 2 is mandatory (that is it is the only supported type). Modify
2842 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2843 use the default md. Update openssl utilities to use the default digest
2844 type for signing if it is not explicitly indicated.
2847 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2848 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2849 signing method from the key type. This effectively removes the link
2850 between digests and public key types.
2853 *) Add an OID cross reference table and utility functions. Its purpose is to
2854 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2855 rsaEncryption. This will allow some of the algorithm specific hackery
2856 needed to use the correct OID to be removed.
2859 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2860 structures for PKCS7_sign(). They are now set up by the relevant public
2864 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2867 *) Add support for key derivation (agreement) in the API, DH method and
2871 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2872 public and private key formats. As a side effect these add additional
2873 command line functionality not previously available: DSA signatures can be
2874 generated and verified using pkeyutl and DH key support and generation in
2879 [Oliver Tappe <zooey@hirschkaefer.de>]
2881 *) New make target "install_html_docs" installs HTML renditions of the
2883 [Oliver Tappe <zooey@hirschkaefer.de>]
2885 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
2886 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2887 support key and parameter generation and add initial key generation
2888 functionality for RSA.
2891 *) Add functions for main EVP_PKEY_method operations. The undocumented
2892 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2893 EVP_PKEY_{encrypt,decrypt}_old.
2896 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2897 key API, doesn't do much yet.
2900 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2901 public key algorithms. New option to openssl utility:
2902 "list-public-key-algorithms" to print out info.
2905 *) Implement the Supported Elliptic Curves Extension for
2906 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2909 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2910 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2913 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2914 utilities such as rsa, dsa, dsaparam etc except they process any key
2918 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2919 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2920 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2924 *) Initial support for pluggable public key ASN1.
2925 De-spaghettify the public key ASN1 handling. Move public and private
2926 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2927 algorithm specific handling to a single module within the relevant
2928 algorithm directory. Add functions to allow (near) opaque processing
2929 of public and private key structures.
2932 *) Implement the Supported Point Formats Extension for
2933 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2936 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2937 for the psk identity [hint] and the psk callback functions to the
2938 SSL_SESSION, SSL and SSL_CTX structure.
2941 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2945 SSL_CTX_use_psk_identity_hint
2946 SSL_get_psk_identity_hint
2947 SSL_get_psk_identity
2948 SSL_use_psk_identity_hint
2950 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2952 *) Add RFC 3161 compliant time stamp request creation, response generation
2953 and response verification functionality.
2954 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2956 *) Add initial support for TLS extensions, specifically for the server_name
2957 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2958 have new members for a host name. The SSL data structure has an
2959 additional member SSL_CTX *initial_ctx so that new sessions can be
2960 stored in that context to allow for session resumption, even after the
2961 SSL has been switched to a new SSL_CTX in reaction to a client's
2962 server_name extension.
2964 New functions (subject to change):
2966 SSL_get_servername()
2967 SSL_get_servername_type()
2970 New CTRL codes and macros (subject to change):
2972 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2973 - SSL_CTX_set_tlsext_servername_callback()
2974 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2975 - SSL_CTX_set_tlsext_servername_arg()
2976 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2978 openssl s_client has a new '-servername ...' option.
2980 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2981 '-key2 ...', '-servername_fatal' (subject to change). This allows
2982 testing the HostName extension for a specific single host name ('-cert'
2983 and '-key' remain fallbacks for handshakes without HostName
2984 negotiation). If the unrecogninzed_name alert has to be sent, this by
2985 default is a warning; it becomes fatal with the '-servername_fatal'
2988 [Peter Sylvester, Remy Allais, Christophe Renou]
2990 *) Whirlpool hash implementation is added.
2993 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2994 bn(64,32). Because of instruction set limitations it doesn't have
2995 any negative impact on performance. This was done mostly in order
2996 to make it possible to share assembler modules, such as bn_mul_mont
2997 implementations, between 32- and 64-bit builds without hassle.
3000 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3001 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3005 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3006 dedicated Montgomery multiplication procedure, is introduced.
3007 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3008 "64-bit" performance on certain 32-bit targets.
3011 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3012 in SSL structures. New SSL ctrl to set maximum send fragment size.
3013 Save memory by seeting the I/O buffer sizes dynamically instead of
3014 using the maximum available value.
3017 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3018 in addition to the text details.
3021 *) Very, very preliminary EXPERIMENTAL support for printing of general
3022 ASN1 structures. This currently produces rather ugly output and doesn't
3023 handle several customised structures at all.
3026 *) Integrated support for PVK file format and some related formats such
3027 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3028 these in the 'rsa' and 'dsa' utilities.
3031 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3034 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3035 place for the (very old) "NETSCAPE" format certificates which are now
3036 handled using new ASN1 code equivalents.
3039 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3040 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3041 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3044 *) Modify CRL distribution points extension code to print out previously
3045 unsupported fields. Enhance extension setting code to allow setting of
3049 *) Add print and set support for Issuing Distribution Point CRL extension.
3052 *) Change 'Configure' script to enable Camellia by default.
3055 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3057 *) When rejecting SSL/TLS records due to an incorrect version number, never
3058 update s->server with a new major version number. As of
3059 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3060 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3061 the previous behavior could result in a read attempt at NULL when
3062 receiving specific incorrect SSL/TLS records once record payload
3063 protection is active. (CVE-2010-0740)
3064 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3066 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3067 could be crashed if the relevant tables were not present (e.g. chrooted).
3068 [Tomas Hoger <thoger@redhat.com>]
3070 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3072 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3073 [Martin Olsson, Neel Mehta]
3075 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3076 accommodate for stack sorting, always a write lock!).
3079 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3080 excessive delays in the RAND_poll(): over a minute. As a workaround
3081 include a time check in the inner Heap32Next loop too.
3084 *) The code that handled flushing of data in SSL/TLS originally used the
3085 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3086 the problem outlined in PR#1949. The fix suggested there however can
3087 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3088 of Apache). So instead simplify the code to flush unconditionally.
3089 This should be fine since flushing with no data to flush is a no op.
3092 *) Handle TLS versions 2.0 and later properly and correctly use the
3093 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3094 off ancient servers have a habit of sticking around for a while...
3097 *) Modify compression code so it frees up structures without using the
3098 ex_data callbacks. This works around a problem where some applications
3099 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3100 restarting) then use compression (e.g. SSL with compression) later.
3101 This results in significant per-connection memory leaks and
3102 has caused some security issues including CVE-2008-1678 and
3106 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3107 change when encrypting or decrypting.
3110 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3111 connect and renegotiate with servers which do not support RI.
3112 Until RI is more widely deployed this option is enabled by default.
3115 *) Add "missing" ssl ctrls to clear options and mode.
3118 *) If client attempts to renegotiate and doesn't support RI respond with
3119 a no_renegotiation alert as required by RFC5746. Some renegotiating
3120 TLS clients will continue a connection gracefully when they receive
3121 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3122 waiting for a server hello which it will never receive. Now we treat a
3123 received no_renegotiation alert as a fatal error. This is because
3124 applications requesting a renegotiation might well expect it to succeed
3125 and would have no code in place to handle the server denying it so the
3126 only safe thing to do is to terminate the connection.
3129 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3130 peer supports secure renegotiation and 0 otherwise. Print out peer
3131 renegotiation support in s_client/s_server.
3134 *) Replace the highly broken and deprecated SPKAC certification method with
3135 the updated NID creation version. This should correctly handle UTF8.
3138 *) Implement RFC5746. Re-enable renegotiation but require the extension
3139 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3140 turns out to be a bad idea. It has been replaced by
3141 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3142 SSL_CTX_set_options(). This is really not recommended unless you
3143 know what you are doing.
3144 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3146 *) Fixes to stateless session resumption handling. Use initial_ctx when
3147 issuing and attempting to decrypt tickets in case it has changed during
3148 servername handling. Use a non-zero length session ID when attempting
3149 stateless session resumption: this makes it possible to determine if
3150 a resumption has occurred immediately after receiving server hello
3151 (several places in OpenSSL subtly assume this) instead of later in
3155 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3156 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3157 fixes for a few places where the return code is not checked
3159 [Julia Lawall <julia@diku.dk>]
3161 *) Add --strict-warnings option to Configure script to include devteam
3162 warnings in other configurations.
3165 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3166 makes it possible to install openssl libraries in locations which
3167 have names other than "lib", for example "/usr/lib64" which some
3169 [Steve Henson, based on patch from Jeremy Utley]
3171 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3172 X690 8.9.12 and can produce some misleading textual output of OIDs.
3173 [Steve Henson, reported by Dan Kaminsky]
3175 *) Delete MD2 from algorithm tables. This follows the recommendation in
3176 several standards that it is not used in new applications due to
3177 several cryptographic weaknesses. For binary compatibility reasons
3178 the MD2 API is still compiled in by default.
3181 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3185 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3186 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3188 [Guenter <lists@gknw.net>]
3190 *) Fix the server certificate chain building code to use X509_verify_cert(),
3191 it used to have an ad-hoc builder which was unable to cope with anything
3192 other than a simple chain.
3193 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3195 *) Don't check self signed certificate signatures in X509_verify_cert()
3196 by default (a flag can override this): it just wastes time without
3197 adding any security. As a useful side effect self signed root CAs
3198 with non-FIPS digests are now usable in FIPS mode.
3201 *) In dtls1_process_out_of_seq_message() the check if the current message
3202 is already buffered was missing. For every new message was memory
3203 allocated, allowing an attacker to perform an denial of service attack
3204 with sending out of seq handshake messages until there is no memory
3205 left. Additionally every future messege was buffered, even if the
3206 sequence number made no sense and would be part of another handshake.
3207 So only messages with sequence numbers less than 10 in advance will be
3208 buffered. (CVE-2009-1378)
3209 [Robin Seggelmann, discovered by Daniel Mentz]
3211 *) Records are buffered if they arrive with a future epoch to be
3212 processed after finishing the corresponding handshake. There is
3213 currently no limitation to this buffer allowing an attacker to perform
3214 a DOS attack with sending records with future epochs until there is no
3215 memory left. This patch adds the pqueue_size() function to detemine
3216 the size of a buffer and limits the record buffer to 100 entries.
3218 [Robin Seggelmann, discovered by Daniel Mentz]
3220 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3221 parent structure is freed. (CVE-2009-1379)
3224 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3225 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3228 [Ilya O. <vrghost@gmail.com>]
3230 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3232 *) Disable renegotiation completely - this fixes a severe security
3233 problem (CVE-2009-3555) at the cost of breaking all
3234 renegotiation. Renegotiation can be re-enabled by setting
3235 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3236 run-time. This is really not recommended unless you know what
3240 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3242 *) Don't set val to NULL when freeing up structures, it is freed up by
3243 underlying code. If sizeof(void *) > sizeof(long) this can result in
3244 zeroing past the valid field. (CVE-2009-0789)
3245 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3247 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3248 checked correctly. This would allow some invalid signed attributes to
3249 appear to verify correctly. (CVE-2009-0591)
3250 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3252 *) Reject UniversalString and BMPString types with invalid lengths. This
3253 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3254 a legal length. (CVE-2009-0590)
3257 *) Set S/MIME signing as the default purpose rather than setting it
3258 unconditionally. This allows applications to override it at the store
3262 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3263 to handle some structures.
3266 *) Improve efficiency of mem_gets: don't search whole buffer each time
3268 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3270 *) New -hex option for openssl rand.
3273 *) Print out UTF8String and NumericString when parsing ASN1.
3276 *) Support NumericString type for name components.
3279 *) Allow CC in the environment to override the automatically chosen
3280 compiler. Note that nothing is done to ensure flags work with the
3284 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3286 *) Properly check EVP_VerifyFinal() and similar return values
3288 [Ben Laurie, Bodo Moeller, Google Security Team]
3290 *) Enable TLS extensions by default.
3293 *) Allow the CHIL engine to be loaded, whether the application is
3294 multithreaded or not. (This does not release the developer from the
3295 obligation to set up the dynamic locking callbacks.)
3296 [Sander Temme <sander@temme.net>]
3298 *) Use correct exit code if there is an error in dgst command.
3299 [Steve Henson; problem pointed out by Roland Dirlewanger]
3301 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3302 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3305 *) Add experimental JPAKE support, including demo authentication in
3306 s_client and s_server.
3309 *) Set the comparison function in v3_addr_canonize().
3310 [Rob Austein <sra@hactrn.net>]
3312 *) Add support for XMPP STARTTLS in s_client.
3313 [Philip Paeps <philip@freebsd.org>]
3315 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3316 to ensure that even with this option, only ciphersuites in the
3317 server's preference list will be accepted. (Note that the option
3318 applies only when resuming a session, so the earlier behavior was
3319 just about the algorithm choice for symmetric cryptography.)
3322 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3324 *) Fix NULL pointer dereference if a DTLS server received
3325 ChangeCipherSpec as first record (CVE-2009-1386).
3328 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
3329 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3332 *) The fix in 0.9.8c that supposedly got rid of unsafe
3333 double-checked locking was incomplete for RSA blinding,
3334 addressing just one layer of what turns out to have been
3335 doubly unsafe triple-checked locking.
3337 So now fix this for real by retiring the MONT_HELPER macro
3338 in crypto/rsa/rsa_eay.c.
3340 [Bodo Moeller; problem pointed out by Marius Schilder]
3342 *) Various precautionary measures:
3344 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3346 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3347 (NB: This would require knowledge of the secret session ticket key
3348 to exploit, in which case you'd be SOL either way.)
3350 - Change bn_nist.c so that it will properly handle input BIGNUMs
3351 outside the expected range.
3353 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3356 [Neel Mehta, Bodo Moeller]
3358 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3359 the load fails. Useful for distros.
3360 [Ben Laurie and the FreeBSD team]
3362 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3365 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3368 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3370 This work was sponsored by Logica.
3373 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3374 keystores. Support for SSL/TLS client authentication too.
3375 Not compiled unless enable-capieng specified to Configure.
3377 This work was sponsored by Logica.
3380 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3381 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3382 attribute creation routines such as certifcate requests and PKCS#12
3386 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3388 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3389 handshake which could lead to a cilent crash as found using the
3390 Codenomicon TLS test suite (CVE-2008-1672)
3391 [Steve Henson, Mark Cox]
3393 *) Fix double free in TLS server name extensions which could lead to
3394 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3397 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3399 Clear the error queue to ensure that error entries left from
3400 older function calls do not interfere with the correct operation.
3401 [Lutz Jaenicke, Erik de Castro Lopo]
3403 *) Remove root CA certificates of commercial CAs:
3405 The OpenSSL project does not recommend any specific CA and does not
3406 have any policy with respect to including or excluding any CA.
3407 Therefore it does not make any sense to ship an arbitrary selection
3408 of root CA certificates with the OpenSSL software.
3411 *) RSA OAEP patches to fix two separate invalid memory reads.
3412 The first one involves inputs when 'lzero' is greater than
3413 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3414 before the beginning of from). The second one involves inputs where
3415 the 'db' section contains nothing but zeroes (there is a one-byte
3416 invalid read after the end of 'db').
3417 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3419 *) Partial backport from 0.9.9-dev:
3421 Introduce bn_mul_mont (dedicated Montgomery multiplication
3422 procedure) as a candidate for BIGNUM assembler implementation.
3423 While 0.9.9-dev uses assembler for various architectures, only
3424 x86_64 is available by default here in the 0.9.8 branch, and
3425 32-bit x86 is available through a compile-time setting.
3427 To try the 32-bit x86 assembler implementation, use Configure
3428 option "enable-montasm" (which exists only for this backport).
3430 As "enable-montasm" for 32-bit x86 disclaims code stability
3431 anyway, in this constellation we activate additional code
3432 backported from 0.9.9-dev for further performance improvements,
3433 namely BN_from_montgomery_word. (To enable this otherwise,
3434 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3436 [Andy Polyakov (backport partially by Bodo Moeller)]
3438 *) Add TLS session ticket callback. This allows an application to set
3439 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3440 values. This is useful for key rollover for example where several key
3441 sets may exist with different names.
3444 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3445 This was broken until now in 0.9.8 releases, such that the only way
3446 a registered ENGINE could be used (assuming it initialises
3447 successfully on the host) was to explicitly set it as the default
3448 for the relevant algorithms. This is in contradiction with 0.9.7
3449 behaviour and the documentation. With this fix, when an ENGINE is
3450 registered into a given algorithm's table of implementations, the
3451 'uptodate' flag is reset so that auto-discovery will be used next
3452 time a new context for that algorithm attempts to select an
3454 [Ian Lister (tweaked by Geoff Thorpe)]
3456 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3457 implemention in the following ways:
3459 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3462 Lack of BER streaming support means one pass streaming processing is
3463 only supported if data is detached: setting the streaming flag is
3464 ignored for embedded content.
3466 CMS support is disabled by default and must be explicitly enabled
3467 with the enable-cms configuration option.
3470 *) Update the GMP engine glue to do direct copies between BIGNUM and
3471 mpz_t when openssl and GMP use the same limb size. Otherwise the
3472 existing "conversion via a text string export" trick is still used.
3473 [Paul Sheer <paulsheer@gmail.com>]
3475 *) Zlib compression BIO. This is a filter BIO which compressed and
3476 uncompresses any data passed through it.
3479 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3480 RFC3394 compatible AES key wrapping.
3483 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3484 sets string data without copying. X509_ALGOR_set0() and
3485 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3486 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3487 from an X509_ATTRIBUTE structure optionally checking it occurs only
3488 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3492 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3493 to get the expected BN_FLG_CONSTTIME behavior.
3494 [Bodo Moeller (Google)]
3498 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3499 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3500 - added some more tests to do_tests.pl
3501 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3502 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3503 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3504 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3505 - various changes to netware.pl to enable gcc-cross builds on Win32
3507 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3508 - various changes to fix missing prototype warnings
3509 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3510 - added AES, WHIRLPOOL and CPUID assembler code to build files
3511 - added missing AES assembler make rules to mk1mf.pl
3512 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3513 [Guenter Knauf <eflash@gmx.net>]
3515 *) Implement certificate status request TLS extension defined in RFC3546.
3516 A client can set the appropriate parameters and receive the encoded
3517 OCSP response via a callback. A server can query the supplied parameters
3518 and set the encoded OCSP response in the callback. Add simplified examples
3519 to s_client and s_server.
3522 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3524 *) Fix various bugs:
3525 + Binary incompatibility of ssl_ctx_st structure
3526 + DTLS interoperation with non-compliant servers
3527 + Don't call get_session_cb() without proposed session
3528 + Fix ia64 assembler code
3529 [Andy Polyakov, Steve Henson]
3531 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3533 *) DTLS Handshake overhaul. There were longstanding issues with
3534 OpenSSL DTLS implementation, which were making it impossible for
3535 RFC 4347 compliant client to communicate with OpenSSL server.
3536 Unfortunately just fixing these incompatibilities would "cut off"
3537 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3538 server keeps tolerating non RFC compliant syntax. The opposite is
3539 not true, 0.9.8f client can not communicate with earlier server.
3540 This update even addresses CVE-2007-4995.
3543 *) Changes to avoid need for function casts in OpenSSL: some compilers
3544 (gcc 4.2 and later) reject their use.
3545 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3548 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3549 RFC4507bis. The encrypted ticket format is an encrypted encoded
3550 SSL_SESSION structure, that way new session features are automatically
3553 If a client application caches session in an SSL_SESSION structure
3554 support is transparent because tickets are now stored in the encoded
3557 The SSL_CTX structure automatically generates keys for ticket
3558 protection in servers so again support should be possible
3559 with no application modification.
3561 If a client or server wishes to disable RFC4507 support then the option
3562 SSL_OP_NO_TICKET can be set.
3564 Add a TLS extension debugging callback to allow the contents of any client
3565 or server extensions to be examined.
3567 This work was sponsored by Google.
3570 *) Add initial support for TLS extensions, specifically for the server_name
3571 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3572 have new members for a host name. The SSL data structure has an
3573 additional member SSL_CTX *initial_ctx so that new sessions can be
3574 stored in that context to allow for session resumption, even after the
3575 SSL has been switched to a new SSL_CTX in reaction to a client's
3576 server_name extension.
3578 New functions (subject to change):
3580 SSL_get_servername()
3581 SSL_get_servername_type()
3584 New CTRL codes and macros (subject to change):
3586 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3587 - SSL_CTX_set_tlsext_servername_callback()
3588 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3589 - SSL_CTX_set_tlsext_servername_arg()
3590 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3592 openssl s_client has a new '-servername ...' option.
3594 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3595 '-key2 ...', '-servername_fatal' (subject to change). This allows
3596 testing the HostName extension for a specific single host name ('-cert'
3597 and '-key' remain fallbacks for handshakes without HostName
3598 negotiation). If the unrecogninzed_name alert has to be sent, this by
3599 default is a warning; it becomes fatal with the '-servername_fatal'
3602 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3604 *) Add AES and SSE2 assembly language support to VC++ build.
3607 *) Mitigate attack on final subtraction in Montgomery reduction.
3610 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3611 (which previously caused an internal error).
3614 *) Squeeze another 10% out of IGE mode when in != out.
3617 *) AES IGE mode speedup.
3618 [Dean Gaudet (Google)]
3620 *) Add the Korean symmetric 128-bit cipher SEED (see
3621 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3622 add SEED ciphersuites from RFC 4162:
3624 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3625 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3626 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3627 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3629 To minimize changes between patchlevels in the OpenSSL 0.9.8
3630 series, SEED remains excluded from compilation unless OpenSSL
3631 is configured with 'enable-seed'.
3632 [KISA, Bodo Moeller]
3634 *) Mitigate branch prediction attacks, which can be practical if a
3635 single processor is shared, allowing a spy process to extract
3636 information. For detailed background information, see
3637 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3638 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3639 and Necessary Software Countermeasures"). The core of the change
3640 are new versions BN_div_no_branch() and
3641 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3642 respectively, which are slower, but avoid the security-relevant
3643 conditional branches. These are automatically called by BN_div()
3644 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3645 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3646 remove a conditional branch.
3648 BN_FLG_CONSTTIME is the new name for the previous
3649 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3650 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3651 in the exponent causes BN_mod_exp_mont() to use the alternative
3652 implementation in BN_mod_exp_mont_consttime().) The old name
3653 remains as a deprecated alias.
3655 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3656 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3657 constant-time implementations for more than just exponentiation.
3658 Here too the old name is kept as a deprecated alias.
3660 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3661 the BN_BLINDING structure gets an independent copy of the
3662 modulus. This means that the previous "BIGNUM *m" argument to
3663 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3664 essentially becomes "const BIGNUM *m", although we can't actually
3665 change this in the header file before 0.9.9. It allows
3666 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3667 enable BN_FLG_CONSTTIME.
3669 [Matthew D Wood (Intel Corp)]
3671 *) In the SSL/TLS server implementation, be strict about session ID
3672 context matching (which matters if an application uses a single
3673 external cache for different purposes). Previously,
3674 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3675 set. This did ensure strict client verification, but meant that,
3676 with applications using a single external cache for quite
3677 different requirements, clients could circumvent ciphersuite
3678 restrictions for a given session ID context by starting a session
3679 in a different context.
3682 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3683 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3684 authentication-only ciphersuites.
3687 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3688 not complete and could lead to a possible single byte overflow
3689 (CVE-2007-5135) [Ben Laurie]
3691 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3693 *) Since AES128 and AES256 (and similarly Camellia128 and
3694 Camellia256) share a single mask bit in the logic of
3695 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3696 kludge to work properly if AES128 is available and AES256 isn't
3697 (or if Camellia128 is available and Camellia256 isn't).
3700 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3701 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3702 When a point or a seed is encoded in a BIT STRING, we need to
3703 prevent the removal of trailing zero bits to get the proper DER
3704 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3705 of a NamedBitList, for which trailing 0 bits need to be removed.)
3708 *) Have SSL/TLS server implementation tolerate "mismatched" record
3709 protocol version while receiving ClientHello even if the
3710 ClientHello is fragmented. (The server can't insist on the
3711 particular protocol version it has chosen before the ServerHello
3712 message has informed the client about his choice.)
3715 *) Add RFC 3779 support.
3716 [Rob Austein for ARIN, Ben Laurie]
3718 *) Load error codes if they are not already present instead of using a
3719 static variable. This allows them to be cleanly unloaded and reloaded.
3720 Improve header file function name parsing.
3723 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3724 or CAPABILITY handshake as required by RFCs.
3727 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3729 *) Introduce limits to prevent malicious keys being able to
3730 cause a denial of service. (CVE-2006-2940)
3731 [Steve Henson, Bodo Moeller]
3733 *) Fix ASN.1 parsing of certain invalid structures that can result
3734 in a denial of service. (CVE-2006-2937) [Steve Henson]
3736 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3737 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3739 *) Fix SSL client code which could crash if connecting to a
3740 malicious SSLv2 server. (CVE-2006-4343)
3741 [Tavis Ormandy and Will Drewry, Google Security Team]
3743 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3744 match only those. Before that, "AES256-SHA" would be interpreted
3745 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3746 the same strength classification in 0.9.7h) as we currently only
3747 have a single AES bit in the ciphersuite description bitmap.
3748 That change, however, also applied to ciphersuite strings such as
3749 "RC4-MD5" that intentionally matched multiple ciphersuites --
3750 namely, SSL 2.0 ciphersuites in addition to the more common ones
3751 from SSL 3.0/TLS 1.0.
3753 So we change the selection algorithm again: Naming an explicit
3754 ciphersuite selects this one ciphersuite, and any other similar
3755 ciphersuite (same bitmap) from *other* protocol versions.
3756 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3757 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3759 Since SSL 2.0 does not have any ciphersuites for which the
3760 128/256 bit distinction would be relevant, this works for now.
3761 The proper fix will be to use different bits for AES128 and
3762 AES256, which would have avoided the problems from the beginning;
3763 however, bits are scarce, so we can only do this in a new release
3764 (not just a patchlevel) when we can change the SSL_CIPHER
3765 definition to split the single 'unsigned long mask' bitmap into
3766 multiple values to extend the available space.
3770 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3772 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3773 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3775 *) Add AES IGE and biIGE modes.
3778 *) Change the Unix randomness entropy gathering to use poll() when
3779 possible instead of select(), since the latter has some
3780 undesirable limitations.
3781 [Darryl Miles via Richard Levitte and Bodo Moeller]
3783 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3784 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3785 cannot be implicitly activated as part of, e.g., the "AES" alias.
3786 However, please upgrade to OpenSSL 0.9.9[-dev] for
3787 non-experimental use of the ECC ciphersuites to get TLS extension
3788 support, which is required for curve and point format negotiation
3789 to avoid potential handshake problems.
3792 *) Disable rogue ciphersuites:
3794 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3795 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3796 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3798 The latter two were purportedly from
3799 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3802 Also deactivate the remaining ciphersuites from
3803 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3804 unofficial, and the ID has long expired.
3807 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3808 dual-core machines) and other potential thread-safety issues.
3811 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3812 versions), which is now available for royalty-free use
3813 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3814 Also, add Camellia TLS ciphersuites from RFC 4132.
3816 To minimize changes between patchlevels in the OpenSSL 0.9.8
3817 series, Camellia remains excluded from compilation unless OpenSSL
3818 is configured with 'enable-camellia'.
3821 *) Disable the padding bug check when compression is in use. The padding
3822 bug check assumes the first packet is of even length, this is not
3823 necessarily true if compresssion is enabled and can result in false
3824 positives causing handshake failure. The actual bug test is ancient
3825 code so it is hoped that implementations will either have fixed it by
3826 now or any which still have the bug do not support compression.
3829 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3831 *) When applying a cipher rule check to see if string match is an explicit
3832 cipher suite and only match that one cipher suite if it is.
3835 *) Link in manifests for VC++ if needed.
3836 [Austin Ziegler <halostatue@gmail.com>]
3838 *) Update support for ECC-based TLS ciphersuites according to
3839 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3840 TLS extensions, which are supported starting with the 0.9.9
3841 branch, not in the OpenSSL 0.9.8 branch).
3844 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3845 opaque EVP_CIPHER_CTX handling.
3848 *) Fixes and enhancements to zlib compression code. We now only use
3849 "zlib1.dll" and use the default __cdecl calling convention on Win32
3850 to conform with the standards mentioned here:
3851 http://www.zlib.net/DLL_FAQ.txt
3852 Static zlib linking now works on Windows and the new --with-zlib-include
3853 --with-zlib-lib options to Configure can be used to supply the location
3854 of the headers and library. Gracefully handle case where zlib library
3858 *) Several fixes and enhancements to the OID generation code. The old code
3859 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3860 handle numbers larger than ULONG_MAX, truncated printing and had a
3861 non standard OBJ_obj2txt() behaviour.
3864 *) Add support for building of engines under engine/ as shared libraries
3865 under VC++ build system.
3868 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3869 Hopefully, we will not see any false combination of paths any more.
3872 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3874 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3875 (part of SSL_OP_ALL). This option used to disable the
3876 countermeasure against man-in-the-middle protocol-version
3877 rollback in the SSL 2.0 server implementation, which is a bad
3878 idea. (CVE-2005-2969)
3880 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3881 for Information Security, National Institute of Advanced Industrial
3882 Science and Technology [AIST], Japan)]
3884 *) Add two function to clear and return the verify parameter flags.
3887 *) Keep cipherlists sorted in the source instead of sorting them at
3888 runtime, thus removing the need for a lock.
3891 *) Avoid some small subgroup attacks in Diffie-Hellman.
3892 [Nick Mathewson and Ben Laurie]
3894 *) Add functions for well-known primes.
3897 *) Extended Windows CE support.
3898 [Satoshi Nakamura and Andy Polyakov]
3900 *) Initialize SSL_METHOD structures at compile time instead of during
3901 runtime, thus removing the need for a lock.
3904 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3905 attempting to decrypt each encrypted key in turn. Add support to
3909 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3911 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3914 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3917 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3918 key into the same file any more.
3921 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3924 *) Add -utf8 command line and config file option to 'ca'.
3925 [Stefan <stf@udoma.org]
3927 *) Removed the macro des_crypt(), as it seems to conflict with some
3928 libraries. Use DES_crypt().
3931 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3932 involves renaming the source and generated shared-libs for
3933 both. The engines will accept the corrected or legacy ids
3934 ('ncipher' and '4758_cca' respectively) when binding. NB,
3935 this only applies when building 'shared'.
3936 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3938 *) Add attribute functions to EVP_PKEY structure. Modify
3939 PKCS12_create() to recognize a CSP name attribute and
3940 use it. Make -CSP option work again in pkcs12 utility.
3943 *) Add new functionality to the bn blinding code:
3944 - automatic re-creation of the BN_BLINDING parameters after
3945 a fixed number of uses (currently 32)
3946 - add new function for parameter creation
3947 - introduce flags to control the update behaviour of the
3948 BN_BLINDING parameters
3949 - hide BN_BLINDING structure
3950 Add a second BN_BLINDING slot to the RSA structure to improve
3951 performance when a single RSA object is shared among several
3955 *) Add support for DTLS.
3956 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3958 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3959 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3962 *) Remove buggy and incompletet DH cert support from
3963 ssl/ssl_rsa.c and ssl/s3_both.c
3966 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3967 the apps/openssl applications.
3970 *) Compile clean with "-Wall -Wmissing-prototypes
3971 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3972 DEBUG_SAFESTACK must also be set.
3975 *) Change ./Configure so that certain algorithms can be disabled by default.
3976 The new counterpiece to "no-xxx" is "enable-xxx".
3978 The patented RC5 and MDC2 algorithms will now be disabled unless
3979 "enable-rc5" and "enable-mdc2", respectively, are specified.
3981 (IDEA remains enabled despite being patented. This is because IDEA
3982 is frequently required for interoperability, and there is no license
3983 fee for non-commercial use. As before, "no-idea" can be used to
3984 avoid this algorithm.)
3988 *) Add processing of proxy certificates (see RFC 3820). This work was
3989 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3990 EGEE (Enabling Grids for E-science in Europe).
3993 *) RC4 performance overhaul on modern architectures/implementations, such
3994 as Intel P4, IA-64 and AMD64.
3997 *) New utility extract-section.pl. This can be used specify an alternative
3998 section number in a pod file instead of having to treat each file as
3999 a separate case in Makefile. This can be done by adding two lines to the
4002 =for comment openssl_section:XXX
4004 The blank line is mandatory.
4008 *) New arguments -certform, -keyform and -pass for s_client and s_server
4009 to allow alternative format key and certificate files and passphrase
4013 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4014 update associated structures and add various utility functions.
4016 Add new policy related verify parameters, include policy checking in
4017 standard verify code. Enhance 'smime' application with extra parameters
4018 to support policy checking and print out.
4021 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4022 Nehemiah processors. These extensions support AES encryption in hardware
4023 as well as RNG (though RNG support is currently disabled).
4024 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4026 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4029 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4030 [Andy Polyakov and a number of other people]
4032 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4033 implementation contributed by IBM.
4034 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4036 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4037 exponent rather than 'unsigned long'. There is a corresponding change to
4038 the new 'rsa_keygen' element of the RSA_METHOD structure.
4039 [Jelte Jansen, Geoff Thorpe]
4041 *) Functionality for creating the initial serial number file is now
4042 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4044 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4045 number file to 1, which is bound to cause problems. To avoid
4046 the problems while respecting compatibility between different 0.9.7
4047 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4048 CA.pl for serial number initialization. With the new release 0.9.8,
4049 we can fix the problem directly in the 'ca' utility.)
4052 *) Reduced header interdepencies by declaring more opaque objects in
4053 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4054 give fewer recursive includes, which could break lazy source code - so
4055 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4056 developers should define this symbol when building and using openssl to
4057 ensure they track the recommended behaviour, interfaces, [etc], but
4058 backwards-compatible behaviour prevails when this isn't defined.
4061 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4064 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4065 This will generate a random key of the appropriate length based on the
4066 cipher context. The EVP_CIPHER can provide its own random key generation
4067 routine to support keys of a specific form. This is used in the des and
4068 3des routines to generate a key of the correct parity. Update S/MIME
4069 code to use new functions and hence generate correct parity DES keys.
4070 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4071 valid (weak or incorrect parity).
4074 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4075 as looking them up. This is useful when the verified structure may contain
4076 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4077 present unless the new PKCS7_NO_CRL flag is asserted.
4080 *) Extend ASN1 oid configuration module. It now additionally accepts the
4083 shortName = some long name, 1.2.3.4
4086 *) Reimplemented the BN_CTX implementation. There is now no more static
4087 limitation on the number of variables it can handle nor the depth of the
4088 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4089 information can now expand as required, and rather than having a single
4090 static array of bignums, BN_CTX now uses a linked-list of such arrays
4091 allowing it to expand on demand whilst maintaining the usefulness of
4092 BN_CTX's "bundling".
4095 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4096 to allow all RSA operations to function using a single BN_CTX.
4099 *) Preliminary support for certificate policy evaluation and checking. This
4100 is initially intended to pass the tests outlined in "Conformance Testing
4101 of Relying Party Client Certificate Path Processing Logic" v1.07.
4104 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4105 remained unused and not that useful. A variety of other little bignum
4106 tweaks and fixes have also been made continuing on from the audit (see
4110 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4111 associated ASN1, EVP and SSL functions and old ASN1 macros.
4114 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4115 and this should never fail. So the return value from the use of
4116 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4117 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4120 *) BN_CTX_get() should return zero-valued bignums, providing the same
4121 initialised value as BN_new().
4122 [Geoff Thorpe, suggested by Ulf Möller]
4124 *) Support for inhibitAnyPolicy certificate extension.
4127 *) An audit of the BIGNUM code is underway, for which debugging code is
4128 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4129 is considered valid when processing BIGNUMs, and causes execution to
4130 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4131 further steps are taken to deliberately pollute unused data in BIGNUM
4132 structures to try and expose faulty code further on. For now, openssl will
4133 (in its default mode of operation) continue to tolerate the inconsistent
4134 forms that it has tolerated in the past, but authors and packagers should
4135 consider trying openssl and their own applications when compiled with
4136 these debugging symbols defined. It will help highlight potential bugs in
4137 their own code, and will improve the test coverage for OpenSSL itself. At
4138 some point, these tighter rules will become openssl's default to improve
4139 maintainability, though the assert()s and other overheads will remain only
4140 in debugging configurations. See bn.h for more details.
4141 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4143 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4144 that can only be obtained through BN_CTX_new() (which implicitly
4145 initialises it). The presence of this function only made it possible
4146 to overwrite an existing structure (and cause memory leaks).
4149 *) Because of the callback-based approach for implementing LHASH as a
4150 template type, lh_insert() adds opaque objects to hash-tables and
4151 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4152 to clean up those corresponding objects before destroying the hash table
4153 (and losing the object pointers). So some over-zealous constifications in
4154 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4155 objects as "const" and the lh_doall[_arg] callback wrappers are not
4156 prototyped to have "const" restrictions on the object pointers they are
4157 given (and so aren't required to cast them away any more).
4160 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4161 (speed) prefers to use its own implementation. The two implementations
4162 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4163 its object type properly exposed (MS_TM) instead of casting to/from "char
4164 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4165 aren't necessarily the greatest nomenclatures - but this is what was used
4166 internally to the implementation so I've used that for now.
4169 *) Ensure that deprecated functions do not get compiled when
4170 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4171 the self-tests were still using deprecated key-generation functions so
4172 these have been updated also.
4175 *) Reorganise PKCS#7 code to separate the digest location functionality
4176 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
4177 New function PKCS7_set_digest() to set the digest type for PKCS#7
4178 digestedData type. Add additional code to correctly generate the
4179 digestedData type and add support for this type in PKCS7 initialization
4183 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4184 structure of type "other".
4187 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4188 sure the loop does correctly stop and breaking ("division by zero")
4189 modulus operations are not performed. The (pre-generated) prime
4190 table crypto/bn/bn_prime.h was already correct, but it could not be
4191 re-generated on some platforms because of the "division by zero"
4192 situation in the script.
4193 [Ralf S. Engelschall]
4195 *) Update support for ECC-based TLS ciphersuites according to
4196 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4197 SHA-1 now is only used for "small" curves (where the
4198 representation of a field element takes up to 24 bytes); for
4199 larger curves, the field element resulting from ECDH is directly
4200 used as premaster secret.
4201 [Douglas Stebila (Sun Microsystems Laboratories)]
4203 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4204 curve secp160r1 to the tests.
4205 [Douglas Stebila (Sun Microsystems Laboratories)]
4207 *) Add the possibility to load symbols globally with DSO.
4208 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4210 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4211 control of the error stack.
4214 *) Add support for STORE in ENGINE.
4217 *) Add the STORE type. The intention is to provide a common interface
4218 to certificate and key stores, be they simple file-based stores, or
4219 HSM-type store, or LDAP stores, or...
4220 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4223 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4224 pass a list of arguments to any function as well as provide a way
4225 for a function to pass data back to the caller.
4228 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4229 works like BUF_strdup() but can be used to duplicate a portion of
4230 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4234 *) Add the function sk_find_ex() which works like sk_find(), but will
4235 return an index to an element even if an exact match couldn't be
4236 found. The index is guaranteed to point at the element where the
4237 searched-for key would be inserted to preserve sorting order.
4240 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4241 takes an extra flags argument for optional functionality. Currently,
4242 the following flags are defined:
4244 OBJ_BSEARCH_VALUE_ON_NOMATCH
4245 This one gets OBJ_bsearch_ex() to return a pointer to the first
4246 element where the comparing function returns a negative or zero
4249 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4250 This one gets OBJ_bsearch_ex() to return a pointer to the first
4251 element where the comparing function returns zero. This is useful
4252 if there are more than one element where the comparing function
4256 *) Make it possible to create self-signed certificates with 'openssl ca'
4257 in such a way that the self-signed certificate becomes part of the
4258 CA database and uses the same mechanisms for serial number generation
4259 as all other certificate signing. The new flag '-selfsign' enables
4260 this functionality. Adapt CA.sh and CA.pl.in.
4263 *) Add functionality to check the public key of a certificate request
4264 against a given private. This is useful to check that a certificate
4265 request can be signed by that key (self-signing).
4268 *) Make it possible to have multiple active certificates with the same
4269 subject in the CA index file. This is done only if the keyword
4270 'unique_subject' is set to 'no' in the main CA section (default
4271 if 'CA_default') of the configuration file. The value is saved
4272 with the database itself in a separate index attribute file,
4273 named like the index file with '.attr' appended to the name.
4276 *) Generate muti valued AVAs using '+' notation in config files for
4280 *) Support for nameConstraints certificate extension.
4283 *) Support for policyConstraints certificate extension.
4286 *) Support for policyMappings certificate extension.
4289 *) Make sure the default DSA_METHOD implementation only uses its
4290 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4291 and change its own handlers to be NULL so as to remove unnecessary
4292 indirection. This lets alternative implementations fallback to the
4293 default implementation more easily.
4296 *) Support for directoryName in GeneralName related extensions
4300 *) Make it possible to link applications using Makefile.shared.
4301 Make that possible even when linking against static libraries!
4304 *) Support for single pass processing for S/MIME signing. This now
4305 means that S/MIME signing can be done from a pipe, in addition
4306 cleartext signing (multipart/signed type) is effectively streaming
4307 and the signed data does not need to be all held in memory.
4309 This is done with a new flag PKCS7_STREAM. When this flag is set
4310 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4311 is done after the data is output (and digests calculated) in
4312 SMIME_write_PKCS7().
4315 *) Add full support for -rpath/-R, both in shared libraries and
4316 applications, at least on the platforms where it's known how
4320 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4321 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4322 will now compute a table of multiples of the generator that
4323 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4324 faster (notably in the case of a single point multiplication,
4325 scalar * generator).
4326 [Nils Larsch, Bodo Moeller]
4328 *) IPv6 support for certificate extensions. The various extensions
4329 which use the IP:a.b.c.d can now take IPv6 addresses using the
4330 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4334 *) Added an ENGINE that implements RSA by performing private key
4335 exponentiations with the GMP library. The conversions to and from
4336 GMP's mpz_t format aren't optimised nor are any montgomery forms
4337 cached, and on x86 it appears OpenSSL's own performance has caught up.
4338 However there are likely to be other architectures where GMP could
4339 provide a boost. This ENGINE is not built in by default, but it can be
4340 specified at Configure time and should be accompanied by the necessary
4341 linker additions, eg;
4342 ./config -DOPENSSL_USE_GMP -lgmp
4345 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4346 testing availability of engines with "-t" - the old behaviour is
4347 produced by increasing the feature's verbosity with "-tt".
4350 *) ECDSA routines: under certain error conditions uninitialized BN objects
4351 could be freed. Solution: make sure initialization is performed early
4352 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4356 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4357 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4358 software implementations. For DSA and DH, parameter generation can
4359 also be overriden by providing the appropriate method callbacks.
4362 *) Change the "progress" mechanism used in key-generation and
4363 primality testing to functions that take a new BN_GENCB pointer in
4364 place of callback/argument pairs. The new API functions have "_ex"
4365 postfixes and the older functions are reimplemented as wrappers for
4366 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4367 declarations of the old functions to help (graceful) attempts to
4368 migrate to the new functions. Also, the new key-generation API
4369 functions operate on a caller-supplied key-structure and return
4370 success/failure rather than returning a key or NULL - this is to
4371 help make "keygen" another member function of RSA_METHOD etc.
4373 Example for using the new callback interface:
4375 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4379 BN_GENCB_set(&my_cb, my_callback, my_arg);
4381 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4382 /* For the meaning of a, b in calls to my_callback(), see the
4383 * documentation of the function that calls the callback.
4384 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4385 * my_callback should return 1 if it wants BN_is_prime_ex()
4386 * to continue, or 0 to stop.
4391 *) Change the ZLIB compression method to be stateful, and make it
4392 available to TLS with the number defined in
4393 draft-ietf-tls-compression-04.txt.
4396 *) Add the ASN.1 structures and functions for CertificatePair, which
4397 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4399 CertificatePair ::= SEQUENCE {
4400 forward [0] Certificate OPTIONAL,
4401 reverse [1] Certificate OPTIONAL,
4402 -- at least one of the pair shall be present -- }
4404 Also implement the PEM functions to read and write certificate
4405 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4407 This needed to be defined, mostly for the sake of the LDAP
4408 attribute crossCertificatePair, but may prove useful elsewhere as
4412 *) Make it possible to inhibit symlinking of shared libraries in
4413 Makefile.shared, for Cygwin's sake.
4416 *) Extend the BIGNUM API by creating a function
4417 void BN_set_negative(BIGNUM *a, int neg);
4418 and a macro that behave like
4419 int BN_is_negative(const BIGNUM *a);
4421 to avoid the need to access 'a->neg' directly in applications.
4424 *) Implement fast modular reduction for pseudo-Mersenne primes
4425 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4426 EC_GROUP_new_curve_GFp() will now automatically use this
4428 [Nils Larsch <nla@trustcenter.de>]
4430 *) Add new lock type (CRYPTO_LOCK_BN).
4433 *) Change the ENGINE framework to automatically load engines
4434 dynamically from specific directories unless they could be
4435 found to already be built in or loaded. Move all the
4436 current engines except for the cryptodev one to a new
4438 The engines in engines/ are built as shared libraries if
4439 the "shared" options was given to ./Configure or ./config.
4440 Otherwise, they are inserted in libcrypto.a.
4441 /usr/local/ssl/engines is the default directory for dynamic
4442 engines, but that can be overriden at configure time through
4443 the usual use of --prefix and/or --openssldir, and at run
4444 time with the environment variable OPENSSL_ENGINES.
4445 [Geoff Thorpe and Richard Levitte]
4447 *) Add Makefile.shared, a helper makefile to build shared
4448 libraries. Addapt Makefile.org.
4451 *) Add version info to Win32 DLLs.
4452 [Peter 'Luna' Runestig" <peter@runestig.com>]
4454 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4455 can be added using this API to created arbitrary PKCS#12
4456 files while avoiding the low level API.
4458 New options to PKCS12_create(), key or cert can be NULL and
4459 will then be omitted from the output file. The encryption
4460 algorithm NIDs can be set to -1 for no encryption, the mac
4461 iteration count can be set to 0 to omit the mac.
4463 Enhance pkcs12 utility by making the -nokeys and -nocerts
4464 options work when creating a PKCS#12 file. New option -nomac
4465 to omit the mac, NONE can be set for an encryption algorithm.
4466 New code is modified to use the enhanced PKCS12_create()
4467 instead of the low level API.
4470 *) Extend ASN1 encoder to support indefinite length constructed
4471 encoding. This can output sequences tags and octet strings in
4472 this form. Modify pk7_asn1.c to support indefinite length
4473 encoding. This is experimental and needs additional code to
4474 be useful, such as an ASN1 bio and some enhanced streaming
4477 Extend template encode functionality so that tagging is passed
4478 down to the template encoder.
4481 *) Let 'openssl req' fail if an argument to '-newkey' is not
4482 recognized instead of using RSA as a default.
4485 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4486 As these are not official, they are not included in "ALL";
4487 the "ECCdraft" ciphersuite group alias can be used to select them.
4488 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4490 *) Add ECDH engine support.
4491 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4493 *) Add ECDH in new directory crypto/ecdh/.
4494 [Douglas Stebila (Sun Microsystems Laboratories)]
4496 *) Let BN_rand_range() abort with an error after 100 iterations
4497 without success (which indicates a broken PRNG).
4500 *) Change BN_mod_sqrt() so that it verifies that the input value
4501 is really the square of the return value. (Previously,
4502 BN_mod_sqrt would show GIGO behaviour.)
4505 *) Add named elliptic curves over binary fields from X9.62, SECG,
4506 and WAP/WTLS; add OIDs that were still missing.
4508 [Sheueling Chang Shantz and Douglas Stebila
4509 (Sun Microsystems Laboratories)]
4511 *) Extend the EC library for elliptic curves over binary fields
4512 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4515 EC_GF2m_simple_method
4519 EC_GROUP_new_curve_GF2m
4520 EC_GROUP_set_curve_GF2m
4521 EC_GROUP_get_curve_GF2m
4522 EC_POINT_set_affine_coordinates_GF2m
4523 EC_POINT_get_affine_coordinates_GF2m
4524 EC_POINT_set_compressed_coordinates_GF2m
4526 Point compression for binary fields is disabled by default for
4527 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4530 As binary polynomials are represented as BIGNUMs, various members
4531 of the EC_GROUP and EC_POINT data structures can be shared
4532 between the implementations for prime fields and binary fields;
4533 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4534 are essentially identical to their ..._GFp counterparts.
4535 (For simplicity, the '..._GFp' prefix has been dropped from
4536 various internal method names.)
4538 An internal 'field_div' method (similar to 'field_mul' and
4539 'field_sqr') has been added; this is used only for binary fields.
4541 [Sheueling Chang Shantz and Douglas Stebila
4542 (Sun Microsystems Laboratories)]
4544 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4545 through methods ('mul', 'precompute_mult').
4547 The generic implementations (now internally called 'ec_wNAF_mul'
4548 and 'ec_wNAF_precomputed_mult') remain the default if these
4549 methods are undefined.
4551 [Sheueling Chang Shantz and Douglas Stebila
4552 (Sun Microsystems Laboratories)]
4554 *) New function EC_GROUP_get_degree, which is defined through
4555 EC_METHOD. For curves over prime fields, this returns the bit
4556 length of the modulus.
4558 [Sheueling Chang Shantz and Douglas Stebila
4559 (Sun Microsystems Laboratories)]
4561 *) New functions EC_GROUP_dup, EC_POINT_dup.
4562 (These simply call ..._new and ..._copy).
4564 [Sheueling Chang Shantz and Douglas Stebila
4565 (Sun Microsystems Laboratories)]
4567 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4568 Polynomials are represented as BIGNUMs (where the sign bit is not
4569 used) in the following functions [macros]:
4572 BN_GF2m_sub [= BN_GF2m_add]
4573 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4574 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4575 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4577 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4578 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4579 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4580 BN_GF2m_cmp [= BN_ucmp]
4582 (Note that only the 'mod' functions are actually for fields GF(2^m).
4583 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4585 For some functions, an the irreducible polynomial defining a
4586 field can be given as an 'unsigned int[]' with strictly
4587 decreasing elements giving the indices of those bits that are set;
4588 i.e., p[] represents the polynomial
4589 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4591 p[0] > p[1] > ... > p[k] = 0.
4592 This applies to the following functions:
4597 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4598 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4600 BN_GF2m_mod_sqrt_arr
4601 BN_GF2m_mod_solve_quad_arr
4605 Conversion can be performed by the following functions:
4610 bntest.c has additional tests for binary polynomial arithmetic.
4612 Two implementations for BN_GF2m_mod_div() are available.
4613 The default algorithm simply uses BN_GF2m_mod_inv() and
4614 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4615 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4616 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4618 [Sheueling Chang Shantz and Douglas Stebila
4619 (Sun Microsystems Laboratories)]
4621 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4622 functionality is disabled at compile-time.
4623 [Douglas Stebila <douglas.stebila@sun.com>]
4625 *) Change default behaviour of 'openssl asn1parse' so that more
4626 information is visible when viewing, e.g., a certificate:
4628 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4629 mode the content of non-printable OCTET STRINGs is output in a
4630 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4631 avoid the appearance of a printable string.
4632 [Nils Larsch <nla@trustcenter.de>]
4634 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4636 EC_GROUP_set_asn1_flag()
4637 EC_GROUP_get_asn1_flag()
4638 EC_GROUP_set_point_conversion_form()
4639 EC_GROUP_get_point_conversion_form()
4640 These control ASN1 encoding details:
4641 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4642 has been set to OPENSSL_EC_NAMED_CURVE.
4643 - Points are encoded in uncompressed form by default; options for
4644 asn1_for are as for point2oct, namely
4645 POINT_CONVERSION_COMPRESSED
4646 POINT_CONVERSION_UNCOMPRESSED
4647 POINT_CONVERSION_HYBRID
4649 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4652 EC_GROUP_get0_seed()
4653 EC_GROUP_get_seed_len()
4654 This is used only for ASN1 purposes (so far).
4655 [Nils Larsch <nla@trustcenter.de>]
4657 *) Add 'field_type' member to EC_METHOD, which holds the NID
4658 of the appropriate field type OID. The new function
4659 EC_METHOD_get_field_type() returns this value.
4660 [Nils Larsch <nla@trustcenter.de>]
4665 EC_POINT_point2hex()
4666 EC_POINT_hex2point()
4667 providing useful interfaces to EC_POINT_point2oct() and
4668 EC_POINT_oct2point().
4669 [Nils Larsch <nla@trustcenter.de>]
4671 *) Change internals of the EC library so that the functions
4672 EC_GROUP_set_generator()
4673 EC_GROUP_get_generator()
4674 EC_GROUP_get_order()
4675 EC_GROUP_get_cofactor()
4676 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4677 to methods, which would lead to unnecessary code duplication when
4678 adding different types of curves.
4679 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4681 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4682 arithmetic, and such that modified wNAFs are generated
4683 (which avoid length expansion in many cases).
4686 *) Add a function EC_GROUP_check_discriminant() (defined via
4687 EC_METHOD) that verifies that the curve discriminant is non-zero.
4689 Add a function EC_GROUP_check() that makes some sanity tests
4690 on a EC_GROUP, its generator and order. This includes
4691 EC_GROUP_check_discriminant().
4692 [Nils Larsch <nla@trustcenter.de>]
4694 *) Add ECDSA in new directory crypto/ecdsa/.
4696 Add applications 'openssl ecparam' and 'openssl ecdsa'
4697 (these are based on 'openssl dsaparam' and 'openssl dsa').
4699 ECDSA support is also included in various other files across the
4700 library. Most notably,
4701 - 'openssl req' now has a '-newkey ecdsa:file' option;
4702 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4703 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4704 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4705 them suitable for ECDSA where domain parameters must be
4706 extracted before the specific public key;
4707 - ECDSA engine support has been added.
4708 [Nils Larsch <nla@trustcenter.de>]
4710 *) Include some named elliptic curves, and add OIDs from X9.62,
4711 SECG, and WAP/WTLS. Each curve can be obtained from the new
4713 EC_GROUP_new_by_curve_name(),
4714 and the list of available named curves can be obtained with
4715 EC_get_builtin_curves().
4716 Also add a 'curve_name' member to EC_GROUP objects, which can be
4718 EC_GROUP_set_curve_name()
4719 EC_GROUP_get_curve_name()
4720 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4722 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4723 was actually never needed) and in BN_mul(). The removal in BN_mul()
4724 required a small change in bn_mul_part_recursive() and the addition
4725 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4726 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4727 bn_sub_words() and bn_add_words() except they take arrays with
4731 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4733 *) Cleanse PEM buffers before freeing them since they may contain
4735 [Benjamin Bennett <ben@psc.edu>]
4737 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4738 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4739 authentication-only ciphersuites.
4742 *) Since AES128 and AES256 share a single mask bit in the logic of
4743 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4744 kludge to work properly if AES128 is available and AES256 isn't.
4747 *) Expand security boundary to match 1.1.1 module.
4750 *) Remove redundant features: hash file source, editing of test vectors
4751 modify fipsld to use external fips_premain.c signature.
4754 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4755 run algorithm test programs.
4758 *) Make algorithm test programs more tolerant of whitespace.
4761 *) Have SSL/TLS server implementation tolerate "mismatched" record
4762 protocol version while receiving ClientHello even if the
4763 ClientHello is fragmented. (The server can't insist on the
4764 particular protocol version it has chosen before the ServerHello
4765 message has informed the client about his choice.)
4768 *) Load error codes if they are not already present instead of using a
4769 static variable. This allows them to be cleanly unloaded and reloaded.
4772 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4774 *) Introduce limits to prevent malicious keys being able to
4775 cause a denial of service. (CVE-2006-2940)
4776 [Steve Henson, Bodo Moeller]
4778 *) Fix ASN.1 parsing of certain invalid structures that can result
4779 in a denial of service. (CVE-2006-2937) [Steve Henson]
4781 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4782 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4784 *) Fix SSL client code which could crash if connecting to a
4785 malicious SSLv2 server. (CVE-2006-4343)
4786 [Tavis Ormandy and Will Drewry, Google Security Team]
4788 *) Change ciphersuite string processing so that an explicit
4789 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4790 will no longer include "AES128-SHA"), and any other similar
4791 ciphersuite (same bitmap) from *other* protocol versions (so that
4792 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4793 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4794 changes from 0.9.8b and 0.9.8d.
4797 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4799 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4800 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4802 *) Change the Unix randomness entropy gathering to use poll() when
4803 possible instead of select(), since the latter has some
4804 undesirable limitations.
4805 [Darryl Miles via Richard Levitte and Bodo Moeller]
4807 *) Disable rogue ciphersuites:
4809 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4810 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4811 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4813 The latter two were purportedly from
4814 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4817 Also deactive the remaining ciphersuites from
4818 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4819 unofficial, and the ID has long expired.
4822 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4823 dual-core machines) and other potential thread-safety issues.
4826 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4828 *) Adapt fipsld and the build system to link against the validated FIPS
4829 module in FIPS mode.
4832 *) Fixes for VC++ 2005 build under Windows.
4835 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4836 from a Windows bash shell such as MSYS. It is autodetected from the
4837 "config" script when run from a VC++ environment. Modify standard VC++
4838 build to use fipscanister.o from the GNU make build.
4841 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4843 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4844 The value now differs depending on if you build for FIPS or not.
4845 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4846 safely run with a non-FIPSed libcrypto, as it may crash because of
4847 the difference induced by this change.
4850 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4852 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4853 (part of SSL_OP_ALL). This option used to disable the
4854 countermeasure against man-in-the-middle protocol-version
4855 rollback in the SSL 2.0 server implementation, which is a bad
4856 idea. (CVE-2005-2969)
4858 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4859 for Information Security, National Institute of Advanced Industrial
4860 Science and Technology [AIST], Japan)]
4862 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4863 mainly for FIPS compliance and not fully integrated at this stage.
4866 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4867 the exponentiation using a fixed-length exponent. (Otherwise,
4868 the information leaked through timing could expose the secret key
4869 after many signatures; cf. Bleichenbacher's attack on DSA with
4873 *) Make a new fixed-window mod_exp implementation the default for
4874 RSA, DSA, and DH private-key operations so that the sequence of
4875 squares and multiplies and the memory access pattern are
4876 independent of the particular secret key. This will mitigate
4877 cache-timing and potential related attacks.
4879 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4880 and this is automatically used by BN_mod_exp_mont() if the new flag
4881 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4882 will use this BN flag for private exponents unless the flag
4883 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4884 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4886 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4888 *) Change the client implementation for SSLv23_method() and
4889 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4890 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4891 (Previously, the SSL 2.0 backwards compatible Client Hello
4892 message format would be used even with SSL_OP_NO_SSLv2.)
4895 *) Add support for smime-type MIME parameter in S/MIME messages which some
4899 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4900 a threadsafe manner. Modify rsa code to use new function and add calls
4901 to dsa and dh code (which had race conditions before).
4904 *) Include the fixed error library code in the C error file definitions
4905 instead of fixing them up at runtime. This keeps the error code
4906 structures constant.
4909 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4911 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4914 *) Fixes for newer kerberos headers. NB: the casts are needed because
4915 the 'length' field is signed on one version and unsigned on another
4916 with no (?) obvious way to tell the difference, without these VC++
4917 complains. Also the "definition" of FAR (blank) is no longer included
4918 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4919 some needed definitions.
4922 *) Undo Cygwin change.
4925 *) Added support for proxy certificates according to RFC 3820.
4926 Because they may be a security thread to unaware applications,
4927 they must be explicitely allowed in run-time. See
4928 docs/HOWTO/proxy_certificates.txt for further information.
4931 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4933 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4934 server and client random values. Previously
4935 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4936 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4938 This change has negligible security impact because:
4940 1. Server and client random values still have 24 bytes of pseudo random
4943 2. Server and client random values are sent in the clear in the initial
4946 3. The master secret is derived using the premaster secret (48 bytes in
4947 size for static RSA ciphersuites) as well as client server and random
4950 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4953 [Stephen Henson, reported by UK NISCC]
4955 *) Use Windows randomness collection on Cygwin.
4958 *) Fix hang in EGD/PRNGD query when communication socket is closed
4959 prematurely by EGD/PRNGD.
4960 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4962 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4965 *) Back-port of selected performance improvements from development
4966 branch, as well as improved support for PowerPC platforms.
4969 *) Add lots of checks for memory allocation failure, error codes to indicate
4970 failure and freeing up memory if a failure occurs.
4971 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4973 *) Add new -passin argument to dgst.
4976 *) Perform some character comparisons of different types in X509_NAME_cmp:
4977 this is needed for some certificates that reencode DNs into UTF8Strings
4978 (in violation of RFC3280) and can't or wont issue name rollover
4982 *) Make an explicit check during certificate validation to see that
4983 the CA setting in each certificate on the chain is correct. As a
4984 side effect always do the following basic checks on extensions,
4985 not just when there's an associated purpose to the check:
4987 - if there is an unhandled critical extension (unless the user
4988 has chosen to ignore this fault)
4989 - if the path length has been exceeded (if one is set at all)
4990 - that certain extensions fit the associated purpose (if one has
4994 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4996 *) Avoid a race condition when CRLs are checked in a multi threaded
4997 environment. This would happen due to the reordering of the revoked
4998 entries during signature checking and serial number lookup. Now the
4999 encoding is cached and the serial number sort performed under a lock.
5000 Add new STACK function sk_is_sorted().
5003 *) Add Delta CRL to the extension code.
5006 *) Various fixes to s3_pkt.c so alerts are sent properly.
5007 [David Holmes <d.holmes@f5.com>]
5009 *) Reduce the chances of duplicate issuer name and serial numbers (in
5010 violation of RFC3280) using the OpenSSL certificate creation utilities.
5011 This is done by creating a random 64 bit value for the initial serial
5012 number when a serial number file is created or when a self signed
5013 certificate is created using 'openssl req -x509'. The initial serial
5014 number file is created using 'openssl x509 -next_serial' in CA.pl
5015 rather than being initialized to 1.
5018 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5020 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5021 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5022 [Joe Orton, Steve Henson]
5024 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5026 [Joe Orton, Steve Henson]
5028 *) Make it possible to have multiple active certificates with the same
5029 subject in the CA index file. This is done only if the keyword
5030 'unique_subject' is set to 'no' in the main CA section (default
5031 if 'CA_default') of the configuration file. The value is saved
5032 with the database itself in a separate index attribute file,
5033 named like the index file with '.attr' appended to the name.
5036 *) X509 verify fixes. Disable broken certificate workarounds when
5037 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5038 keyUsage extension present. Don't accept CRLs with unhandled critical
5039 extensions: since verify currently doesn't process CRL extensions this
5040 rejects a CRL with *any* critical extensions. Add new verify error codes
5044 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5045 A clarification of RFC2560 will require the use of OCTET STRINGs and
5046 some implementations cannot handle the current raw format. Since OpenSSL
5047 copies and compares OCSP nonces as opaque blobs without any attempt at
5048 parsing them this should not create any compatibility issues.
5051 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5052 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5053 this HMAC (and other) operations are several times slower than OpenSSL
5057 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5058 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5060 *) Use the correct content when signing type "other".
5063 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5065 *) Fix various bugs revealed by running the NISCC test suite:
5067 Stop out of bounds reads in the ASN1 code when presented with
5068 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5070 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5072 If verify callback ignores invalid public key errors don't try to check
5073 certificate signature with the NULL public key.
5077 *) New -ignore_err option in ocsp application to stop the server
5078 exiting on the first error in a request.
5081 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5082 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5086 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5087 extra data after the compression methods not only for TLS 1.0
5088 but also for SSL 3.0 (as required by the specification).
5089 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5091 *) Change X509_certificate_type() to mark the key as exported/exportable
5092 when it's 512 *bits* long, not 512 bytes.
5095 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5096 blocks during encryption.
5099 *) Various fixes to base64 BIO and non blocking I/O. On write
5100 flushes were not handled properly if the BIO retried. On read
5101 data was not being buffered properly and had various logic bugs.
5102 This also affects blocking I/O when the data being decoded is a
5106 *) Various S/MIME bugfixes and compatibility changes:
5107 output correct application/pkcs7 MIME type if
5108 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5109 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5110 of files as .eml work). Correctly handle very long lines in MIME
5114 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5116 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5117 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5118 a protocol version number mismatch like a decryption error
5119 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5122 *) Turn on RSA blinding by default in the default implementation
5123 to avoid a timing attack. Applications that don't want it can call
5124 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5125 They would be ill-advised to do so in most cases.
5126 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5128 *) Change RSA blinding code so that it works when the PRNG is not
5129 seeded (in this case, the secret RSA exponent is abused as
5130 an unpredictable seed -- if it is not unpredictable, there
5131 is no point in blinding anyway). Make RSA blinding thread-safe
5132 by remembering the creator's thread ID in rsa->blinding and
5133 having all other threads use local one-time blinding factors
5134 (this requires more computation than sharing rsa->blinding, but
5135 avoids excessive locking; and if an RSA object is not shared
5136 between threads, blinding will still be very fast).
5139 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5140 ENGINE as defaults for all supported algorithms irrespective of
5141 the 'flags' parameter. 'flags' is now honoured, so applications
5142 should make sure they are passing it correctly.
5145 *) Target "mingw" now allows native Windows code to be generated in
5146 the Cygwin environment as well as with the MinGW compiler.
5149 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5151 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5152 via timing by performing a MAC computation even if incorrrect
5153 block cipher padding has been found. This is a countermeasure
5154 against active attacks where the attacker has to distinguish
5155 between bad padding and a MAC verification error. (CVE-2003-0078)
5157 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5158 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5159 Martin Vuagnoux (EPFL, Ilion)]
5161 *) Make the no-err option work as intended. The intention with no-err
5162 is not to have the whole error stack handling routines removed from
5163 libcrypto, it's only intended to remove all the function name and
5164 reason texts, thereby removing some of the footprint that may not
5165 be interesting if those errors aren't displayed anyway.
5167 NOTE: it's still possible for any application or module to have it's
5168 own set of error texts inserted. The routines are there, just not
5169 used by default when no-err is given.
5172 *) Add support for FreeBSD on IA64.
5173 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5175 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5176 Kerberos function mit_des_cbc_cksum(). Before this change,
5177 the value returned by DES_cbc_cksum() was like the one from
5178 mit_des_cbc_cksum(), except the bytes were swapped.
5179 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5181 *) Allow an application to disable the automatic SSL chain building.
5182 Before this a rather primitive chain build was always performed in
5183 ssl3_output_cert_chain(): an application had no way to send the
5184 correct chain if the automatic operation produced an incorrect result.
5186 Now the chain builder is disabled if either:
5188 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5190 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5192 The reasoning behind this is that an application would not want the
5193 auto chain building to take place if extra chain certificates are
5194 present and it might also want a means of sending no additional
5195 certificates (for example the chain has two certificates and the
5199 *) Add the possibility to build without the ENGINE framework.
5200 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5202 *) Under Win32 gmtime() can return NULL: check return value in
5203 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5206 *) DSA routines: under certain error conditions uninitialized BN objects
5207 could be freed. Solution: make sure initialization is performed early
5208 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5209 Nils Larsch <nla@trustcenter.de> via PR#459)
5212 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5213 checked on reconnect on the client side, therefore session resumption
5214 could still fail with a "ssl session id is different" error. This
5215 behaviour is masked when SSL_OP_ALL is used due to
5216 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5217 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5218 followup to PR #377.
5221 *) IA-32 assembler support enhancements: unified ELF targets, support
5222 for SCO/Caldera platforms, fix for Cygwin shared build.
5225 *) Add support for FreeBSD on sparc64. As a consequence, support for
5226 FreeBSD on non-x86 processors is separate from x86 processors on
5227 the config script, much like the NetBSD support.
5228 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5230 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5232 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5235 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5236 code (06) was taken as the first octet of the session ID and the last
5237 octet was ignored consequently. As a result SSLv2 client side session
5238 caching could not have worked due to the session ID mismatch between
5240 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5244 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5245 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5249 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5250 seems that in spite of existing for more than a year, many application
5251 author have done nothing to provide the necessary callbacks, which
5252 means that this particular engine will not work properly anywhere.
5253 This is a very unfortunate situation which forces us, in the name
5254 of usability, to give the hw_ncipher.c a static lock, which is part
5256 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5257 appear in 0.9.8 or later. We EXPECT application authors to have
5258 dealt properly with this when 0.9.8 is released (unless we actually
5259 make such changes in the libcrypto locking code that changes will
5260 have to be made anyway).
5263 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5264 octets have been read, EOF or an error occurs. Without this change
5265 some truncated ASN1 structures will not produce an error.
5268 *) Disable Heimdal support, since it hasn't been fully implemented.
5269 Still give the possibility to force the use of Heimdal, but with
5270 warnings and a request that patches get sent to openssl-dev.
5273 *) Add the VC-CE target, introduce the WINCE sysname, and add
5274 INSTALL.WCE and appropriate conditionals to make it build.
5275 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5277 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5278 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5279 edit numbers of the version.
5280 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5282 *) Introduce safe string copy and catenation functions
5283 (BUF_strlcpy() and BUF_strlcat()).
5284 [Ben Laurie (CHATS) and Richard Levitte]
5286 *) Avoid using fixed-size buffers for one-line DNs.
5287 [Ben Laurie (CHATS)]
5289 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5290 resizing buffers containing secrets, and use where appropriate.
5291 [Ben Laurie (CHATS)]
5293 *) Avoid using fixed size buffers for configuration file location.
5294 [Ben Laurie (CHATS)]
5296 *) Avoid filename truncation for various CA files.
5297 [Ben Laurie (CHATS)]
5299 *) Use sizeof in preference to magic numbers.
5300 [Ben Laurie (CHATS)]
5302 *) Avoid filename truncation in cert requests.
5303 [Ben Laurie (CHATS)]
5305 *) Add assertions to check for (supposedly impossible) buffer
5307 [Ben Laurie (CHATS)]
5309 *) Don't cache truncated DNS entries in the local cache (this could
5310 potentially lead to a spoofing attack).
5311 [Ben Laurie (CHATS)]
5313 *) Fix various buffers to be large enough for hex/decimal
5314 representations in a platform independent manner.
5315 [Ben Laurie (CHATS)]
5317 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5318 resizing buffers containing secrets, and use where appropriate.
5319 [Ben Laurie (CHATS)]
5321 *) Add BIO_indent() to avoid much slightly worrying code to do
5323 [Ben Laurie (CHATS)]
5325 *) Convert sprintf()/BIO_puts() to BIO_printf().
5326 [Ben Laurie (CHATS)]
5328 *) buffer_gets() could terminate with the buffer only half
5330 [Ben Laurie (CHATS)]
5332 *) Add assertions to prevent user-supplied crypto functions from
5333 overflowing internal buffers by having large block sizes, etc.
5334 [Ben Laurie (CHATS)]
5336 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5338 [Ben Laurie (CHATS)]
5340 *) Eliminate unused copy of key in RC4.
5341 [Ben Laurie (CHATS)]
5343 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5344 [Ben Laurie (CHATS)]
5346 *) Fix off-by-one error in EGD path.
5347 [Ben Laurie (CHATS)]
5349 *) If RANDFILE path is too long, ignore instead of truncating.
5350 [Ben Laurie (CHATS)]
5352 *) Eliminate unused and incorrectly sized X.509 structure
5354 [Ben Laurie (CHATS)]
5356 *) Eliminate unused and dangerous function knumber().
5357 [Ben Laurie (CHATS)]
5359 *) Eliminate unused and dangerous structure, KSSL_ERR.
5360 [Ben Laurie (CHATS)]
5362 *) Protect against overlong session ID context length in an encoded
5363 session object. Since these are local, this does not appear to be
5365 [Ben Laurie (CHATS)]
5367 *) Change from security patch (see 0.9.6e below) that did not affect
5368 the 0.9.6 release series:
5370 Remote buffer overflow in SSL3 protocol - an attacker could
5371 supply an oversized master key in Kerberos-enabled versions.
5373 [Ben Laurie (CHATS)]
5375 *) Change the SSL kerb5 codes to match RFC 2712.
5378 *) Make -nameopt work fully for req and add -reqopt switch.
5379 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5381 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5382 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5384 *) Make sure tests can be performed even if the corresponding algorithms
5385 have been removed entirely. This was also the last step to make
5386 OpenSSL compilable with DJGPP under all reasonable conditions.
5387 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5389 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5390 to allow version independent disabling of normally unselected ciphers,
5391 which may be activated as a side-effect of selecting a single cipher.
5393 (E.g., cipher list string "RSA" enables ciphersuites that are left
5394 out of "ALL" because they do not provide symmetric encryption.
5395 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5396 [Lutz Jaenicke, Bodo Moeller]
5398 *) Add appropriate support for separate platform-dependent build
5399 directories. The recommended way to make a platform-dependent
5400 build directory is the following (tested on Linux), maybe with
5403 # Place yourself outside of the OpenSSL source tree. In
5404 # this example, the environment variable OPENSSL_SOURCE
5405 # is assumed to contain the absolute OpenSSL source directory.
5406 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5407 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5408 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5409 mkdir -p `dirname $F`
5410 ln -s $OPENSSL_SOURCE/$F $F
5413 To be absolutely sure not to disturb the source tree, a "make clean"
5414 is a good thing. If it isn't successfull, don't worry about it,
5415 it probably means the source directory is very clean.
5418 *) Make sure any ENGINE control commands make local copies of string
5419 pointers passed to them whenever necessary. Otherwise it is possible
5420 the caller may have overwritten (or deallocated) the original string
5421 data when a later ENGINE operation tries to use the stored values.
5422 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5424 *) Improve diagnostics in file reading and command-line digests.
5425 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5427 *) Add AES modes CFB and OFB to the object database. Correct an
5428 error in AES-CFB decryption.
5431 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5432 allows existing EVP_CIPHER_CTX structures to be reused after
5433 calling EVP_*Final(). This behaviour is used by encryption
5434 BIOs and some applications. This has the side effect that
5435 applications must explicitly clean up cipher contexts with
5436 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5439 *) Check the values of dna and dnb in bn_mul_recursive before calling
5440 bn_mul_comba (a non zero value means the a or b arrays do not contain
5441 n2 elements) and fallback to bn_mul_normal if either is not zero.
5444 *) Fix escaping of non-ASCII characters when using the -subj option
5445 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5448 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5449 form for "surname", serialNumber has no short form.
5450 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5451 therefore remove "mail" short name for "internet 7".
5452 The OID for unique identifiers in X509 certificates is
5453 x500UniqueIdentifier, not uniqueIdentifier.
5454 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5457 *) Add an "init" command to the ENGINE config module and auto initialize
5458 ENGINEs. Without any "init" command the ENGINE will be initialized
5459 after all ctrl commands have been executed on it. If init=1 the
5460 ENGINE is initailized at that point (ctrls before that point are run
5461 on the uninitialized ENGINE and after on the initialized one). If
5462 init=0 then the ENGINE will not be iniatialized at all.
5465 *) Fix the 'app_verify_callback' interface so that the user-defined
5466 argument is actually passed to the callback: In the
5467 SSL_CTX_set_cert_verify_callback() prototype, the callback
5468 declaration has been changed from
5471 int (*cb)(X509_STORE_CTX *,void *);
5472 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5473 i=s->ctx->app_verify_callback(&ctx)
5474 has been changed into
5475 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5477 To update applications using SSL_CTX_set_cert_verify_callback(),
5478 a dummy argument can be added to their callback functions.
5479 [D. K. Smetters <smetters@parc.xerox.com>]
5481 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5482 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5484 *) Add and OPENSSL_LOAD_CONF define which will cause
5485 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5486 This allows older applications to transparently support certain
5487 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5488 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5489 load the config file and OPENSSL_add_all_algorithms_conf() which will
5490 always load it have also been added.
5493 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5494 Adjust NIDs and EVP layer.
5495 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5497 *) Config modules support in openssl utility.
5499 Most commands now load modules from the config file,
5500 though in a few (such as version) this isn't done
5501 because it couldn't be used for anything.
5503 In the case of ca and req the config file used is
5504 the same as the utility itself: that is the -config
5505 command line option can be used to specify an
5509 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5510 use "openssl_conf" if filename is NULL use default openssl config file.
5513 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5514 config section name. Add a new flag to tolerate a missing config file
5515 and move code to CONF_modules_load_file().
5518 *) Support for crypto accelerator cards from Accelerated Encryption
5519 Processing, www.aep.ie. (Use engine 'aep')
5520 The support was copied from 0.9.6c [engine] and adapted/corrected
5521 to work with the new engine framework.
5522 [AEP Inc. and Richard Levitte]
5524 *) Support for SureWare crypto accelerator cards from Baltimore
5525 Technologies. (Use engine 'sureware')
5526 The support was copied from 0.9.6c [engine] and adapted
5527 to work with the new engine framework.
5530 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5531 make the newer ENGINE framework commands for the CHIL engine work.
5532 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5534 *) Make it possible to produce shared libraries on ReliantUNIX.
5535 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5537 *) Add the configuration target debug-linux-ppro.
5538 Make 'openssl rsa' use the general key loading routines
5539 implemented in apps.c, and make those routines able to
5540 handle the key format FORMAT_NETSCAPE and the variant
5542 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5544 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5545 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5547 *) Add -keyform to rsautl, and document -engine.
5548 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5550 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5551 BIO_R_NO_SUCH_FILE error code rather than the generic
5552 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5555 *) Add new functions
5557 ERR_peek_last_error_line
5558 ERR_peek_last_error_line_data.
5559 These are similar to
5562 ERR_peek_error_line_data,
5563 but report on the latest error recorded rather than the first one
5564 still in the error queue.
5565 [Ben Laurie, Bodo Moeller]
5567 *) default_algorithms option in ENGINE config module. This allows things
5569 default_algorithms = ALL
5570 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5573 *) Prelminary ENGINE config module.
5576 *) New experimental application configuration code.
5579 *) Change the AES code to follow the same name structure as all other
5580 symmetric ciphers, and behave the same way. Move everything to
5581 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5582 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5584 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5585 [Ben Laurie and Theo de Raadt]
5587 *) Add option to output public keys in req command.
5588 [Massimiliano Pala madwolf@openca.org]
5590 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5591 (up to about 10% better than before for P-192 and P-224).
5594 *) New functions/macros
5596 SSL_CTX_set_msg_callback(ctx, cb)
5597 SSL_CTX_set_msg_callback_arg(ctx, arg)
5598 SSL_set_msg_callback(ssl, cb)
5599 SSL_set_msg_callback_arg(ssl, arg)
5601 to request calling a callback function
5603 void cb(int write_p, int version, int content_type,
5604 const void *buf, size_t len, SSL *ssl, void *arg)
5606 whenever a protocol message has been completely received
5607 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5608 protocol version according to which the SSL library interprets
5609 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5610 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5611 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5612 specification (change_cipher_spec(20), alert(21), handshake(22)).
5613 'buf' and 'len' point to the actual message, 'ssl' to the
5614 SSL object, and 'arg' is the application-defined value set by
5615 SSL[_CTX]_set_msg_callback_arg().
5617 'openssl s_client' and 'openssl s_server' have new '-msg' options
5618 to enable a callback that displays all protocol messages.
5621 *) Change the shared library support so shared libraries are built as
5622 soon as the corresponding static library is finished, and thereby get
5623 openssl and the test programs linked against the shared library.
5624 This still only happens when the keyword "shard" has been given to
5625 the configuration scripts.
5627 NOTE: shared library support is still an experimental thing, and
5628 backward binary compatibility is still not guaranteed.
5629 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5631 *) Add support for Subject Information Access extension.
5632 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5634 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5635 additional bytes when new memory had to be allocated, not just
5636 when reusing an existing buffer.
5639 *) New command line and configuration option 'utf8' for the req command.
5640 This allows field values to be specified as UTF8 strings.
5643 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5644 runs for the former and machine-readable output for the latter.
5647 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5648 of the e-mail address in the DN (i.e., it will go into a certificate
5649 extension only). The new configuration file option 'email_in_dn = no'
5650 has the same effect.
5651 [Massimiliano Pala madwolf@openca.org]
5653 *) Change all functions with names starting with des_ to be starting
5654 with DES_ instead. Add wrappers that are compatible with libdes,
5655 but are named _ossl_old_des_*. Finally, add macros that map the
5656 des_* symbols to the corresponding _ossl_old_des_* if libdes
5657 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5658 desired, the des_* symbols will be mapped to DES_*, with one
5661 Since we provide two compatibility mappings, the user needs to
5662 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5663 compatibility is desired. The default (i.e., when that macro
5664 isn't defined) is OpenSSL 0.9.6c compatibility.
5666 There are also macros that enable and disable the support of old
5667 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5668 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5669 are defined, the default will apply: to support the old des routines.
5671 In either case, one must include openssl/des.h to get the correct
5672 definitions. Do not try to just include openssl/des_old.h, that
5675 NOTE: This is a major break of an old API into a new one. Software
5676 authors are encouraged to switch to the DES_ style functions. Some
5677 time in the future, des_old.h and the libdes compatibility functions
5678 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5679 default), and then completely removed.
5682 *) Test for certificates which contain unsupported critical extensions.
5683 If such a certificate is found during a verify operation it is
5684 rejected by default: this behaviour can be overridden by either
5685 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5686 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5687 X509_supported_extension() has also been added which returns 1 if a
5688 particular extension is supported.
5691 *) Modify the behaviour of EVP cipher functions in similar way to digests
5692 to retain compatibility with existing code.
5695 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5696 compatibility with existing code. In particular the 'ctx' parameter does
5697 not have to be to be initialized before the call to EVP_DigestInit() and
5698 it is tidied up after a call to EVP_DigestFinal(). New function
5699 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5700 EVP_MD_CTX_copy() changed to not require the destination to be
5701 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5702 requires the destination to be valid.
5704 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5705 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5708 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5709 so that complete 'Handshake' protocol structures are kept in memory
5710 instead of overwriting 'msg_type' and 'length' with 'body' data.
5713 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5714 [Massimo Santin via Richard Levitte]
5716 *) Major restructuring to the underlying ENGINE code. This includes
5717 reduction of linker bloat, separation of pure "ENGINE" manipulation
5718 (initialisation, etc) from functionality dealing with implementations
5719 of specific crypto iterfaces. This change also introduces integrated
5720 support for symmetric ciphers and digest implementations - so ENGINEs
5721 can now accelerate these by providing EVP_CIPHER and EVP_MD
5722 implementations of their own. This is detailed in crypto/engine/README
5723 as it couldn't be adequately described here. However, there are a few
5724 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5725 were changed in the original introduction of ENGINE code have now
5726 reverted back - the hooking from this code to ENGINE is now a good
5727 deal more passive and at run-time, operations deal directly with
5728 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5729 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5730 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5731 they were not being used by the framework as there is no concept of a
5732 BIGNUM_METHOD and they could not be generalised to the new
5733 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5734 ENGINE_cpy() has been removed as it cannot be consistently defined in
5738 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5741 *) Change mkdef.pl to sort symbols that get the same entry number,
5742 and make sure the automatically generated functions ERR_load_*
5743 become part of libeay.num as well.
5746 *) New function SSL_renegotiate_pending(). This returns true once
5747 renegotiation has been requested (either SSL_renegotiate() call
5748 or HelloRequest/ClientHello receveived from the peer) and becomes
5749 false once a handshake has been completed.
5750 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5751 sends a HelloRequest, but does not ensure that a handshake takes
5752 place. SSL_renegotiate_pending() is useful for checking if the
5753 client has followed the request.)
5756 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5757 By default, clients may request session resumption even during
5758 renegotiation (if session ID contexts permit); with this option,
5759 session resumption is possible only in the first handshake.
5761 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5762 more bits available for options that should not be part of
5763 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5766 *) Add some demos for certificate and certificate request creation.
5769 *) Make maximum certificate chain size accepted from the peer application
5770 settable (SSL*_get/set_max_cert_list()), as proposed by
5771 "Douglas E. Engert" <deengert@anl.gov>.
5774 *) Add support for shared libraries for Unixware-7
5775 (Boyd Lynn Gerber <gerberb@zenez.com>).
5778 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5779 be done prior to destruction. Use this to unload error strings from
5780 ENGINEs that load their own error strings. NB: This adds two new API
5781 functions to "get" and "set" this destroy handler in an ENGINE.
5784 *) Alter all existing ENGINE implementations (except "openssl" and
5785 "openbsd") to dynamically instantiate their own error strings. This
5786 makes them more flexible to be built both as statically-linked ENGINEs
5787 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5788 Also, add stub code to each that makes building them as self-contained
5789 shared-libraries easier (see README.ENGINE).
5792 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5793 implementations into applications that are completely implemented in
5794 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5795 commands that can be used to configure what shared-library to load and
5796 to control aspects of the way it is handled. Also, made an update to
5797 the README.ENGINE file that brings its information up-to-date and
5798 provides some information and instructions on the "dynamic" ENGINE
5799 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5802 *) Make it possible to unload ranges of ERR strings with a new
5803 "ERR_unload_strings" function.
5806 *) Add a copy() function to EVP_MD.
5809 *) Make EVP_MD routines take a context pointer instead of just the
5810 md_data void pointer.
5813 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5814 that the digest can only process a single chunk of data
5815 (typically because it is provided by a piece of
5816 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5817 is only going to provide a single chunk of data, and hence the
5818 framework needn't accumulate the data for oneshot drivers.
5821 *) As with "ERR", make it possible to replace the underlying "ex_data"
5822 functions. This change also alters the storage and management of global
5823 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5824 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5825 index counters. The API functions that use this state have been changed
5826 to take a "class_index" rather than pointers to the class's local STACK
5827 and counter, and there is now an API function to dynamically create new
5828 classes. This centralisation allows us to (a) plug a lot of the
5829 thread-safety problems that existed, and (b) makes it possible to clean
5830 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5831 such data would previously have always leaked in application code and
5832 workarounds were in place to make the memory debugging turn a blind eye
5833 to it. Application code that doesn't use this new function will still
5834 leak as before, but their memory debugging output will announce it now
5835 rather than letting it slide.
5837 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5838 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5839 has a return value to indicate success or failure.
5842 *) Make it possible to replace the underlying "ERR" functions such that the
5843 global state (2 LHASH tables and 2 locks) is only used by the "default"
5844 implementation. This change also adds two functions to "get" and "set"
5845 the implementation prior to it being automatically set the first time
5846 any other ERR function takes place. Ie. an application can call "get",
5847 pass the return value to a module it has just loaded, and that module
5848 can call its own "set" function using that value. This means the
5849 module's "ERR" operations will use (and modify) the error state in the
5850 application and not in its own statically linked copy of OpenSSL code.
5853 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5854 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5855 the operation, and provides a more encapsulated way for external code
5856 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5857 to use these functions rather than manually incrementing the counts.
5859 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5862 *) Add EVP test program.
5865 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5868 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5869 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5870 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5871 These allow a CRL to be built without having to access X509_CRL fields
5872 directly. Modify 'ca' application to use new functions.
5875 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5876 bug workarounds. Rollback attack detection is a security feature.
5877 The problem will only arise on OpenSSL servers when TLSv1 is not
5878 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5879 Software authors not wanting to support TLSv1 will have special reasons
5880 for their choice and can explicitly enable this option.
5881 [Bodo Moeller, Lutz Jaenicke]
5883 *) Rationalise EVP so it can be extended: don't include a union of
5884 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5885 (similar to those existing for EVP_CIPHER_CTX).
5890 EVP_MD_CTX_init(&md); /* new function call */
5891 EVP_DigestInit(&md, EVP_sha1());
5892 EVP_DigestUpdate(&md, in, len);
5893 EVP_DigestFinal(&md, out, NULL);
5894 EVP_MD_CTX_cleanup(&md); /* new function call */
5898 *) Make DES key schedule conform to the usual scheme, as well as
5899 correcting its structure. This means that calls to DES functions
5900 now have to pass a pointer to a des_key_schedule instead of a
5901 plain des_key_schedule (which was actually always a pointer
5904 des_key_schedule ks;
5906 des_set_key_checked(..., &ks);
5907 des_ncbc_encrypt(..., &ks, ...);
5909 (Note that a later change renames 'des_...' into 'DES_...'.)
5912 *) Initial reduction of linker bloat: the use of some functions, such as
5913 PEM causes large amounts of unused functions to be linked in due to
5914 poor organisation. For example pem_all.c contains every PEM function
5915 which has a knock on effect of linking in large amounts of (unused)
5916 ASN1 code. Grouping together similar functions and splitting unrelated
5917 functions prevents this.
5920 *) Cleanup of EVP macros.
5923 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5924 correct _ecb suffix.
5927 *) Add initial OCSP responder support to ocsp application. The
5928 revocation information is handled using the text based index
5929 use by the ca application. The responder can either handle
5930 requests generated internally, supplied in files (for example
5931 via a CGI script) or using an internal minimal server.
5934 *) Add configuration choices to get zlib compression for TLS.
5937 *) Changes to Kerberos SSL for RFC 2712 compliance:
5938 1. Implemented real KerberosWrapper, instead of just using
5939 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5940 2. Implemented optional authenticator field of KerberosWrapper.
5942 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5943 and authenticator structs; see crypto/krb5/.
5945 Generalized Kerberos calls to support multiple Kerberos libraries.
5946 [Vern Staats <staatsvr@asc.hpc.mil>,
5947 Jeffrey Altman <jaltman@columbia.edu>
5948 via Richard Levitte]
5950 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5951 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5952 values for each of the key sizes rather than having just
5953 parameters (and 'speed' generating keys each time).
5956 *) Speed up EVP routines.
5959 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5960 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5961 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5962 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5964 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5965 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5966 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5969 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5971 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5974 *) Added the OS2-EMX target.
5975 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5977 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5978 to support NCONF routines in extension code. New function CONF_set_nconf()
5979 to allow functions which take an NCONF to also handle the old LHASH
5980 structure: this means that the old CONF compatible routines can be
5981 retained (in particular wrt extensions) without having to duplicate the
5982 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5985 *) Enhance the general user interface with mechanisms for inner control
5986 and with possibilities to have yes/no kind of prompts.
5989 *) Change all calls to low level digest routines in the library and
5990 applications to use EVP. Add missing calls to HMAC_cleanup() and
5991 don't assume HMAC_CTX can be copied using memcpy().
5992 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5994 *) Add the possibility to control engines through control names but with
5995 arbitrary arguments instead of just a string.
5996 Change the key loaders to take a UI_METHOD instead of a callback
5997 function pointer. NOTE: this breaks binary compatibility with earlier
5998 versions of OpenSSL [engine].
5999 Adapt the nCipher code for these new conditions and add a card insertion
6003 *) Enhance the general user interface with mechanisms to better support
6004 dialog box interfaces, application-defined prompts, the possibility
6005 to use defaults (for example default passwords from somewhere else)
6006 and interrupts/cancellations.
6009 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6010 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6013 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6014 tidy up some unnecessarily weird code in 'sk_new()').
6015 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6017 *) Change the key loading routines for ENGINEs to use the same kind
6018 callback (pem_password_cb) as all other routines that need this
6022 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6023 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6024 than this minimum value is recommended.
6027 *) New random seeder for OpenVMS, using the system process statistics
6028 that are easily reachable.
6031 *) Windows apparently can't transparently handle global
6032 variables defined in DLLs. Initialisations such as:
6034 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6036 wont compile. This is used by the any applications that need to
6037 declare their own ASN1 modules. This was fixed by adding the option
6038 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6039 needed for static libraries under Win32.
6042 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6043 setting of purpose and trust fields. New X509_STORE trust and
6044 purpose functions and tidy up setting in other SSL functions.
6047 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6048 structure. These are inherited by X509_STORE_CTX when it is
6049 initialised. This allows various defaults to be set in the
6050 X509_STORE structure (such as flags for CRL checking and custom
6051 purpose or trust settings) for functions which only use X509_STORE_CTX
6052 internally such as S/MIME.
6054 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6055 trust settings if they are not set in X509_STORE. This allows X509_STORE
6056 purposes and trust (in S/MIME for example) to override any set by default.
6058 Add command line options for CRL checking to smime, s_client and s_server
6062 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6063 are set then the CRL is looked up in the X509_STORE structure and
6064 its validity and signature checked, then if the certificate is found
6065 in the CRL the verify fails with a revoked error.
6067 Various new CRL related callbacks added to X509_STORE_CTX structure.
6069 Command line options added to 'verify' application to support this.
6071 This needs some additional work, such as being able to handle multiple
6072 CRLs with different times, extension based lookup (rather than just
6073 by subject name) and ultimately more complete V2 CRL extension
6077 *) Add a general user interface API (crypto/ui/). This is designed
6078 to replace things like des_read_password and friends (backward
6079 compatibility functions using this new API are provided).
6080 The purpose is to remove prompting functions from the DES code
6081 section as well as provide for prompting through dialog boxes in
6082 a window system and the like.
6085 *) Add "ex_data" support to ENGINE so implementations can add state at a
6086 per-structure level rather than having to store it globally.
6089 *) Make it possible for ENGINE structures to be copied when retrieved by
6090 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6091 This causes the "original" ENGINE structure to act like a template,
6092 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6093 operational state can be localised to each ENGINE structure, despite the
6094 fact they all share the same "methods". New ENGINE structures returned in
6095 this case have no functional references and the return value is the single
6096 structural reference. This matches the single structural reference returned
6097 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6101 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6102 needs to match any other type at all we need to manually clear the
6106 *) Changes to the "openssl engine" utility to include;
6107 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6108 about an ENGINE's available control commands.
6109 - executing control commands from command line arguments using the
6110 '-pre' and '-post' switches. '-post' is only used if '-t' is
6111 specified and the ENGINE is successfully initialised. The syntax for
6112 the individual commands are colon-separated, for example;
6113 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6116 *) New dynamic control command support for ENGINEs. ENGINEs can now
6117 declare their own commands (numbers), names (strings), descriptions,
6118 and input types for run-time discovery by calling applications. A
6119 subset of these commands are implicitly classed as "executable"
6120 depending on their input type, and only these can be invoked through
6121 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6122 can be based on user input, config files, etc). The distinction is
6123 that "executable" commands cannot return anything other than a boolean
6124 result and can only support numeric or string input, whereas some
6125 discoverable commands may only be for direct use through
6126 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6127 pointers, or other custom uses. The "executable" commands are to
6128 support parameterisations of ENGINE behaviour that can be
6129 unambiguously defined by ENGINEs and used consistently across any
6130 OpenSSL-based application. Commands have been added to all the
6131 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6132 control over shared-library paths without source code alterations.
6135 *) Changed all ENGINE implementations to dynamically allocate their
6136 ENGINEs rather than declaring them statically. Apart from this being
6137 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6138 this also allows the implementations to compile without using the
6139 internal engine_int.h header.
6142 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6143 'const' value. Any code that should be able to modify a RAND_METHOD
6144 should already have non-const pointers to it (ie. they should only
6145 modify their own ones).
6148 *) Made a variety of little tweaks to the ENGINE code.
6149 - "atalla" and "ubsec" string definitions were moved from header files
6150 to C code. "nuron" string definitions were placed in variables
6151 rather than hard-coded - allowing parameterisation of these values
6152 later on via ctrl() commands.
6153 - Removed unused "#if 0"'d code.
6154 - Fixed engine list iteration code so it uses ENGINE_free() to release
6155 structural references.
6156 - Constified the RAND_METHOD element of ENGINE structures.
6157 - Constified various get/set functions as appropriate and added
6158 missing functions (including a catch-all ENGINE_cpy that duplicates
6159 all ENGINE values onto a new ENGINE except reference counts/state).
6160 - Removed NULL parameter checks in get/set functions. Setting a method
6161 or function to NULL is a way of cancelling out a previously set
6162 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6163 and doesn't justify the extra error symbols and code.
6164 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6165 flags from engine_int.h to engine.h.
6166 - Changed prototypes for ENGINE handler functions (init(), finish(),
6167 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6170 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6171 to the algorithm using long division. The binary algorithm can be
6172 used only if the modulus is odd. On 32-bit systems, it is faster
6173 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6174 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6175 up to 450 bits. In 64-bit environments, the binary algorithm
6176 appears to be advantageous for much longer moduli; here we use it
6177 for moduli up to 2048 bits.
6180 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6181 could not support the combine flag in choice fields.
6184 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6185 extensions from a certificate request to the certificate.
6188 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6189 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6190 file: this allows the display of the certificate about to be
6191 signed to be customised, to allow certain fields to be included
6192 or excluded and extension details. The old system didn't display
6193 multicharacter strings properly, omitted fields not in the policy
6194 and couldn't display additional details such as extensions.
6197 *) Function EC_POINTs_mul for multiple scalar multiplication
6198 of an arbitrary number of elliptic curve points
6199 \sum scalars[i]*points[i],
6200 optionally including the generator defined for the EC_GROUP:
6201 scalar*generator + \sum scalars[i]*points[i].
6203 EC_POINT_mul is a simple wrapper function for the typical case
6204 that the point list has just one item (besides the optional
6208 *) First EC_METHODs for curves over GF(p):
6210 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6211 operations and provides various method functions that can also
6212 operate with faster implementations of modular arithmetic.
6214 EC_GFp_mont_method() reuses most functions that are part of
6215 EC_GFp_simple_method, but uses Montgomery arithmetic.
6217 [Bodo Moeller; point addition and point doubling
6218 implementation directly derived from source code provided by
6219 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6221 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6222 crypto/ec/ec_lib.c):
6224 Curves are EC_GROUP objects (with an optional group generator)
6225 based on EC_METHODs that are built into the library.
6227 Points are EC_POINT objects based on EC_GROUP objects.
6229 Most of the framework would be able to handle curves over arbitrary
6230 finite fields, but as there are no obvious types for fields other
6231 than GF(p), some functions are limited to that for now.
6234 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6235 that the file contains a complete HTTP response.
6238 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6239 change the def and num file printf format specifier from "%-40sXXX"
6240 to "%-39s XXX". The latter will always guarantee a space after the
6241 field while the former will cause them to run together if the field
6242 is 40 of more characters long.
6245 *) Constify the cipher and digest 'method' functions and structures
6246 and modify related functions to take constant EVP_MD and EVP_CIPHER
6250 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6251 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6254 *) Modify EVP_Digest*() routines so they now return values. Although the
6255 internal software routines can never fail additional hardware versions
6259 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6261 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6262 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6265 ERR_R_NESTED_ASN1_ERROR
6267 ERR_R_MISSING_ASN1_EOS
6268 were 4 .. 9, conflicting with
6269 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6271 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6272 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6274 Add new error code 'ERR_R_INTERNAL_ERROR'.
6277 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6281 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6282 sets the subject name for a new request or supersedes the
6283 subject name in a given request. Formats that can be parsed are
6284 'CN=Some Name, OU=myOU, C=IT'
6286 'CN=Some Name/OU=myOU/C=IT'.
6288 Add options '-batch' and '-verbose' to 'openssl req'.
6289 [Massimiliano Pala <madwolf@hackmasters.net>]
6291 *) Introduce the possibility to access global variables through
6292 functions on platform were that's the best way to handle exporting
6293 global variables in shared libraries. To enable this functionality,
6294 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6295 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6296 is normally done by Configure or something similar).
6298 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6299 in the source file (foo.c) like this:
6301 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6302 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6304 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6305 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6307 OPENSSL_DECLARE_GLOBAL(int,foo);
6308 #define foo OPENSSL_GLOBAL_REF(foo)
6309 OPENSSL_DECLARE_GLOBAL(double,bar);
6310 #define bar OPENSSL_GLOBAL_REF(bar)
6312 The #defines are very important, and therefore so is including the
6313 header file everywhere where the defined globals are used.
6315 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6316 of ASN.1 items, but that structure is a bit different.
6318 The largest change is in util/mkdef.pl which has been enhanced with
6319 better and easier to understand logic to choose which symbols should
6320 go into the Windows .def files as well as a number of fixes and code
6321 cleanup (among others, algorithm keywords are now sorted
6322 lexicographically to avoid constant rewrites).
6325 *) In BN_div() keep a copy of the sign of 'num' before writing the
6326 result to 'rm' because if rm==num the value will be overwritten
6327 and produce the wrong result if 'num' is negative: this caused
6328 problems with BN_mod() and BN_nnmod().
6331 *) Function OCSP_request_verify(). This checks the signature on an
6332 OCSP request and verifies the signer certificate. The signer
6333 certificate is just checked for a generic purpose and OCSP request
6337 *) Add OCSP_check_validity() function to check the validity of OCSP
6338 responses. OCSP responses are prepared in real time and may only
6339 be a few seconds old. Simply checking that the current time lies
6340 between thisUpdate and nextUpdate max reject otherwise valid responses
6341 caused by either OCSP responder or client clock inaccuracy. Instead
6342 we allow thisUpdate and nextUpdate to fall within a certain period of
6343 the current time. The age of the response can also optionally be
6344 checked. Two new options -validity_period and -status_age added to
6348 *) If signature or public key algorithm is unrecognized print out its
6349 OID rather that just UNKNOWN.
6352 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6353 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6354 ID to be generated from the issuer certificate alone which can then be
6355 passed to OCSP_id_issuer_cmp().
6358 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6359 ASN1 modules to export functions returning ASN1_ITEM pointers
6360 instead of the ASN1_ITEM structures themselves. This adds several
6361 new macros which allow the underlying ASN1 function/structure to
6362 be accessed transparently. As a result code should not use ASN1_ITEM
6363 references directly (such as &X509_it) but instead use the relevant
6364 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6365 use of the new ASN1 code on platforms where exporting structures
6366 is problematical (for example in shared libraries) but exporting
6367 functions returning pointers to structures is not.
6370 *) Add support for overriding the generation of SSL/TLS session IDs.
6371 These callbacks can be registered either in an SSL_CTX or per SSL.
6372 The purpose of this is to allow applications to control, if they wish,
6373 the arbitrary values chosen for use as session IDs, particularly as it
6374 can be useful for session caching in multiple-server environments. A
6375 command-line switch for testing this (and any client code that wishes
6376 to use such a feature) has been added to "s_server".
6377 [Geoff Thorpe, Lutz Jaenicke]
6379 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6380 of the form '#if defined(...) || defined(...) || ...' and
6381 '#if !defined(...) && !defined(...) && ...'. This also avoids
6382 the growing number of special cases it was previously handling.
6385 *) Make all configuration macros available for application by making
6386 sure they are available in opensslconf.h, by giving them names starting
6387 with "OPENSSL_" to avoid conflicts with other packages and by making
6388 sure e_os2.h will cover all platform-specific cases together with
6390 Additionally, it is now possible to define configuration/platform-
6391 specific names (called "system identities"). In the C code, these
6392 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6393 macro with the name beginning with "OPENSSL_SYS_", which is determined
6394 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6398 *) New option -set_serial to 'req' and 'x509' this allows the serial
6399 number to use to be specified on the command line. Previously self
6400 signed certificates were hard coded with serial number 0 and the
6401 CA options of 'x509' had to use a serial number in a file which was
6405 *) New options to 'ca' utility to support V2 CRL entry extensions.
6406 Currently CRL reason, invalidity date and hold instruction are
6407 supported. Add new CRL extensions to V3 code and some new objects.
6410 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6411 disable standard block padding (aka PKCS#5 padding) in the EVP
6412 API, which was previously mandatory. This means that the data is
6413 not padded in any way and so the total length much be a multiple
6414 of the block size, otherwise an error occurs.
6417 *) Initial (incomplete) OCSP SSL support.
6420 *) New function OCSP_parse_url(). This splits up a URL into its host,
6421 port and path components: primarily to parse OCSP URLs. New -url
6422 option to ocsp utility.
6425 *) New nonce behavior. The return value of OCSP_check_nonce() now
6426 reflects the various checks performed. Applications can decide
6427 whether to tolerate certain situations such as an absent nonce
6428 in a response when one was present in a request: the ocsp application
6429 just prints out a warning. New function OCSP_add1_basic_nonce()
6430 this is to allow responders to include a nonce in a response even if
6431 the request is nonce-less.
6434 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6435 skipped when using openssl x509 multiple times on a single input file,
6436 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6439 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6440 set string type: to handle setting ASN1_TIME structures. Fix ca
6441 utility to correctly initialize revocation date of CRLs.
6444 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6445 the clients preferred ciphersuites and rather use its own preferences.
6446 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6447 Internet Explorer by ensuring unchanged hash method during stepup.
6448 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6451 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6452 to aes and add a new 'exist' option to print out symbols that don't
6456 *) Additional options to ocsp utility to allow flags to be set and
6457 additional certificates supplied.
6460 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6461 OCSP client a number of certificate to only verify the response
6465 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6466 handle the new API. Currently only ECB, CBC modes supported. Add new
6469 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6470 Encryption Standard (AES) Ciphersuites for Transport Layer
6471 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6472 not enabled by default and were not part of the "ALL" ciphersuite
6473 alias because they were not yet official; they could be
6474 explicitly requested by specifying the "AESdraft" ciphersuite
6475 group alias. In the final release of OpenSSL 0.9.7, the group
6476 alias is called "AES" and is part of "ALL".)
6477 [Ben Laurie, Steve Henson, Bodo Moeller]
6479 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6480 request to response.
6483 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6484 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6485 extract information from a certificate request. OCSP_response_create()
6486 creates a response and optionally adds a basic response structure.
6487 OCSP_basic_add1_status() adds a complete single response to a basic
6488 response and returns the OCSP_SINGLERESP structure just added (to allow
6489 extensions to be included for example). OCSP_basic_add1_cert() adds a
6490 certificate to a basic response and OCSP_basic_sign() signs a basic
6491 response with various flags. New helper functions ASN1_TIME_check()
6492 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6493 (converts ASN1_TIME to GeneralizedTime).
6496 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6497 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6498 structure from a certificate. X509_pubkey_digest() digests the public_key
6499 contents: this is used in various key identifiers.
6502 *) Make sk_sort() tolerate a NULL argument.
6503 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6505 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6506 passed by the function are trusted implicitly. If any of them signed the
6507 response then it is assumed to be valid and is not verified.
6510 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6511 to data. This was previously part of the PKCS7 ASN1 code. This
6512 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6513 [Steve Henson, reported by Kenneth R. Robinette
6514 <support@securenetterm.com>]
6516 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6517 routines: without these tracing memory leaks is very painful.
6518 Fix leaks in PKCS12 and PKCS7 routines.
6521 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6522 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6523 effectively meant GeneralizedTime would never be used. Now it
6524 is initialised to -1 but X509_time_adj() now has to check the value
6525 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6526 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6527 [Steve Henson, reported by Kenneth R. Robinette
6528 <support@securenetterm.com>]
6530 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6531 result in a zero length in the ASN1_INTEGER structure which was
6532 not consistent with the structure when d2i_ASN1_INTEGER() was used
6533 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6534 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6535 where it did not print out a minus for negative ASN1_INTEGER.
6538 *) Add summary printout to ocsp utility. The various functions which
6539 convert status values to strings have been renamed to:
6540 OCSP_response_status_str(), OCSP_cert_status_str() and
6541 OCSP_crl_reason_str() and are no longer static. New options
6542 to verify nonce values and to disable verification. OCSP response
6543 printout format cleaned up.
6546 *) Add additional OCSP certificate checks. These are those specified
6547 in RFC2560. This consists of two separate checks: the CA of the
6548 certificate being checked must either be the OCSP signer certificate
6549 or the issuer of the OCSP signer certificate. In the latter case the
6550 OCSP signer certificate must contain the OCSP signing extended key
6551 usage. This check is performed by attempting to match the OCSP
6552 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6553 in the OCSP_CERTID structures of the response.
6556 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6557 and related routines. This uses the standard OpenSSL certificate
6558 verify routines to perform initial checks (just CA validity) and
6559 to obtain the certificate chain. Then additional checks will be
6560 performed on the chain. Currently the root CA is checked to see
6561 if it is explicitly trusted for OCSP signing. This is used to set
6562 a root CA as a global signing root: that is any certificate that
6563 chains to that CA is an acceptable OCSP signing certificate.
6566 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6567 extensions from a separate configuration file.
6568 As when reading extensions from the main configuration file,
6569 the '-extensions ...' option may be used for specifying the
6571 [Massimiliano Pala <madwolf@comune.modena.it>]
6573 *) New OCSP utility. Allows OCSP requests to be generated or
6574 read. The request can be sent to a responder and the output
6575 parsed, outputed or printed in text form. Not complete yet:
6576 still needs to check the OCSP response validity.
6579 *) New subcommands for 'openssl ca':
6580 'openssl ca -status <serial>' prints the status of the cert with
6581 the given serial number (according to the index file).
6582 'openssl ca -updatedb' updates the expiry status of certificates
6584 [Massimiliano Pala <madwolf@comune.modena.it>]
6586 *) New '-newreq-nodes' command option to CA.pl. This is like
6587 '-newreq', but calls 'openssl req' with the '-nodes' option
6588 so that the resulting key is not encrypted.
6589 [Damien Miller <djm@mindrot.org>]
6591 *) New configuration for the GNU Hurd.
6592 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6594 *) Initial code to implement OCSP basic response verify. This
6595 is currently incomplete. Currently just finds the signer's
6596 certificate and verifies the signature on the response.
6599 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6600 value of OPENSSLDIR. This is available via the new '-d' option
6601 to 'openssl version', and is also included in 'openssl version -a'.
6604 *) Allowing defining memory allocation callbacks that will be given
6605 file name and line number information in additional arguments
6606 (a const char* and an int). The basic functionality remains, as
6607 well as the original possibility to just replace malloc(),
6608 realloc() and free() by functions that do not know about these
6609 additional arguments. To register and find out the current
6610 settings for extended allocation functions, the following
6611 functions are provided:
6613 CRYPTO_set_mem_ex_functions
6614 CRYPTO_set_locked_mem_ex_functions
6615 CRYPTO_get_mem_ex_functions
6616 CRYPTO_get_locked_mem_ex_functions
6618 These work the same way as CRYPTO_set_mem_functions and friends.
6619 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6620 extended allocation function is enabled.
6621 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6622 a conventional allocation function is enabled.
6623 [Richard Levitte, Bodo Moeller]
6625 *) Finish off removing the remaining LHASH function pointer casts.
6626 There should no longer be any prototype-casting required when using
6627 the LHASH abstraction, and any casts that remain are "bugs". See
6628 the callback types and macros at the head of lhash.h for details
6629 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6632 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6633 If /dev/[u]random devices are not available or do not return enough
6634 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6636 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6637 /etc/entropy will be queried once each in this sequence, quering stops
6638 when enough entropy was collected without querying more sockets.
6641 *) Change the Unix RAND_poll() variant to be able to poll several
6642 random devices, as specified by DEVRANDOM, until a sufficient amount
6643 of data has been collected. We spend at most 10 ms on each file
6644 (select timeout) and read in non-blocking mode. DEVRANDOM now
6645 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6646 (previously it was just the string "/dev/urandom"), so on typical
6647 platforms the 10 ms delay will never occur.
6648 Also separate out the Unix variant to its own file, rand_unix.c.
6649 For VMS, there's a currently-empty rand_vms.c.
6652 *) Move OCSP client related routines to ocsp_cl.c. These
6653 provide utility functions which an application needing
6654 to issue a request to an OCSP responder and analyse the
6655 response will typically need: as opposed to those which an
6656 OCSP responder itself would need which will be added later.
6658 OCSP_request_sign() signs an OCSP request with an API similar
6659 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6660 response. OCSP_response_get1_basic() extracts basic response
6661 from response. OCSP_resp_find_status(): finds and extracts status
6662 information from an OCSP_CERTID structure (which will be created
6663 when the request structure is built). These are built from lower
6664 level functions which work on OCSP_SINGLERESP structures but
6665 wont normally be used unless the application wishes to examine
6666 extensions in the OCSP response for example.
6668 Replace nonce routines with a pair of functions.
6669 OCSP_request_add1_nonce() adds a nonce value and optionally
6670 generates a random value. OCSP_check_nonce() checks the
6671 validity of the nonce in an OCSP response.
6674 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6675 This doesn't copy the supplied OCSP_CERTID and avoids the
6676 need to free up the newly created id. Change return type
6677 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6678 This can then be used to add extensions to the request.
6679 Deleted OCSP_request_new(), since most of its functionality
6680 is now in OCSP_REQUEST_new() (and the case insensitive name
6681 clash) apart from the ability to set the request name which
6682 will be added elsewhere.
6685 *) Update OCSP API. Remove obsolete extensions argument from
6686 various functions. Extensions are now handled using the new
6687 OCSP extension code. New simple OCSP HTTP function which
6688 can be used to send requests and parse the response.
6691 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6692 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6693 uses the special reorder version of SET OF to sort the attributes
6694 and reorder them to match the encoded order. This resolves a long
6695 standing problem: a verify on a PKCS7 structure just after signing
6696 it used to fail because the attribute order did not match the
6697 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6698 it uses the received order. This is necessary to tolerate some broken
6699 software that does not order SET OF. This is handled by encoding
6700 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6701 to produce the required SET OF.
6704 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6705 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6706 files to get correct declarations of the ASN.1 item variables.
6709 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6710 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6711 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6712 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6713 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6714 ASN1_ITEM and no wrapper functions.
6717 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6718 replace the old function pointer based I/O routines. Change most of
6719 the *_d2i_bio() and *_d2i_fp() functions to use these.
6722 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6723 lines, recognice more "algorithms" that can be deselected, and make
6724 it complain about algorithm deselection that isn't recognised.
6727 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6728 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6729 to use new functions. Add NO_ASN1_OLD which can be set to remove
6730 some old style ASN1 functions: this can be used to determine if old
6731 code will still work when these eventually go away.
6734 *) New extension functions for OCSP structures, these follow the
6735 same conventions as certificates and CRLs.
6738 *) New function X509V3_add1_i2d(). This automatically encodes and
6739 adds an extension. Its behaviour can be customised with various
6740 flags to append, replace or delete. Various wrappers added for
6741 certifcates and CRLs.
6744 *) Fix to avoid calling the underlying ASN1 print routine when
6745 an extension cannot be parsed. Correct a typo in the
6746 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6749 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
6750 entries for variables.
6753 *) Add functionality to apps/openssl.c for detecting locking
6754 problems: As the program is single-threaded, all we have
6755 to do is register a locking callback using an array for
6756 storing which locks are currently held by the program.
6759 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6760 SSL_get_ex_data_X509_STORE_idx(), which is used in
6761 ssl_verify_cert_chain() and thus can be called at any time
6762 during TLS/SSL handshakes so that thread-safety is essential.
6763 Unfortunately, the ex_data design is not at all suited
6764 for multi-threaded use, so it probably should be abolished.
6767 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6768 [Broadcom, tweaked and integrated by Geoff Thorpe]
6770 *) Move common extension printing code to new function
6771 X509V3_print_extensions(). Reorganise OCSP print routines and
6772 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6775 *) New function X509_signature_print() to remove duplication in some
6779 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6780 set (this was treated exactly the same as SET OF previously). This
6781 is used to reorder the STACK representing the structure to match the
6782 encoding. This will be used to get round a problem where a PKCS7
6783 structure which was signed could not be verified because the STACK
6784 order did not reflect the encoded order.
6787 *) Reimplement the OCSP ASN1 module using the new code.
6790 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6791 for its ASN1 operations. The old style function pointers still exist
6792 for now but they will eventually go away.
6795 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6796 completely replaces the old ASN1 functionality with a table driven
6797 encoder and decoder which interprets an ASN1_ITEM structure describing
6798 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6799 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6800 has also been converted to the new form.
6803 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6804 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6805 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6806 for negative moduli.
6809 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6810 of not touching the result's sign bit.
6813 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6817 *) Changed the LHASH code to use prototypes for callbacks, and created
6818 macros to declare and implement thin (optionally static) functions
6819 that provide type-safety and avoid function pointer casting for the
6820 type-specific callbacks.
6823 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6825 [Veers Staats <staatsvr@asc.hpc.mil>,
6826 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6828 *) Reformat the FAQ so the different questions and answers can be divided
6829 in sections depending on the subject.
6832 *) Have the zlib compression code load ZLIB.DLL dynamically under
6836 *) New function BN_mod_sqrt for computing square roots modulo a prime
6837 (using the probabilistic Tonelli-Shanks algorithm unless
6838 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6839 be handled deterministically).
6840 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6842 *) Make BN_mod_inverse faster by explicitly handling small quotients
6843 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6844 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6847 *) New function BN_kronecker.
6850 *) Fix BN_gcd so that it works on negative inputs; the result is
6851 positive unless both parameters are zero.
6852 Previously something reasonably close to an infinite loop was
6853 possible because numbers could be growing instead of shrinking
6854 in the implementation of Euclid's algorithm.
6857 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6858 sign of the number in question.
6860 Fix BN_is_word(a,w) to work correctly for w == 0.
6862 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6863 because its test if the absolute value of 'a' equals 'w'.
6864 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6865 it exists mostly for use in the implementations of BN_is_zero(),
6866 BN_is_one(), and BN_is_word().
6869 *) New function BN_swap.
6872 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6873 the exponentiation functions are more likely to produce reasonable
6874 results on negative inputs.
6877 *) Change BN_mod_mul so that the result is always non-negative.
6878 Previously, it could be negative if one of the factors was negative;
6879 I don't think anyone really wanted that behaviour.
6882 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6883 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6884 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6885 and add new functions:
6894 BN_mod_lshift1_quick
6898 These functions always generate non-negative results.
6900 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6901 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6903 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6904 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6905 be reduced modulo m.
6906 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6909 The following entry accidentily appeared in the CHANGES file
6910 distributed with OpenSSL 0.9.7. The modifications described in
6911 it do *not* apply to OpenSSL 0.9.7.
6913 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6914 was actually never needed) and in BN_mul(). The removal in BN_mul()
6915 required a small change in bn_mul_part_recursive() and the addition
6916 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6917 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6918 bn_sub_words() and bn_add_words() except they take arrays with
6923 *) In 'openssl passwd', verify passwords read from the terminal
6924 unless the '-salt' option is used (which usually means that
6925 verification would just waste user's time since the resulting
6926 hash is going to be compared with some given password hash)
6927 or the new '-noverify' option is used.
6929 This is an incompatible change, but it does not affect
6930 non-interactive use of 'openssl passwd' (passwords on the command
6931 line, '-stdin' option, '-in ...' option) and thus should not
6935 *) Remove all references to RSAref, since there's no more need for it.
6938 *) Make DSO load along a path given through an environment variable
6939 (SHLIB_PATH) with shl_load().
6942 *) Constify the ENGINE code as a result of BIGNUM constification.
6943 Also constify the RSA code and most things related to it. In a
6944 few places, most notable in the depth of the ASN.1 code, ugly
6945 casts back to non-const were required (to be solved at a later
6949 *) Make it so the openssl application has all engines loaded by default.
6952 *) Constify the BIGNUM routines a little more.
6955 *) Add the following functions:
6957 ENGINE_load_cswift()
6959 ENGINE_load_atalla()
6961 ENGINE_load_builtin_engines()
6963 That way, an application can itself choose if external engines that
6964 are built-in in OpenSSL shall ever be used or not. The benefit is
6965 that applications won't have to be linked with libdl or other dso
6966 libraries unless it's really needed.
6968 Changed 'openssl engine' to load all engines on demand.
6969 Changed the engine header files to avoid the duplication of some
6970 declarations (they differed!).
6973 *) 'openssl engine' can now list capabilities.
6976 *) Better error reporting in 'openssl engine'.
6979 *) Never call load_dh_param(NULL) in s_server.
6982 *) Add engine application. It can currently list engines by name and
6983 identity, and test if they are actually available.
6986 *) Improve RPM specification file by forcing symbolic linking and making
6987 sure the installed documentation is also owned by root.root.
6988 [Damien Miller <djm@mindrot.org>]
6990 *) Give the OpenSSL applications more possibilities to make use of
6991 keys (public as well as private) handled by engines.
6994 *) Add OCSP code that comes from CertCo.
6997 *) Add VMS support for the Rijndael code.
7000 *) Added untested support for Nuron crypto accelerator.
7003 *) Add support for external cryptographic devices. This code was
7004 previously distributed separately as the "engine" branch.
7005 [Geoff Thorpe, Richard Levitte]
7007 *) Rework the filename-translation in the DSO code. It is now possible to
7008 have far greater control over how a "name" is turned into a filename
7009 depending on the operating environment and any oddities about the
7010 different shared library filenames on each system.
7013 *) Support threads on FreeBSD-elf in Configure.
7016 *) Fix for SHA1 assembly problem with MASM: it produces
7017 warnings about corrupt line number information when assembling
7018 with debugging information. This is caused by the overlapping
7020 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7023 NCONF_get_number() has no error checking at all. As a replacement,
7024 NCONF_get_number_e() is defined (_e for "error checking") and is
7025 promoted strongly. The old NCONF_get_number is kept around for
7026 binary backward compatibility.
7027 Make it possible for methods to load from something other than a BIO,
7028 by providing a function pointer that is given a name instead of a BIO.
7029 For example, this could be used to load configuration data from an
7033 *) Fix for non blocking accept BIOs. Added new I/O special reason
7034 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7035 with non blocking I/O was not possible because no retry code was
7036 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7040 *) Added the beginnings of Rijndael support.
7043 *) Fix for bug in DirectoryString mask setting. Add support for
7044 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7045 to allow certificate printing to more controllable, additional
7046 'certopt' option to 'x509' to allow new printing options to be
7050 *) Clean old EAY MD5 hack from e_os.h.
7053 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7055 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7056 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7057 [Joe Orton, Steve Henson]
7059 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7061 *) Fix additional bug revealed by the NISCC test suite:
7063 Stop bug triggering large recursion when presented with
7064 certain ASN.1 tags (CVE-2003-0851)
7067 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7069 *) Fix various bugs revealed by running the NISCC test suite:
7071 Stop out of bounds reads in the ASN1 code when presented with
7072 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7074 If verify callback ignores invalid public key errors don't try to check
7075 certificate signature with the NULL public key.
7079 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7080 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7084 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7085 extra data after the compression methods not only for TLS 1.0
7086 but also for SSL 3.0 (as required by the specification).
7087 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7089 *) Change X509_certificate_type() to mark the key as exported/exportable
7090 when it's 512 *bits* long, not 512 bytes.
7093 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7095 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7096 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7097 a protocol version number mismatch like a decryption error
7098 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7101 *) Turn on RSA blinding by default in the default implementation
7102 to avoid a timing attack. Applications that don't want it can call
7103 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7104 They would be ill-advised to do so in most cases.
7105 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7107 *) Change RSA blinding code so that it works when the PRNG is not
7108 seeded (in this case, the secret RSA exponent is abused as
7109 an unpredictable seed -- if it is not unpredictable, there
7110 is no point in blinding anyway). Make RSA blinding thread-safe
7111 by remembering the creator's thread ID in rsa->blinding and
7112 having all other threads use local one-time blinding factors
7113 (this requires more computation than sharing rsa->blinding, but
7114 avoids excessive locking; and if an RSA object is not shared
7115 between threads, blinding will still be very fast).
7118 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7120 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7121 via timing by performing a MAC computation even if incorrrect
7122 block cipher padding has been found. This is a countermeasure
7123 against active attacks where the attacker has to distinguish
7124 between bad padding and a MAC verification error. (CVE-2003-0078)
7126 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7127 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7128 Martin Vuagnoux (EPFL, Ilion)]
7130 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7132 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7133 memory from it's contents. This is done with a counter that will
7134 place alternating values in each byte. This can be used to solve
7135 two issues: 1) the removal of calls to memset() by highly optimizing
7136 compilers, and 2) cleansing with other values than 0, since those can
7137 be read through on certain media, for example a swap space on disk.
7140 *) Bugfix: client side session caching did not work with external caching,
7141 because the session->cipher setting was not restored when reloading
7142 from the external cache. This problem was masked, when
7143 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7144 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7147 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7148 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7149 [Zeev Lieber <zeev-l@yahoo.com>]
7151 *) Undo an undocumented change introduced in 0.9.6e which caused
7152 repeated calls to OpenSSL_add_all_ciphers() and
7153 OpenSSL_add_all_digests() to be ignored, even after calling
7157 *) Change the default configuration reader to deal with last line not
7158 being properly terminated.
7161 *) Change X509_NAME_cmp() so it applies the special rules on handling
7162 DN values that are of type PrintableString, as well as RDNs of type
7163 emailAddress where the value has the type ia5String.
7164 [stefank@valicert.com via Richard Levitte]
7166 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7167 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7168 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7169 the bitwise-OR of the two for use by the majority of applications
7170 wanting this behaviour, and update the docs. The documented
7171 behaviour and actual behaviour were inconsistent and had been
7172 changing anyway, so this is more a bug-fix than a behavioural
7174 [Geoff Thorpe, diagnosed by Nadav Har'El]
7176 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7177 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7180 *) Fix initialization code race conditions in
7181 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7182 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7183 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7184 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7185 ssl2_get_cipher_by_char(),
7186 ssl3_get_cipher_by_char().
7187 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7189 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7190 the cached sessions are flushed, as the remove_cb() might use ex_data
7191 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7192 (see [openssl.org #212]).
7193 [Geoff Thorpe, Lutz Jaenicke]
7195 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7196 length, instead of the encoding length to d2i_ASN1_OBJECT.
7199 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7201 *) [In 0.9.6g-engine release:]
7202 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7203 [Lynn Gazis <lgazis@rainbow.com>]
7205 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7207 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7208 and get fix the header length calculation.
7209 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7210 Alon Kantor <alonk@checkpoint.com> (and others),
7213 *) Use proper error handling instead of 'assertions' in buffer
7214 overflow checks added in 0.9.6e. This prevents DoS (the
7215 assertions could call abort()).
7216 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7218 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7220 *) Add various sanity checks to asn1_get_length() to reject
7221 the ASN1 length bytes if they exceed sizeof(long), will appear
7222 negative or the content length exceeds the length of the
7224 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7226 *) Fix cipher selection routines: ciphers without encryption had no flags
7227 for the cipher strength set and where therefore not handled correctly
7228 by the selection routines (PR #130).
7231 *) Fix EVP_dsa_sha macro.
7235 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7236 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7237 that was added in OpenSSL 0.9.6d.
7239 As the countermeasure turned out to be incompatible with some
7240 broken SSL implementations, the new option is part of SSL_OP_ALL.
7241 SSL_OP_ALL is usually employed when compatibility with weird SSL
7242 implementations is desired (e.g. '-bugs' option to 's_client' and
7243 's_server'), so the new option is automatically set in many
7247 *) Changes in security patch:
7249 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7250 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7251 Air Force Materiel Command, USAF, under agreement number
7254 *) Add various sanity checks to asn1_get_length() to reject
7255 the ASN1 length bytes if they exceed sizeof(long), will appear
7256 negative or the content length exceeds the length of the
7257 supplied buffer. (CVE-2002-0659)
7258 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7260 *) Assertions for various potential buffer overflows, not known to
7262 [Ben Laurie (CHATS)]
7264 *) Various temporary buffers to hold ASCII versions of integers were
7265 too small for 64 bit platforms. (CVE-2002-0655)
7266 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7268 *) Remote buffer overflow in SSL3 protocol - an attacker could
7269 supply an oversized session ID to a client. (CVE-2002-0656)
7270 [Ben Laurie (CHATS)]
7272 *) Remote buffer overflow in SSL2 protocol - an attacker could
7273 supply an oversized client master key. (CVE-2002-0656)
7274 [Ben Laurie (CHATS)]
7276 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7278 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7279 encoded as NULL) with id-dsa-with-sha1.
7280 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7282 *) Check various X509_...() return values in apps/req.c.
7283 [Nils Larsch <nla@trustcenter.de>]
7285 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7286 an end-of-file condition would erronously be flagged, when the CRLF
7287 was just at the end of a processed block. The bug was discovered when
7288 processing data through a buffering memory BIO handing the data to a
7289 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7290 <ptsekov@syntrex.com> and Nedelcho Stanev.
7293 *) Implement a countermeasure against a vulnerability recently found
7294 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7295 before application data chunks to avoid the use of known IVs
7296 with data potentially chosen by the attacker.
7299 *) Fix length checks in ssl3_get_client_hello().
7302 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7303 to prevent ssl3_read_internal() from incorrectly assuming that
7304 ssl3_read_bytes() found application data while handshake
7305 processing was enabled when in fact s->s3->in_read_app_data was
7306 merely automatically cleared during the initial handshake.
7307 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7309 *) Fix object definitions for Private and Enterprise: they were not
7310 recognized in their shortname (=lowercase) representation. Extend
7311 obj_dat.pl to issue an error when using undefined keywords instead
7312 of silently ignoring the problem (Svenning Sorensen
7313 <sss@sss.dnsalias.net>).
7316 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7317 generators, i.e. generators other than 2 and 5. (Previously, the
7318 code did not properly initialise the 'add' and 'rem' values to
7319 BN_generate_prime().)
7321 In the new general case, we do not insist that 'generator' is
7322 actually a primitive root: This requirement is rather pointless;
7323 a generator of the order-q subgroup is just as good, if not
7327 *) Map new X509 verification errors to alerts. Discovered and submitted by
7328 Tom Wu <tom@arcot.com>.
7331 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7332 returning non-zero before the data has been completely received
7333 when using non-blocking I/O.
7334 [Bodo Moeller; problem pointed out by John Hughes]
7336 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7337 [Ben Laurie, Lutz Jaenicke]
7339 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7340 Yoram Zahavi <YoramZ@gilian.com>).
7343 *) Add information about CygWin 1.3 and on, and preserve proper
7344 configuration for the versions before that.
7345 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7347 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7348 check whether we deal with a copy of a session and do not delete from
7349 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7350 <izhar@checkpoint.com>.
7353 *) Do not store session data into the internal session cache, if it
7354 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7355 flag is set). Proposed by Aslam <aslam@funk.com>.
7358 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7362 *) [In 0.9.6d-engine release:]
7363 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7364 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7366 *) Add the configuration target linux-s390x.
7367 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7369 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7370 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7371 variable as an indication that a ClientHello message has been
7372 received. As the flag value will be lost between multiple
7373 invocations of ssl3_accept when using non-blocking I/O, the
7374 function may not be aware that a handshake has actually taken
7375 place, thus preventing a new session from being added to the
7378 To avoid this problem, we now set s->new_session to 2 instead of
7379 using a local variable.
7380 [Lutz Jaenicke, Bodo Moeller]
7382 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7383 if the SSL_R_LENGTH_MISMATCH error is detected.
7384 [Geoff Thorpe, Bodo Moeller]
7386 *) New 'shared_ldflag' column in Configure platform table.
7389 *) Fix EVP_CIPHER_mode macro.
7390 ["Dan S. Camper" <dan@bti.net>]
7392 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7393 type, we must throw them away by setting rr->length to 0.
7394 [D P Chang <dpc@qualys.com>]
7396 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7398 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7399 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7400 worked incorrectly for those cases where range = 10..._2 and
7401 3*range is two bits longer than range.)
7404 *) Only add signing time to PKCS7 structures if it is not already
7408 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7409 OBJ_ld_ce should be OBJ_id_ce.
7410 Also some ip-pda OIDs in crypto/objects/objects.txt were
7411 incorrect (cf. RFC 3039).
7412 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7414 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7415 returns early because it has nothing to do.
7416 [Andy Schneider <andy.schneider@bjss.co.uk>]
7418 *) [In 0.9.6c-engine release:]
7419 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7420 [Andy Schneider <andy.schneider@bjss.co.uk>]
7422 *) [In 0.9.6c-engine release:]
7423 Add support for Cryptographic Appliance's keyserver technology.
7424 (Use engine 'keyclient')
7425 [Cryptographic Appliances and Geoff Thorpe]
7427 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7428 is called via tools/c89.sh because arguments have to be
7429 rearranged (all '-L' options must appear before the first object
7431 [Richard Shapiro <rshapiro@abinitio.com>]
7433 *) [In 0.9.6c-engine release:]
7434 Add support for Broadcom crypto accelerator cards, backported
7436 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7438 *) [In 0.9.6c-engine release:]
7439 Add support for SureWare crypto accelerator cards from
7440 Baltimore Technologies. (Use engine 'sureware')
7441 [Baltimore Technologies and Mark Cox]
7443 *) [In 0.9.6c-engine release:]
7444 Add support for crypto accelerator cards from Accelerated
7445 Encryption Processing, www.aep.ie. (Use engine 'aep')
7446 [AEP Inc. and Mark Cox]
7448 *) Add a configuration entry for gcc on UnixWare.
7449 [Gary Benson <gbenson@redhat.com>]
7451 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7452 messages are stored in a single piece (fixed-length part and
7453 variable-length part combined) and fix various bugs found on the way.
7456 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7457 instead. BIO_gethostbyname() does not know what timeouts are
7458 appropriate, so entries would stay in cache even when they have
7460 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7462 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7463 faced with a pathologically small ClientHello fragment that does
7464 not contain client_version: Instead of aborting with an error,
7465 simply choose the highest available protocol version (i.e.,
7466 TLS 1.0 unless it is disabled). In practice, ClientHello
7467 messages are never sent like this, but this change gives us
7468 strictly correct behaviour at least for TLS.
7471 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7472 never resets s->method to s->ctx->method when called from within
7473 one of the SSL handshake functions.
7474 [Bodo Moeller; problem pointed out by Niko Baric]
7476 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7477 (sent using the client's version number) if client_version is
7478 smaller than the protocol version in use. Also change
7479 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7480 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7481 the client will at least see that alert.
7484 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7488 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7489 client receives HelloRequest while in a handshake.
7490 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7492 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7493 should end in 'break', not 'goto end' which circuments various
7494 cleanups done in state SSL_ST_OK. But session related stuff
7495 must be disabled for SSL_ST_OK in the case that we just sent a
7498 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7499 before just sending a HelloRequest.
7500 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7502 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7503 reveal whether illegal block cipher padding was found or a MAC
7504 verification error occured. (Neither SSLerr() codes nor alerts
7505 are directly visible to potential attackers, but the information
7506 may leak via logfiles.)
7508 Similar changes are not required for the SSL 2.0 implementation
7509 because the number of padding bytes is sent in clear for SSL 2.0,
7510 and the extra bytes are just ignored. However ssl/s2_pkt.c
7511 failed to verify that the purported number of padding bytes is in
7515 *) Add OpenUNIX-8 support including shared libraries
7516 (Boyd Lynn Gerber <gerberb@zenez.com>).
7519 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7520 'wristwatch attack' using huge encoding parameters (cf.
7521 James H. Manger's CRYPTO 2001 paper). Note that the
7522 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7523 encoding parameters and hence was not vulnerable.
7526 *) BN_sqr() bug fix.
7527 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7529 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7530 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7531 followed by modular reduction.
7532 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7534 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7535 equivalent based on BN_pseudo_rand() instead of BN_rand().
7538 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7539 This function was broken, as the check for a new client hello message
7540 to handle SGC did not allow these large messages.
7541 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7544 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7547 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7548 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7551 *) Rework the configuration and shared library support for Tru64 Unix.
7552 The configuration part makes use of modern compiler features and
7553 still retains old compiler behavior for those that run older versions
7554 of the OS. The shared library support part includes a variant that
7555 uses the RPATH feature, and is available through the special
7556 configuration target "alpha-cc-rpath", which will never be selected
7558 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7560 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7561 with the same message size as in ssl3_get_certificate_request().
7562 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7563 messages might inadvertently be reject as too long.
7564 [Petr Lampa <lampa@fee.vutbr.cz>]
7566 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7569 *) Modified SSL library such that the verify_callback that has been set
7570 specificly for an SSL object with SSL_set_verify() is actually being
7571 used. Before the change, a verify_callback set with this function was
7572 ignored and the verify_callback() set in the SSL_CTX at the time of
7573 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7574 to allow the necessary settings.
7577 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7578 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7579 done automatically (in contradiction to the requirements of the C
7580 standard). This made problems when used from OpenSSH.
7583 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7584 dh->length and always used
7586 BN_rand_range(priv_key, dh->p).
7588 BN_rand_range() is not necessary for Diffie-Hellman, and this
7589 specific range makes Diffie-Hellman unnecessarily inefficient if
7590 dh->length (recommended exponent length) is much smaller than the
7591 length of dh->p. We could use BN_rand_range() if the order of
7592 the subgroup was stored in the DH structure, but we only have
7597 BN_rand(priv_key, l, ...)
7599 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7605 RSA_eay_public_encrypt
7606 RSA_eay_private_decrypt
7607 RSA_eay_private_encrypt (signing)
7608 RSA_eay_public_decrypt (signature verification)
7610 (default implementations for RSA_public_encrypt,
7611 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7612 always reject numbers >= n.
7615 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7616 to synchronize access to 'locking_thread'. This is necessary on
7617 systems where access to 'locking_thread' (an 'unsigned long'
7618 variable) is not atomic.
7621 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7622 *before* setting the 'crypto_lock_rand' flag. The previous code had
7623 a race condition if 0 is a valid thread ID.
7624 [Travis Vitek <vitek@roguewave.com>]
7626 *) Add support for shared libraries under Irix.
7627 [Albert Chin-A-Young <china@thewrittenword.com>]
7629 *) Add configuration option to build on Linux on both big-endian and
7631 [Ralf Baechle <ralf@uni-koblenz.de>]
7633 *) Add the possibility to create shared libraries on HP-UX.
7636 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7638 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7639 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7640 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7641 PRNG state recovery was possible based on the output of
7642 one PRNG request appropriately sized to gain knowledge on
7643 'md' followed by enough consecutive 1-byte PRNG requests
7644 to traverse all of 'state'.
7646 1. When updating 'md_local' (the current thread's copy of 'md')
7647 during PRNG output generation, hash all of the previous
7648 'md_local' value, not just the half used for PRNG output.
7650 2. Make the number of bytes from 'state' included into the hash
7651 independent from the number of PRNG bytes requested.
7653 The first measure alone would be sufficient to avoid
7654 Markku-Juhani's attack. (Actually it had never occurred
7655 to me that the half of 'md_local' used for chaining was the
7656 half from which PRNG output bytes were taken -- I had always
7657 assumed that the secret half would be used.) The second
7658 measure makes sure that additional data from 'state' is never
7659 mixed into 'md_local' in small portions; this heuristically
7660 further strengthens the PRNG.
7663 *) Fix crypto/bn/asm/mips3.s.
7666 *) When only the key is given to "enc", the IV is undefined. Print out
7667 an error message in this case.
7670 *) Handle special case when X509_NAME is empty in X509 printing routines.
7673 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7674 positive and less than q.
7677 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7678 used: it isn't thread safe and the add_lock_callback should handle
7680 [Paul Rose <Paul.Rose@bridge.com>]
7682 *) Verify that incoming data obeys the block size in
7683 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7687 [Ulf Möller, Bodo Möller]
7689 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7690 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7691 when fixing the server behaviour for backwards-compatible 'client
7692 hello' messages. (Note that the attack is impractical against
7693 SSL 3.0 and TLS 1.0 anyway because length and version checking
7694 means that the probability of guessing a valid ciphertext is
7695 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7698 Before 0.9.5, the countermeasure (hide the error by generating a
7699 random 'decryption result') did not work properly because
7700 ERR_clear_error() was missing, meaning that SSL_get_error() would
7701 detect the supposedly ignored error.
7703 Both problems are now fixed.
7706 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7707 (previously it was 1024).
7710 *) Fix for compatibility mode trust settings: ignore trust settings
7711 unless some valid trust or reject settings are present.
7714 *) Fix for blowfish EVP: its a variable length cipher.
7717 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7718 parameters in DSA public key structures and return an error in the
7719 DSA routines if parameters are absent.
7722 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7723 in the current directory if neither $RANDFILE nor $HOME was set.
7724 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7725 caused some confusion to Windows users who haven't defined $HOME.
7726 Thus RAND_file_name() is changed again: e_os.h can define a
7727 DEFAULT_HOME, which will be used if $HOME is not set.
7728 For Windows, we use "C:"; on other platforms, we still require
7729 environment variables.
7731 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7732 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7733 having multiple threads call RAND_poll() concurrently.
7736 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7737 combination of a flag and a thread ID variable.
7738 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7739 flag), *other* threads can enter ssleay_add_bytes without obeying
7740 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7741 that they do not hold after the first thread unsets add_do_not_lock).
7744 *) Change bctest again: '-x' expressions are not available in all
7748 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7750 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7751 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7753 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7754 the default extension for executables, if any. Also, make the perl
7755 scripts that use symlink() to test if it really exists and use "cp"
7756 if it doesn't. All this made OpenSSL compilable and installable in
7760 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7761 If SEQUENCE is length is indefinite just set c->slen to the total
7762 amount of data available.
7763 [Steve Henson, reported by shige@FreeBSD.org]
7764 [This change does not apply to 0.9.7.]
7766 *) Change bctest to avoid here-documents inside command substitution
7767 (workaround for FreeBSD /bin/sh bug).
7768 For compatibility with Ultrix, avoid shell functions (introduced
7769 in the bctest version that searches along $PATH).
7772 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7773 with des_encrypt() defined on some operating systems, like Solaris
7777 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7778 On the Importance of Eliminating Errors in Cryptographic
7779 Computations, J. Cryptology 14 (2001) 2, 101-119,
7780 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7783 *) MIPS assembler BIGNUM division bug fix.
7786 *) Disabled incorrect Alpha assembler code.
7789 *) Fix PKCS#7 decode routines so they correctly update the length
7790 after reading an EOC for the EXPLICIT tag.
7792 [This change does not apply to 0.9.7.]
7794 *) Fix bug in PKCS#12 key generation routines. This was triggered
7795 if a 3DES key was generated with a 0 initial byte. Include
7796 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7797 (but broken) behaviour.
7800 *) Enhance bctest to search for a working bc along $PATH and print
7802 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7804 *) Fix memory leaks in err.c: free err_data string if necessary;
7805 don't write to the wrong index in ERR_set_error_data.
7808 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7812 *) Replace rdtsc with _emit statements for VC++ version 5.
7813 [Jeremy Cooper <jeremy@baymoo.org>]
7815 *) Make it possible to reuse SSLv2 sessions.
7818 *) In copy_email() check for >= 0 as a return value for
7819 X509_NAME_get_index_by_NID() since 0 is a valid index.
7820 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7822 *) Avoid coredump with unsupported or invalid public keys by checking if
7823 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7824 PKCS7_verify() fails with non detached data.
7827 *) Don't use getenv in library functions when run as setuid/setgid.
7828 New function OPENSSL_issetugid().
7831 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7832 due to incorrect handling of multi-threading:
7834 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7836 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7838 3. Count how many times MemCheck_off() has been called so that
7839 nested use can be treated correctly. This also avoids
7840 inband-signalling in the previous code (which relied on the
7841 assumption that thread ID 0 is impossible).
7844 *) Add "-rand" option also to s_client and s_server.
7847 *) Fix CPU detection on Irix 6.x.
7848 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7849 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7851 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7854 [This change does not apply to 0.9.7.]
7856 *) Use the cached encoding of an X509_NAME structure rather than
7857 copying it. This is apparently the reason for the libsafe "errors"
7858 but the code is actually correct.
7861 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7862 Bleichenbacher's DSA attack.
7863 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7864 to be set and top=0 forces the highest bit to be set; top=-1 is new
7865 and leaves the highest bit random.
7866 [Ulf Moeller, Bodo Moeller]
7868 *) In the NCONF_...-based implementations for CONF_... queries
7869 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7870 a temporary CONF structure with the data component set to NULL
7871 (which gives segmentation faults in lh_retrieve).
7872 Instead, use NULL for the CONF pointer in CONF_get_string and
7873 CONF_get_number (which may use environment variables) and directly
7874 return NULL from CONF_get_section.
7877 *) Fix potential buffer overrun for EBCDIC.
7880 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7881 keyUsage if basicConstraints absent for a CA.
7884 *) Make SMIME_write_PKCS7() write mail header values with a format that
7885 is more generally accepted (no spaces before the semicolon), since
7886 some programs can't parse those values properly otherwise. Also make
7887 sure BIO's that break lines after each write do not create invalid
7891 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7892 macros previously used would not encode an empty SEQUENCE OF
7893 and break the signature.
7895 [This change does not apply to 0.9.7.]
7897 *) Zero the premaster secret after deriving the master secret in
7901 *) Add some EVP_add_digest_alias registrations (as found in
7902 OpenSSL_add_all_digests()) to SSL_library_init()
7903 aka OpenSSL_add_ssl_algorithms(). This provides improved
7904 compatibility with peers using X.509 certificates
7905 with unconventional AlgorithmIdentifier OIDs.
7908 *) Fix for Irix with NO_ASM.
7909 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7911 *) ./config script fixes.
7912 [Ulf Moeller, Richard Levitte]
7914 *) Fix 'openssl passwd -1'.
7917 *) Change PKCS12_key_gen_asc() so it can cope with non null
7918 terminated strings whose length is passed in the passlen
7919 parameter, for example from PEM callbacks. This was done
7920 by adding an extra length parameter to asc2uni().
7921 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7923 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7924 call failed, free the DSA structure.
7927 *) Fix to uni2asc() to cope with zero length Unicode strings.
7928 These are present in some PKCS#12 files.
7931 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7932 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7933 when writing a 32767 byte record.
7934 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7936 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7937 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7939 (RSA objects have a reference count access to which is protected
7940 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7941 so they are meant to be shared between threads.)
7942 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7943 "Reddie, Steven" <Steven.Reddie@ca.com>]
7945 *) Fix a deadlock in CRYPTO_mem_leaks().
7948 *) Use better test patterns in bntest.
7951 *) rand_win.c fix for Borland C.
7954 *) BN_rshift bugfix for n == 0.
7957 *) Add a 'bctest' script that checks for some known 'bc' bugs
7958 so that 'make test' does not abort just because 'bc' is broken.
7961 *) Store verify_result within SSL_SESSION also for client side to
7962 avoid potential security hole. (Re-used sessions on the client side
7963 always resulted in verify_result==X509_V_OK, not using the original
7964 result of the server certificate verification.)
7967 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7968 SSL3_RT_APPLICATION_DATA, return 0.
7969 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7973 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7974 releases, have been re-implemented by renaming the previous
7975 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7976 and ssl3_read_internal, respectively, and adding 'peek' parameters
7977 to them. The new ssl[23]_{read,peek} functions are calls to
7978 ssl[23]_read_internal with the 'peek' flag set appropriately.
7979 A 'peek' parameter has also been added to ssl3_read_bytes, which
7980 does the actual work for ssl3_read_internal.
7983 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7984 the method-specific "init()" handler. Also clean up ex_data after
7985 calling the method-specific "finish()" handler. Previously, this was
7986 happening the other way round.
7989 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7990 The previous value, 12, was not always sufficient for BN_mod_exp().
7993 *) Make sure that shared libraries get the internal name engine with
7994 the full version number and not just 0. This should mark the
7995 shared libraries as not backward compatible. Of course, this should
7996 be changed again when we can guarantee backward binary compatibility.
7999 *) Fix typo in get_cert_by_subject() in by_dir.c
8000 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8002 *) Rework the system to generate shared libraries:
8004 - Make note of the expected extension for the shared libraries and
8005 if there is a need for symbolic links from for example libcrypto.so.0
8006 to libcrypto.so.0.9.7. There is extended info in Configure for
8009 - Make as few rebuilds of the shared libraries as possible.
8011 - Still avoid linking the OpenSSL programs with the shared libraries.
8013 - When installing, install the shared libraries separately from the
8017 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8019 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8020 and not in SSL_clear because the latter is also used by the
8021 accept/connect functions; previously, the settings made by
8022 SSL_set_read_ahead would be lost during the handshake.
8023 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8025 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8026 Previously, it would create entries for disableed algorithms no
8030 *) Added several new manual pages for SSL_* function.
8033 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8035 *) In ssl23_get_client_hello, generate an error message when faced
8036 with an initial SSL 3.0/TLS record that is too small to contain the
8037 first two bytes of the ClientHello message, i.e. client_version.
8038 (Note that this is a pathologic case that probably has never happened
8039 in real life.) The previous approach was to use the version number
8040 from the record header as a substitute; but our protocol choice
8041 should not depend on that one because it is not authenticated
8042 by the Finished messages.
8045 *) More robust randomness gathering functions for Windows.
8046 [Jeffrey Altman <jaltman@columbia.edu>]
8048 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8049 not set then we don't setup the error code for issuer check errors
8050 to avoid possibly overwriting other errors which the callback does
8051 handle. If an application does set the flag then we assume it knows
8052 what it is doing and can handle the new informational codes
8056 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8057 a general "ANY" type, as such it should be able to decode anything
8058 including tagged types. However it didn't check the class so it would
8059 wrongly interpret tagged types in the same way as their universal
8060 counterpart and unknown types were just rejected. Changed so that the
8061 tagged and unknown types are handled in the same way as a SEQUENCE:
8062 that is the encoding is stored intact. There is also a new type
8063 "V_ASN1_OTHER" which is used when the class is not universal, in this
8064 case we have no idea what the actual type is so we just lump them all
8068 *) On VMS, stdout may very well lead to a file that is written to
8069 in a record-oriented fashion. That means that every write() will
8070 write a separate record, which will be read separately by the
8071 programs trying to read from it. This can be very confusing.
8073 The solution is to put a BIO filter in the way that will buffer
8074 text until a linefeed is reached, and then write everything a
8075 line at a time, so every record written will be an actual line,
8076 not chunks of lines and not (usually doesn't happen, but I've
8077 seen it once) several lines in one record. BIO_f_linebuffer() is
8080 Currently, it's a VMS-only method, because that's where it has
8081 been tested well enough.
8084 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8085 it can return incorrect results.
8086 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8087 but it was in 0.9.6-beta[12].)
8090 *) Disable the check for content being present when verifying detached
8091 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8092 include zero length content when signing messages.
8095 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8096 BIO_ctrl (for BIO pairs).
8099 *) Add DSO method for VMS.
8102 *) Bug fix: Montgomery multiplication could produce results with the
8106 *) Add RPM specification openssl.spec and modify it to build three
8107 packages. The default package contains applications, application
8108 documentation and run-time libraries. The devel package contains
8109 include files, static libraries and function documentation. The
8110 doc package contains the contents of the doc directory. The original
8111 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8114 *) Add a large number of documentation files for many SSL routines.
8115 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8117 *) Add a configuration entry for Sony News 4.
8118 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8120 *) Don't set the two most significant bits to one when generating a
8121 random number < q in the DSA library.
8124 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8125 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8126 the underlying transport is blocking) if a handshake took place.
8127 (The default behaviour is needed by applications such as s_client
8128 and s_server that use select() to determine when to use SSL_read;
8129 but for applications that know in advance when to expect data, it
8130 just makes things more complicated.)
8133 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8137 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8138 work better on such systems.
8139 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8141 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8142 Update PKCS12_parse() so it copies the friendlyName and the
8143 keyid to the certificates aux info.
8146 *) Fix bug in PKCS7_verify() which caused an infinite loop
8147 if there was more than one signature.
8148 [Sven Uszpelkat <su@celocom.de>]
8150 *) Major change in util/mkdef.pl to include extra information
8151 about each symbol, as well as presentig variables as well
8152 as functions. This change means that there's n more need
8153 to rebuild the .num files when some algorithms are excluded.
8156 *) Allow the verify time to be set by an application,
8157 rather than always using the current time.
8160 *) Phase 2 verify code reorganisation. The certificate
8161 verify code now looks up an issuer certificate by a
8162 number of criteria: subject name, authority key id
8163 and key usage. It also verifies self signed certificates
8164 by the same criteria. The main comparison function is
8165 X509_check_issued() which performs these checks.
8167 Lot of changes were necessary in order to support this
8168 without completely rewriting the lookup code.
8170 Authority and subject key identifier are now cached.
8172 The LHASH 'certs' is X509_STORE has now been replaced
8173 by a STACK_OF(X509_OBJECT). This is mainly because an
8174 LHASH can't store or retrieve multiple objects with
8175 the same hash value.
8177 As a result various functions (which were all internal
8178 use only) have changed to handle the new X509_STORE
8179 structure. This will break anything that messed round
8180 with X509_STORE internally.
8182 The functions X509_STORE_add_cert() now checks for an
8183 exact match, rather than just subject name.
8185 The X509_STORE API doesn't directly support the retrieval
8186 of multiple certificates matching a given criteria, however
8187 this can be worked round by performing a lookup first
8188 (which will fill the cache with candidate certificates)
8189 and then examining the cache for matches. This is probably
8190 the best we can do without throwing out X509_LOOKUP
8191 entirely (maybe later...).
8193 The X509_VERIFY_CTX structure has been enhanced considerably.
8195 All certificate lookup operations now go via a get_issuer()
8196 callback. Although this currently uses an X509_STORE it
8197 can be replaced by custom lookups. This is a simple way
8198 to bypass the X509_STORE hackery necessary to make this
8199 work and makes it possible to use more efficient techniques
8200 in future. A very simple version which uses a simple
8201 STACK for its trusted certificate store is also provided
8202 using X509_STORE_CTX_trusted_stack().
8204 The verify_cb() and verify() callbacks now have equivalents
8205 in the X509_STORE_CTX structure.
8207 X509_STORE_CTX also has a 'flags' field which can be used
8208 to customise the verify behaviour.
8211 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8212 excludes S/MIME capabilities.
8215 *) When a certificate request is read in keep a copy of the
8216 original encoding of the signed data and use it when outputing
8217 again. Signatures then use the original encoding rather than
8218 a decoded, encoded version which may cause problems if the
8219 request is improperly encoded.
8222 *) For consistency with other BIO_puts implementations, call
8223 buffer_write(b, ...) directly in buffer_puts instead of calling
8226 In BIO_puts, increment b->num_write as in BIO_write.
8227 [Peter.Sylvester@EdelWeb.fr]
8229 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8230 BN_zero, we may not return a BIGNUM with an array consisting of
8234 *) Avoid calling abort() from within the library when problems are
8235 detected, except if preprocessor symbols have been defined
8236 (such as REF_CHECK, BN_DEBUG etc.).
8239 *) New openssl application 'rsautl'. This utility can be
8240 used for low level RSA operations. DER public key
8241 BIO/fp routines also added.
8244 *) New Configure entry and patches for compiling on QNX 4.
8245 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8247 *) A demo state-machine implementation was sponsored by
8248 Nuron (http://www.nuron.com/) and is now available in
8249 demos/state_machine.
8252 *) New options added to the 'dgst' utility for signature
8253 generation and verification.
8256 *) Unrecognized PKCS#7 content types are now handled via a
8257 catch all ASN1_TYPE structure. This allows unsupported
8258 types to be stored as a "blob" and an application can
8259 encode and decode it manually.
8262 *) Fix various signed/unsigned issues to make a_strex.c
8264 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8266 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8267 length if passed a buffer. ASN1_INTEGER_to_BN failed
8268 if passed a NULL BN and its argument was negative.
8269 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8271 *) Modification to PKCS#7 encoding routines to output definite
8272 length encoding. Since currently the whole structures are in
8273 memory there's not real point in using indefinite length
8274 constructed encoding. However if OpenSSL is compiled with
8275 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8278 *) Added BIO_vprintf() and BIO_vsnprintf().
8281 *) Added more prefixes to parse for in the the strings written
8282 through a logging bio, to cover all the levels that are available
8283 through syslog. The prefixes are now:
8285 PANIC, EMERG, EMR => LOG_EMERG
8286 ALERT, ALR => LOG_ALERT
8287 CRIT, CRI => LOG_CRIT
8288 ERROR, ERR => LOG_ERR
8289 WARNING, WARN, WAR => LOG_WARNING
8290 NOTICE, NOTE, NOT => LOG_NOTICE
8291 INFO, INF => LOG_INFO
8292 DEBUG, DBG => LOG_DEBUG
8294 and as before, if none of those prefixes are present at the
8295 beginning of the string, LOG_ERR is chosen.
8297 On Win32, the LOG_* levels are mapped according to this:
8299 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8300 LOG_WARNING => EVENTLOG_WARNING_TYPE
8301 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8305 *) Made it possible to reconfigure with just the configuration
8306 argument "reconf" or "reconfigure". The command line arguments
8307 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8308 and are retrieved from there when reconfiguring.
8312 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8314 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8317 *) The obj_dat.pl script was messing up the sorting of object
8318 names. The reason was that it compared the quoted version
8319 of strings as a result "OCSP" > "OCSP Signing" because
8320 " > SPACE. Changed script to store unquoted versions of
8321 names and add quotes on output. It was also omitting some
8322 names from the lookup table if they were given a default
8323 value (that is if SN is missing it is given the same
8324 value as LN and vice versa), these are now added on the
8325 grounds that if an object has a name we should be able to
8326 look it up. Finally added warning output when duplicate
8327 short or long names are found.
8330 *) Changes needed for Tandem NSK.
8331 [Scott Uroff <scott@xypro.com>]
8333 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8334 RSA_padding_check_SSLv23(), special padding was never detected
8335 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8336 version rollback attacks was not effective.
8338 In s23_clnt.c, don't use special rollback-attack detection padding
8339 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8340 client; similarly, in s23_srvr.c, don't do the rollback check if
8341 SSL 2.0 is the only protocol enabled in the server.
8344 *) Make it possible to get hexdumps of unprintable data with 'openssl
8345 asn1parse'. By implication, the functions ASN1_parse_dump() and
8346 BIO_dump_indent() are added.
8349 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8350 these print out strings and name structures based on various
8351 flags including RFC2253 support and proper handling of
8352 multibyte characters. Added options to the 'x509' utility
8353 to allow the various flags to be set.
8356 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8357 Also change the functions X509_cmp_current_time() and
8358 X509_gmtime_adj() work with an ASN1_TIME structure,
8359 this will enable certificates using GeneralizedTime in validity
8360 dates to be checked.
8363 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8364 negative public key encodings) on by default,
8365 NO_NEG_PUBKEY_BUG can be set to disable it.
8368 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8369 content octets. An i2c_ASN1_OBJECT is unnecessary because
8370 the encoding can be trivially obtained from the structure.
8373 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8374 not read locks (CRYPTO_r_[un]lock).
8377 *) A first attempt at creating official support for shared
8378 libraries through configuration. I've kept it so the
8379 default is static libraries only, and the OpenSSL programs
8380 are always statically linked for now, but there are
8381 preparations for dynamic linking in place.
8382 This has been tested on Linux and Tru64.
8385 *) Randomness polling function for Win9x, as described in:
8386 Peter Gutmann, Software Generation of Practically Strong
8390 *) Fix so PRNG is seeded in req if using an already existing
8394 *) New options to smime application. -inform and -outform
8395 allow alternative formats for the S/MIME message including
8396 PEM and DER. The -content option allows the content to be
8397 specified separately. This should allow things like Netscape
8398 form signing output easier to verify.
8401 *) Fix the ASN1 encoding of tags using the 'long form'.
8404 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8405 STRING types. These convert content octets to and from the
8406 underlying type. The actual tag and length octets are
8407 already assumed to have been read in and checked. These
8408 are needed because all other string types have virtually
8409 identical handling apart from the tag. By having versions
8410 of the ASN1 functions that just operate on content octets
8411 IMPLICIT tagging can be handled properly. It also allows
8412 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8413 and ASN1_INTEGER are identical apart from the tag.
8416 *) Change the handling of OID objects as follows:
8418 - New object identifiers are inserted in objects.txt, following
8419 the syntax given in objects.README.
8420 - objects.pl is used to process obj_mac.num and create a new
8422 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8425 This is currently kind of a hack, and the perl code in objects.pl
8426 isn't very elegant, but it works as I intended. The simplest way
8427 to check that it worked correctly is to look in obj_dat.h and
8428 check the array nid_objs and make sure the objects haven't moved
8429 around (this is important!). Additions are OK, as well as
8430 consistent name changes.
8433 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8436 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8437 The given file adds to whatever has already been seeded into the
8438 random pool through the RANDFILE configuration file option or
8439 environment variable, or the default random state file.
8442 *) mkstack.pl now sorts each macro group into lexical order.
8443 Previously the output order depended on the order the files
8444 appeared in the directory, resulting in needless rewriting
8448 *) Patches to make OpenSSL compile under Win32 again. Mostly
8449 work arounds for the VC++ problem that it treats func() as
8450 func(void). Also stripped out the parts of mkdef.pl that
8451 added extra typesafe functions: these no longer exist.
8454 *) Reorganisation of the stack code. The macros are now all
8455 collected in safestack.h . Each macro is defined in terms of
8456 a "stack macro" of the form SKM_<name>(type, a, b). The
8457 DEBUG_SAFESTACK is now handled in terms of function casts,
8458 this has the advantage of retaining type safety without the
8459 use of additional functions. If DEBUG_SAFESTACK is not defined
8460 then the non typesafe macros are used instead. Also modified the
8461 mkstack.pl script to handle the new form. Needs testing to see
8462 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8463 the default if no major problems. Similar behaviour for ASN1_SET_OF
8464 and PKCS12_STACK_OF.
8467 *) When some versions of IIS use the 'NET' form of private key the
8468 key derivation algorithm is different. Normally MD5(password) is
8469 used as a 128 bit RC4 key. In the modified case
8470 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
8471 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8472 as the old Netscape_RSA functions except they have an additional
8473 'sgckey' parameter which uses the modified algorithm. Also added
8474 an -sgckey command line option to the rsa utility. Thanks to
8475 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8476 algorithm to openssl-dev.
8479 *) The evp_local.h macros were using 'c.##kname' which resulted in
8480 invalid expansion on some systems (SCO 5.0.5 for example).
8481 Corrected to 'c.kname'.
8482 [Phillip Porch <root@theporch.com>]
8484 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8485 a STACK of email addresses from a certificate or request, these look
8486 in the subject name and the subject alternative name extensions and
8487 omit any duplicate addresses.
8490 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8491 This makes DSA verification about 2 % faster.
8494 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8495 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8496 plus overhead for 1024 bit moduli).
8497 This makes exponentiations about 0.5 % faster for 1024 bit
8498 exponents (as measured by "openssl speed rsa2048").
8501 *) Rename memory handling macros to avoid conflicts with other
8503 Malloc => OPENSSL_malloc
8504 Malloc_locked => OPENSSL_malloc_locked
8505 Realloc => OPENSSL_realloc
8506 Free => OPENSSL_free
8509 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8510 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8513 *) CygWin32 support.
8514 [John Jarvie <jjarvie@newsguy.com>]
8516 *) The type-safe stack code has been rejigged. It is now only compiled
8517 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8518 by default all type-specific stack functions are "#define"d back to
8519 standard stack functions. This results in more streamlined output
8520 but retains the type-safety checking possibilities of the original
8524 *) The STACK code has been cleaned up, and certain type declarations
8525 that didn't make a lot of sense have been brought in line. This has
8526 also involved a cleanup of sorts in safestack.h to more correctly
8527 map type-safe stack functions onto their plain stack counterparts.
8528 This work has also resulted in a variety of "const"ifications of
8529 lots of the code, especially "_cmp" operations which should normally
8530 be prototyped with "const" parameters anyway.
8533 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8534 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8535 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8536 where all of 'md' is used each time the PRNG is used, but 'state'
8537 is used only indexed by a cyclic counter. As entropy may not be
8538 well distributed from the beginning, 'md' is important as a
8539 chaining variable. However, the output function chains only half
8540 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8541 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8542 in all of 'state' being rewritten, with the new values depending
8543 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8546 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8547 the handshake is continued after ssl_verify_cert_chain();
8548 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8549 can lead to 'unexplainable' connection aborts later.
8550 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8552 *) Major EVP API cipher revision.
8553 Add hooks for extra EVP features. This allows various cipher
8554 parameters to be set in the EVP interface. Support added for variable
8555 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8556 setting of RC2 and RC5 parameters.
8558 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8561 Remove lots of duplicated code from the EVP library. For example *every*
8562 cipher init() function handles the 'iv' in the same way according to the
8563 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8564 for CFB and OFB modes they zero ctx->num.
8566 New functionality allows removal of S/MIME code RC2 hack.
8568 Most of the routines have the same form and so can be declared in terms
8571 By shifting this to the top level EVP_CipherInit() it can be removed from
8572 all individual ciphers. If the cipher wants to handle IVs or keys
8573 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8576 Change lots of functions like EVP_EncryptUpdate() to now return a
8577 value: although software versions of the algorithms cannot fail
8578 any installed hardware versions can.
8581 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8582 this option is set, tolerate broken clients that send the negotiated
8583 protocol version number instead of the requested protocol version
8587 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8588 i.e. non-zero for export ciphersuites, zero otherwise.
8589 Previous versions had this flag inverted, inconsistent with
8590 rsa_tmp_cb (..._TMP_RSA_CB).
8591 [Bodo Moeller; problem reported by Amit Chopra]
8593 *) Add missing DSA library text string. Work around for some IIS
8594 key files with invalid SEQUENCE encoding.
8597 *) Add a document (doc/standards.txt) that list all kinds of standards
8598 and so on that are implemented in OpenSSL.
8601 *) Enhance c_rehash script. Old version would mishandle certificates
8602 with the same subject name hash and wouldn't handle CRLs at all.
8603 Added -fingerprint option to crl utility, to support new c_rehash
8607 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8610 *) Fix for SSL server purpose checking. Server checking was
8611 rejecting certificates which had extended key usage present
8612 but no ssl client purpose.
8613 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8615 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8616 is a little unclear about how a blank password is handled.
8617 Since the password in encoded as a BMPString with terminating
8618 double NULL a zero length password would end up as just the
8619 double NULL. However no password at all is different and is
8620 handled differently in the PKCS#12 key generation code. NS
8621 treats a blank password as zero length. MSIE treats it as no
8622 password on export: but it will try both on import. We now do
8623 the same: PKCS12_parse() tries zero length and no password if
8624 the password is set to "" or NULL (NULL is now a valid password:
8625 it wasn't before) as does the pkcs12 application.
8628 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8629 perror when PEM_read_bio_X509_REQ fails, the error message must
8630 be obtained from the error queue.
8633 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8634 it in ERR_remove_state if appropriate, and change ERR_get_state
8635 accordingly to avoid race conditions (this is necessary because
8636 thread_hash is no longer constant once set).
8639 *) Bugfix for linux-elf makefile.one.
8642 *) RSA_get_default_method() will now cause a default
8643 RSA_METHOD to be chosen if one doesn't exist already.
8644 Previously this was only set during a call to RSA_new()
8645 or RSA_new_method(NULL) meaning it was possible for
8646 RSA_get_default_method() to return NULL.
8649 *) Added native name translation to the existing DSO code
8650 that will convert (if the flag to do so is set) filenames
8651 that are sufficiently small and have no path information
8652 into a canonical native form. Eg. "blah" converted to
8653 "libblah.so" or "blah.dll" etc.
8656 *) New function ERR_error_string_n(e, buf, len) which is like
8657 ERR_error_string(e, buf), but writes at most 'len' bytes
8658 including the 0 terminator. For ERR_error_string_n, 'buf'
8660 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8662 *) CONF library reworked to become more general. A new CONF
8663 configuration file reader "class" is implemented as well as a
8664 new functions (NCONF_*, for "New CONF") to handle it. The now
8665 old CONF_* functions are still there, but are reimplemented to
8666 work in terms of the new functions. Also, a set of functions
8667 to handle the internal storage of the configuration data is
8668 provided to make it easier to write new configuration file
8669 reader "classes" (I can definitely see something reading a
8670 configuration file in XML format, for example), called _CONF_*,
8671 or "the configuration storage API"...
8673 The new configuration file reading functions are:
8675 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8676 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8678 NCONF_default, NCONF_WIN32
8680 NCONF_dump_fp, NCONF_dump_bio
8682 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8683 NCONF_new creates a new CONF object. This works in the same way
8684 as other interfaces in OpenSSL, like the BIO interface.
8685 NCONF_dump_* dump the internal storage of the configuration file,
8686 which is useful for debugging. All other functions take the same
8687 arguments as the old CONF_* functions wth the exception of the
8688 first that must be a `CONF *' instead of a `LHASH *'.
8690 To make it easer to use the new classes with the old CONF_* functions,
8691 the function CONF_set_default_method is provided.
8694 *) Add '-tls1' option to 'openssl ciphers', which was already
8695 mentioned in the documentation but had not been implemented.
8696 (This option is not yet really useful because even the additional
8697 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8700 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8701 OpenSSL-based applications) load shared libraries and bind to
8702 them in a portable way.
8703 [Geoff Thorpe, with contributions from Richard Levitte]
8705 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8707 *) Make sure _lrotl and _lrotr are only used with MSVC.
8709 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8710 (the default implementation of RAND_status).
8712 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8713 to '-clrext' (= clear extensions), as intended and documented.
8714 [Bodo Moeller; inconsistency pointed out by Michael Attili
8717 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8718 was larger than the MD block size.
8719 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8721 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8722 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8723 using the passed key: if the passed key was a private key the result
8724 of X509_print(), for example, would be to print out all the private key
8728 *) des_quad_cksum() byte order bug fix.
8729 [Ulf Möller, using the problem description in krb4-0.9.7, where
8730 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8732 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8734 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8736 *) For easily testing in shell scripts whether some command
8737 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8738 returns with exit code 0 iff no command of the given name is available.
8739 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8740 the output goes to stdout and nothing is printed to stderr.
8741 Additional arguments are always ignored.
8743 Since for each cipher there is a command of the same name,
8744 the 'no-cipher' compilation switches can be tested this way.
8746 ('openssl no-XXX' is not able to detect pseudo-commands such
8747 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8750 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8753 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8754 is set; it will be thrown away anyway because each handshake creates
8756 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8757 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8758 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
8759 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8762 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8763 'Q' and 'R' lose their special meanings (quit/renegotiate).
8764 This is part of what -quiet does; unlike -quiet, -ign_eof
8765 does not suppress any output.
8768 *) Add compatibility options to the purpose and trust code. The
8769 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8770 accepts a certificate or CA, this was the previous behaviour,
8771 with all the associated security issues.
8773 X509_TRUST_COMPAT is the old trust behaviour: only and
8774 automatically trust self signed roots in certificate store. A
8775 new trust setting X509_TRUST_DEFAULT is used to specify that
8776 a purpose has no associated trust setting and it should instead
8777 use the value in the default purpose.
8780 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8781 and fix a memory leak.
8784 *) In util/mkerr.pl (which implements 'make errors'), preserve
8785 reason strings from the previous version of the .c file, as
8786 the default to have only downcase letters (and digits) in
8787 automatically generated reasons codes is not always appropriate.
8790 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8791 using strerror. Previously, ERR_reason_error_string() returned
8792 library names as reason strings for SYSerr; but SYSerr is a special
8793 case where small numbers are errno values, not library numbers.
8796 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8797 converts DSA parameters into DH parameters. (When creating parameters,
8798 DSA_generate_parameters is used.)
8801 *) Include 'length' (recommended exponent length) in C code generated
8802 by 'openssl dhparam -C'.
8805 *) The second argument to set_label in perlasm was already being used
8806 so couldn't be used as a "file scope" flag. Moved to third argument
8810 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8811 instead of RAND_bytes for encryption IVs and salts.
8814 *) Include RAND_status() into RAND_METHOD instead of implementing
8815 it only for md_rand.c Otherwise replacing the PRNG by calling
8816 RAND_set_rand_method would be impossible.
8819 *) Don't let DSA_generate_key() enter an infinite loop if the random
8820 number generation fails.
8823 *) New 'rand' application for creating pseudo-random output.
8826 *) Added configuration support for Linux/IA64
8827 [Rolf Haberrecker <rolf@suse.de>]
8829 *) Assembler module support for Mingw32.
8832 *) Shared library support for HPUX (in shlib/).
8833 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8835 *) Shared library support for Solaris gcc.
8836 [Lutz Behnke <behnke@trustcenter.de>]
8838 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8840 *) PKCS7_encrypt() was adding text MIME headers twice because they
8841 were added manually and by SMIME_crlf_copy().
8844 *) In bntest.c don't call BN_rand with zero bits argument.
8845 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8847 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8848 case was implemented. This caused BN_div_recp() to fail occasionally.
8851 *) Add an optional second argument to the set_label() in the perl
8852 assembly language builder. If this argument exists and is set
8853 to 1 it signals that the assembler should use a symbol whose
8854 scope is the entire file, not just the current function. This
8855 is needed with MASM which uses the format label:: for this scope.
8856 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8858 *) Change the ASN1 types so they are typedefs by default. Before
8859 almost all types were #define'd to ASN1_STRING which was causing
8860 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8864 *) Change names of new functions to the new get1/get0 naming
8865 convention: After 'get1', the caller owns a reference count
8866 and has to call ..._free; 'get0' returns a pointer to some
8867 data structure without incrementing reference counters.
8868 (Some of the existing 'get' functions increment a reference
8869 counter, some don't.)
8870 Similarly, 'set1' and 'add1' functions increase reference
8871 counters or duplicate objects.
8874 *) Allow for the possibility of temp RSA key generation failure:
8875 the code used to assume it always worked and crashed on failure.
8878 *) Fix potential buffer overrun problem in BIO_printf().
8879 [Ulf Möller, using public domain code by Patrick Powell; problem
8880 pointed out by David Sacerdote <das33@cornell.edu>]
8882 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8883 RAND_egd() and RAND_status(). In the command line application,
8884 the EGD socket can be specified like a seed file using RANDFILE
8888 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8889 Some CAs (e.g. Verisign) distribute certificates in this form.
8892 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8893 list to exclude them. This means that no special compilation option
8894 is needed to use anonymous DH: it just needs to be included in the
8898 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8899 EVP_MD_type. The old functionality is available in a new macro called
8900 EVP_MD_md(). Change code that uses it and update docs.
8903 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8904 where the 'void *' argument is replaced by a function pointer argument.
8905 Previously 'void *' was abused to point to functions, which works on
8906 many platforms, but is not correct. As these functions are usually
8907 called by macros defined in OpenSSL header files, most source code
8908 should work without changes.
8911 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8912 sections with information on -D... compiler switches used for
8913 compiling the library so that applications can see them. To enable
8914 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8915 must be defined. E.g.,
8916 #define OPENSSL_ALGORITHM_DEFINES
8917 #include <openssl/opensslconf.h>
8918 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8919 [Richard Levitte, Ulf and Bodo Möller]
8921 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8925 *) Change the 'other' type in certificate aux info to a STACK_OF
8926 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8927 the required ASN1 format: arbitrary types determined by an OID.
8930 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8931 argument to 'req'. This is not because the function is newer or
8932 better than others it just uses the work 'NEW' in the certificate
8933 request header lines. Some software needs this.
8936 *) Reorganise password command line arguments: now passwords can be
8937 obtained from various sources. Delete the PEM_cb function and make
8938 it the default behaviour: i.e. if the callback is NULL and the
8939 usrdata argument is not NULL interpret it as a null terminated pass
8940 phrase. If usrdata and the callback are NULL then the pass phrase
8941 is prompted for as usual.
8944 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8945 the support is automatically enabled. The resulting binaries will
8946 autodetect the card and use it if present.
8947 [Ben Laurie and Compaq Inc.]
8949 *) Work around for Netscape hang bug. This sends certificate request
8950 and server done in one record. Since this is perfectly legal in the
8951 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8952 the bugs/SSLv3 entry for more info.
8955 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8958 *) Add -rand argument to smime and pkcs12 applications and read/write
8962 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8965 *) Add command line password options to the remaining applications.
8968 *) Bug fix for BN_div_recp() for numerators with an even number of
8972 *) More tests in bntest.c, and changed test_bn output.
8975 *) ./config recognizes MacOS X now.
8978 *) Bug fix for BN_div() when the first words of num and divsor are
8979 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8982 *) Add support for various broken PKCS#8 formats, and command line
8983 options to produce them.
8986 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8987 get temporary BIGNUMs from a BN_CTX.
8990 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8994 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8995 include a #define from the old name to the new. The original intent
8996 was that statically linked binaries could for example just call
8997 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8998 link with digests. This never worked becayse SSLeay_add_all_digests()
8999 and SSLeay_add_all_ciphers() were in the same source file so calling
9000 one would link with the other. They are now in separate source files.
9003 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9006 *) Use a less unusual form of the Miller-Rabin primality test (it used
9007 a binary algorithm for exponentiation integrated into the Miller-Rabin
9008 loop, our standard modexp algorithms are faster).
9011 *) Support for the EBCDIC character set completed.
9012 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9014 *) Source code cleanups: use const where appropriate, eliminate casts,
9015 use void * instead of char * in lhash.
9018 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9019 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9020 this the server could overwrite ephemeral keys that the client
9024 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9025 using 50 iterations of the Rabin-Miller test.
9027 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9028 iterations of the Rabin-Miller test as required by the appendix
9029 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9030 As BN_is_prime_fasttest includes trial division, DSA parameter
9031 generation becomes much faster.
9033 This implies a change for the callback functions in DSA_is_prime
9034 and DSA_generate_parameters: The callback function is called once
9035 for each positive witness in the Rabin-Miller test, not just
9036 occasionally in the inner loop; and the parameters to the
9037 callback function now provide an iteration count for the outer
9038 loop rather than for the current invocation of the inner loop.
9039 DSA_generate_parameters additionally can call the callback
9040 function with an 'iteration count' of -1, meaning that a
9041 candidate has passed the trial division test (when q is generated
9042 from an application-provided seed, trial division is skipped).
9045 *) New function BN_is_prime_fasttest that optionally does trial
9046 division before starting the Rabin-Miller test and has
9047 an additional BN_CTX * argument (whereas BN_is_prime always
9048 has to allocate at least one BN_CTX).
9049 'callback(1, -1, cb_arg)' is called when a number has passed the
9050 trial division stage.
9053 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9057 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9060 *) New function BN_pseudo_rand().
9063 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9064 bignum version of BN_from_montgomery() with the working code from
9065 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9069 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9070 made it impossible to use the same SSL_SESSION data structure in
9071 SSL2 clients in multiple threads.
9074 *) The return value of RAND_load_file() no longer counts bytes obtained
9075 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9076 to seed the PRNG (previously an explicit byte count was required).
9077 [Ulf Möller, Bodo Möller]
9079 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9080 used (char *) instead of (void *) and had casts all over the place.
9083 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9086 *) Retain source code compatibility for BN_prime_checks macro:
9087 BN_is_prime(..., BN_prime_checks, ...) now uses
9088 BN_prime_checks_for_size to determine the appropriate number of
9089 Rabin-Miller iterations.
9092 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9093 DH_CHECK_P_NOT_SAFE_PRIME.
9094 (Check if this is true? OpenPGP calls them "strong".)
9097 *) Merge the functionality of "dh" and "gendh" programs into a new program
9098 "dhparam". The old programs are retained for now but will handle DH keys
9099 (instead of parameters) in future.
9102 *) Make the ciphers, s_server and s_client programs check the return values
9103 when a new cipher list is set.
9106 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9107 ciphers. Before when the 56bit ciphers were enabled the sorting was
9110 The syntax for the cipher sorting has been extended to support sorting by
9111 cipher-strength (using the strength_bits hard coded in the tables).
9112 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9114 Fix a bug in the cipher-command parser: when supplying a cipher command
9115 string with an "undefined" symbol (neither command nor alphanumeric
9116 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9117 an error is flagged.
9119 Due to the strength-sorting extension, the code of the
9120 ssl_create_cipher_list() function was completely rearranged. I hope that
9121 the readability was also increased :-)
9122 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9124 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9125 for the first serial number and places 2 in the serial number file. This
9126 avoids problems when the root CA is created with serial number zero and
9127 the first user certificate has the same issuer name and serial number
9131 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9132 the new code. Add documentation for this stuff.
9135 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9136 X509_*() to X509at_*() on the grounds that they don't handle X509
9137 structures and behave in an analagous way to the X509v3 functions:
9138 they shouldn't be called directly but wrapper functions should be used
9141 So we also now have some wrapper functions that call the X509at functions
9142 when passed certificate requests. (TO DO: similar things can be done with
9143 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9144 things. Some of these need some d2i or i2d and print functionality
9145 because they handle more complex structures.)
9148 *) Add missing #ifndefs that caused missing symbols when building libssl
9149 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9150 NO_RSA in ssl/s2*.c.
9151 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9153 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9154 has a return value which indicates the quality of the random data
9155 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9156 error queue. New function RAND_pseudo_bytes() generates output that is
9157 guaranteed to be unique but not unpredictable. RAND_add is like
9158 RAND_seed, but takes an extra argument for an entropy estimate
9159 (RAND_seed always assumes full entropy).
9162 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9163 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9164 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9165 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9166 false-positive rate of at most 2^-80 for random input.
9169 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9172 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9173 in the 0.9.5 release), this returns the chain
9174 from an X509_CTX structure with a dup of the stack and all
9175 the X509 reference counts upped: so the stack will exist
9176 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9179 Also make SSL_SESSION_print() print out the verify return
9183 *) Add manpage for the pkcs12 command. Also change the default
9184 behaviour so MAC iteration counts are used unless the new
9185 -nomaciter option is used. This improves file security and
9186 only older versions of MSIE (4.0 for example) need it.
9189 *) Honor the no-xxx Configure options when creating .DEF files.
9192 *) Add PKCS#10 attributes to field table: challengePassword,
9193 unstructuredName and unstructuredAddress. These are taken from
9194 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9195 international characters are used.
9197 More changes to X509_ATTRIBUTE code: allow the setting of types
9198 based on strings. Remove the 'loc' parameter when adding
9199 attributes because these will be a SET OF encoding which is sorted
9203 *) Initial changes to the 'req' utility to allow request generation
9204 automation. This will allow an application to just generate a template
9205 file containing all the field values and have req construct the
9208 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9209 used all over the place including certificate requests and PKCS#7
9210 structures. They are currently handled manually where necessary with
9211 some primitive wrappers for PKCS#7. The new functions behave in a
9212 manner analogous to the X509 extension functions: they allow
9213 attributes to be looked up by NID and added.
9215 Later something similar to the X509V3 code would be desirable to
9216 automatically handle the encoding, decoding and printing of the
9217 more complex types. The string types like challengePassword can
9218 be handled by the string table functions.
9220 Also modified the multi byte string table handling. Now there is
9221 a 'global mask' which masks out certain types. The table itself
9222 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9223 is useful when for example there is only one permissible type
9224 (as in countryName) and using the mask might result in no valid
9228 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9229 SSL_get_peer_finished to allow applications to obtain the latest
9230 Finished messages sent to the peer or expected from the peer,
9231 respectively. (SSL_get_peer_finished is usually the Finished message
9232 actually received from the peer, otherwise the protocol will be aborted.)
9234 As the Finished message are message digests of the complete handshake
9235 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9236 be used for external authentication procedures when the authentication
9237 provided by SSL/TLS is not desired or is not enough.
9240 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9241 the host supports BWX extension and if Compaq C is present on the
9242 $PATH. Just exploiting of the BWX extension results in 20-30%
9243 performance kick for some algorithms, e.g. DES and RC4 to mention
9244 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9248 *) Add support for MS "fast SGC". This is arguably a violation of the
9249 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9250 weak crypto and after checking the certificate is SGC a second one
9251 with strong crypto. MS SGC stops the first handshake after receiving
9252 the server certificate message and sends a second client hello. Since
9253 a server will typically do all the time consuming operations before
9254 expecting any further messages from the client (server key exchange
9255 is the most expensive) there is little difference between the two.
9257 To get OpenSSL to support MS SGC we have to permit a second client
9258 hello message after we have sent server done. In addition we have to
9259 reset the MAC if we do get this second client hello.
9262 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9263 if a DER encoded private key is RSA or DSA traditional format. Changed
9264 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9265 format DER encoded private key. Newer code should use PKCS#8 format which
9266 has the key type encoded in the ASN1 structure. Added DER private key
9267 support to pkcs8 application.
9270 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9271 ciphersuites has been selected (as required by the SSL 3/TLS 1
9272 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9273 is set, we interpret this as a request to violate the specification
9274 (the worst that can happen is a handshake failure, and 'correct'
9275 behaviour would result in a handshake failure anyway).
9278 *) In SSL_CTX_add_session, take into account that there might be multiple
9279 SSL_SESSION structures with the same session ID (e.g. when two threads
9280 concurrently obtain them from an external cache).
9281 The internal cache can handle only one SSL_SESSION with a given ID,
9282 so if there's a conflict, we now throw out the old one to achieve
9286 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9287 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9288 some routines that use cipher OIDs: some ciphers do not have OIDs
9289 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9293 *) Simplify the trust setting structure and code. Now we just have
9294 two sequences of OIDs for trusted and rejected settings. These will
9295 typically have values the same as the extended key usage extension
9296 and any application specific purposes.
9298 The trust checking code now has a default behaviour: it will just
9299 check for an object with the same NID as the passed id. Functions can
9300 be provided to override either the default behaviour or the behaviour
9301 for a given id. SSL client, server and email already have functions
9302 in place for compatibility: they check the NID and also return "trusted"
9303 if the certificate is self signed.
9306 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9307 traditional format into an EVP_PKEY structure.
9310 *) Add a password callback function PEM_cb() which either prompts for
9311 a password if usr_data is NULL or otherwise assumes it is a null
9312 terminated password. Allow passwords to be passed on command line
9313 environment or config files in a few more utilities.
9316 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9317 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9318 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9319 Update documentation.
9322 *) Support for ASN1 "NULL" type. This could be handled before by using
9323 ASN1_TYPE but there wasn't any function that would try to read a NULL
9324 and produce an error if it couldn't. For compatibility we also have
9325 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9326 don't allocate anything because they don't need to.
9329 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9331 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9333 *) Rebuild of the memory allocation routines used by OpenSSL code and
9334 possibly others as well. The purpose is to make an interface that
9335 provide hooks so anyone can build a separate set of allocation and
9336 deallocation routines to be used by OpenSSL, for example memory
9337 pool implementations, or something else, which was previously hard
9338 since Malloc(), Realloc() and Free() were defined as macros having
9339 the values malloc, realloc and free, respectively (except for Win32
9340 compilations). The same is provided for memory debugging code.
9341 OpenSSL already comes with functionality to find memory leaks, but
9342 this gives people a chance to debug other memory problems.
9344 With these changes, a new set of functions and macros have appeared:
9346 CRYPTO_set_mem_debug_functions() [F]
9347 CRYPTO_get_mem_debug_functions() [F]
9348 CRYPTO_dbg_set_options() [F]
9349 CRYPTO_dbg_get_options() [F]
9350 CRYPTO_malloc_debug_init() [M]
9352 The memory debug functions are NULL by default, unless the library
9353 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9354 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9355 gives the standard debugging functions that come with OpenSSL) or
9356 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9357 provided by the library user) must be used. When the standard
9358 debugging functions are used, CRYPTO_dbg_set_options can be used to
9359 request additional information:
9360 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9361 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9363 Also, things like CRYPTO_set_mem_functions will always give the
9364 expected result (the new set of functions is used for allocation
9365 and deallocation) at all times, regardless of platform and compiler
9368 To finish it up, some functions that were never use in any other
9369 way than through macros have a new API and new semantic:
9372 CRYPTO_dbg_realloc()
9375 All macros of value have retained their old syntax.
9376 [Richard Levitte and Bodo Moeller]
9378 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9379 ordering of SMIMECapabilities wasn't in "strength order" and there
9380 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9384 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9385 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9386 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9388 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9389 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9390 functionality to handle multipart/signed properly) and a utility
9391 called 'smime' to call all this stuff. This is based on code I
9392 originally wrote for Celo who have kindly allowed it to be
9393 included in OpenSSL.
9396 *) Add variants des_set_key_checked and des_set_key_unchecked of
9397 des_set_key (aka des_key_sched). Global variable des_check_key
9398 decides which of these is called by des_set_key; this way
9399 des_check_key behaves as it always did, but applications and
9400 the library itself, which was buggy for des_check_key == 1,
9401 have a cleaner way to pick the version they need.
9404 *) New function PKCS12_newpass() which changes the password of a
9408 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9409 dynamic mix. In both cases the ids can be used as an index into the
9410 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9411 functions so they accept a list of the field values and the
9412 application doesn't need to directly manipulate the X509_TRUST
9416 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9420 *) Modify the way the V3 extension code looks up extensions. This now
9421 works in a similar way to the object code: we have some "standard"
9422 extensions in a static table which is searched with OBJ_bsearch()
9423 and the application can add dynamic ones if needed. The file
9424 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9425 updated whenever a new extension is added to the core code and kept
9426 in ext_nid order. There is a simple program 'tabtest.c' which checks
9427 this. New extensions are not added too often so this file can readily
9428 be maintained manually.
9430 There are two big advantages in doing things this way. The extensions
9431 can be looked up immediately and no longer need to be "added" using
9432 X509V3_add_standard_extensions(): this function now does nothing.
9433 [Side note: I get *lots* of email saying the extension code doesn't
9434 work because people forget to call this function]
9435 Also no dynamic allocation is done unless new extensions are added:
9436 so if we don't add custom extensions there is no need to call
9437 X509V3_EXT_cleanup().
9440 *) Modify enc utility's salting as follows: make salting the default. Add a
9441 magic header, so unsalted files fail gracefully instead of just decrypting
9442 to garbage. This is because not salting is a big security hole, so people
9443 should be discouraged from doing it.
9446 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9447 digest to be passed on the command line but it only used this
9448 parameter when signing a certificate. Modified so all relevant
9449 operations are affected by the digest parameter including the
9450 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9451 DSA key was used because it didn't fix the digest.
9454 *) Initial certificate chain verify code. Currently tests the untrusted
9455 certificates for consistency with the verify purpose (which is set
9456 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9458 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9459 this is because it will reject chains with invalid extensions whereas
9460 every previous version of OpenSSL and SSLeay made no checks at all.
9462 Trust code: checks the root CA for the relevant trust settings. Trust
9463 settings have an initial value consistent with the verify purpose: e.g.
9464 if the verify purpose is for SSL client use it expects the CA to be
9465 trusted for SSL client use. However the default value can be changed to
9466 permit custom trust settings: one example of this would be to only trust
9467 certificates from a specific "secure" set of CAs.
9469 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9470 which should be used for version portability: especially since the
9471 verify structure is likely to change more often now.
9473 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9474 to set them. If not set then assume SSL clients will verify SSL servers
9477 Two new options to the verify program: -untrusted allows a set of
9478 untrusted certificates to be passed in and -purpose which sets the
9479 intended purpose of the certificate. If a purpose is set then the
9480 new chain verify code is used to check extension consistency.
9483 *) Support for the authority information access extension.
9486 *) Modify RSA and DSA PEM read routines to transparently handle
9487 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9488 public keys in a format compatible with certificate
9489 SubjectPublicKeyInfo structures. Unfortunately there were already
9490 functions called *_PublicKey_* which used various odd formats so
9491 these are retained for compatibility: however the DSA variants were
9492 never in a public release so they have been deleted. Changed dsa/rsa
9493 utilities to handle the new format: note no releases ever handled public
9494 keys so we should be OK.
9496 The primary motivation for this change is to avoid the same fiasco
9497 that dogs private keys: there are several incompatible private key
9498 formats some of which are standard and some OpenSSL specific and
9499 require various evil hacks to allow partial transparent handling and
9500 even then it doesn't work with DER formats. Given the option anything
9501 other than PKCS#8 should be dumped: but the other formats have to
9502 stay in the name of compatibility.
9504 With public keys and the benefit of hindsight one standard format
9505 is used which works with EVP_PKEY, RSA or DSA structures: though
9506 it clearly returns an error if you try to read the wrong kind of key.
9508 Added a -pubkey option to the 'x509' utility to output the public key.
9509 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9510 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9511 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9512 that do the same as the EVP_PKEY_assign_*() except they up the
9513 reference count of the added key (they don't "swallow" the
9517 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9518 CRLs would fail if the file contained no certificates or no CRLs:
9519 added a new function to read in both types and return the number
9520 read: this means that if none are read it will be an error. The
9521 DER versions of the certificate and CRL reader would always fail
9522 because it isn't possible to mix certificates and CRLs in DER format
9523 without choking one or the other routine. Changed this to just read
9524 a certificate: this is the best we can do. Also modified the code
9525 in apps/verify.c to take notice of return codes: it was previously
9526 attempting to read in certificates from NULL pointers and ignoring
9527 any errors: this is one reason why the cert and CRL reader seemed
9528 to work. It doesn't check return codes from the default certificate
9529 routines: these may well fail if the certificates aren't installed.
9532 *) Code to support otherName option in GeneralName.
9535 *) First update to verify code. Change the verify utility
9536 so it warns if it is passed a self signed certificate:
9537 for consistency with the normal behaviour. X509_verify
9538 has been modified to it will now verify a self signed
9539 certificate if *exactly* the same certificate appears
9540 in the store: it was previously impossible to trust a
9541 single self signed certificate. This means that:
9542 openssl verify ss.pem
9543 now gives a warning about a self signed certificate but
9544 openssl verify -CAfile ss.pem ss.pem
9548 *) For servers, store verify_result in SSL_SESSION data structure
9549 (and add it to external session representation).
9550 This is needed when client certificate verifications fails,
9551 but an application-provided verification callback (set by
9552 SSL_CTX_set_cert_verify_callback) allows accepting the session
9553 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9554 but returns 1): When the session is reused, we have to set
9555 ssl->verify_result to the appropriate error code to avoid
9557 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9559 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9560 case in PKCS7_dataInit() where the signed PKCS7 structure
9561 didn't contain any existing data because it was being created.
9562 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9564 *) Add a salt to the key derivation routines in enc.c. This
9565 forms the first 8 bytes of the encrypted file. Also add a
9566 -S option to allow a salt to be input on the command line.
9569 *) New function X509_cmp(). Oddly enough there wasn't a function
9570 to compare two certificates. We do this by working out the SHA1
9571 hash and comparing that. X509_cmp() will be needed by the trust
9575 *) SSL_get1_session() is like SSL_get_session(), but increments
9576 the reference count in the SSL_SESSION returned.
9577 [Geoff Thorpe <geoff@eu.c2.net>]
9579 *) Fix for 'req': it was adding a null to request attributes.
9580 Also change the X509_LOOKUP and X509_INFO code to handle
9581 certificate auxiliary information.
9584 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9588 *) Add the possibility to add extra information to the memory leak
9589 detecting output, to form tracebacks, showing from where each
9590 allocation was originated: CRYPTO_push_info("constant string") adds
9591 the string plus current file name and line number to a per-thread
9592 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9593 is like calling CYRPTO_pop_info() until the stack is empty.
9594 Also updated memory leak detection code to be multi-thread-safe.
9597 *) Add options -text and -noout to pkcs7 utility and delete the
9598 encryption options which never did anything. Update docs.
9601 *) Add options to some of the utilities to allow the pass phrase
9602 to be included on either the command line (not recommended on
9603 OSes like Unix) or read from the environment. Update the
9604 manpages and fix a few bugs.
9607 *) Add a few manpages for some of the openssl commands.
9610 *) Fix the -revoke option in ca. It was freeing up memory twice,
9611 leaking and not finding already revoked certificates.
9614 *) Extensive changes to support certificate auxiliary information.
9615 This involves the use of X509_CERT_AUX structure and X509_AUX
9616 functions. An X509_AUX function such as PEM_read_X509_AUX()
9617 can still read in a certificate file in the usual way but it
9618 will also read in any additional "auxiliary information". By
9619 doing things this way a fair degree of compatibility can be
9620 retained: existing certificates can have this information added
9621 using the new 'x509' options.
9623 Current auxiliary information includes an "alias" and some trust
9624 settings. The trust settings will ultimately be used in enhanced
9625 certificate chain verification routines: currently a certificate
9626 can only be trusted if it is self signed and then it is trusted
9630 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9631 The problem was that one of the replacement routines had not been working
9632 since SSLeay releases. For now the offending routine has been replaced
9633 with non-optimised assembler. Even so, this now gives around 95%
9634 performance improvement for 1024 bit RSA signs.
9637 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9638 handling. Most clients have the effective key size in bits equal to
9639 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9640 A few however don't do this and instead use the size of the decrypted key
9641 to determine the RC2 key length and the AlgorithmIdentifier to determine
9642 the effective key length. In this case the effective key length can still
9643 be 40 bits but the key length can be 168 bits for example. This is fixed
9644 by manually forcing an RC2 key into the EVP_PKEY structure because the
9645 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9646 the key length and effective key length are equal.
9649 *) Add a bunch of functions that should simplify the creation of
9650 X509_NAME structures. Now you should be able to do:
9651 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9652 and have it automatically work out the correct field type and fill in
9653 the structures. The more adventurous can try:
9654 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9655 and it will (hopefully) work out the correct multibyte encoding.
9658 *) Change the 'req' utility to use the new field handling and multibyte
9659 copy routines. Before the DN field creation was handled in an ad hoc
9660 way in req, ca, and x509 which was rather broken and didn't support
9661 BMPStrings or UTF8Strings. Since some software doesn't implement
9662 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9663 using the dirstring_type option. See the new comment in the default
9664 openssl.cnf for more info.
9667 *) Make crypto/rand/md_rand.c more robust:
9668 - Assure unique random numbers after fork().
9669 - Make sure that concurrent threads access the global counter and
9670 md serializably so that we never lose entropy in them
9671 or use exactly the same state in multiple threads.
9672 Access to the large state is not always serializable because
9673 the additional locking could be a performance killer, and
9674 md should be large enough anyway.
9677 *) New file apps/app_rand.c with commonly needed functionality
9678 for handling the random seed file.
9680 Use the random seed file in some applications that previously did not:
9682 dsaparam -genkey (which also ignored its '-rand' option),
9685 x509 (when signing).
9686 Except on systems with /dev/urandom, it is crucial to have a random
9687 seed file at least for key creation, DSA signing, and for DH exchanges;
9688 for RSA signatures we could do without one.
9690 gendh and gendsa (unlike genrsa) used to read only the first byte
9691 of each file listed in the '-rand' option. The function as previously
9692 found in genrsa is now in app_rand.c and is used by all programs
9693 that support '-rand'.
9696 *) In RAND_write_file, use mode 0600 for creating files;
9697 don't just chmod when it may be too late.
9700 *) Report an error from X509_STORE_load_locations
9701 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9704 *) New function ASN1_mbstring_copy() this copies a string in either
9705 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9706 into an ASN1_STRING type. A mask of permissible types is passed
9707 and it chooses the "minimal" type to use or an error if not type
9711 *) Add function equivalents to the various macros in asn1.h. The old
9712 macros are retained with an M_ prefix. Code inside the library can
9713 use the M_ macros. External code (including the openssl utility)
9714 should *NOT* in order to be "shared library friendly".
9717 *) Add various functions that can check a certificate's extensions
9718 to see if it usable for various purposes such as SSL client,
9719 server or S/MIME and CAs of these types. This is currently
9720 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9721 verification. Also added a -purpose flag to x509 utility to
9722 print out all the purposes.
9725 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9729 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9730 for, obtain and decode and extension and obtain its critical flag.
9731 This allows all the necessary extension code to be handled in a
9732 single function call.
9735 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9736 platforms. See crypto/rc4/rc4_enc.c for further details.
9739 *) New -noout option to asn1parse. This causes no output to be produced
9740 its main use is when combined with -strparse and -out to extract data
9741 from a file (which may not be in ASN.1 format).
9744 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9745 when producing the local key id.
9746 [Richard Levitte <levitte@stacken.kth.se>]
9748 *) New option -dhparam in s_server. This allows a DH parameter file to be
9749 stated explicitly. If it is not stated then it tries the first server
9750 certificate file. The previous behaviour hard coded the filename
9754 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9755 a public key to be input or output. For example:
9756 openssl rsa -in key.pem -pubout -out pubkey.pem
9757 Also added necessary DSA public key functions to handle this.
9760 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9761 in the message. This was handled by allowing
9762 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9763 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9765 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9766 to the end of the strings whereas this didn't. This would cause problems
9767 if strings read with d2i_ASN1_bytes() were later modified.
9768 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9770 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9771 data and it contains EOF it will end up returning an error. This is
9772 caused by input 46 bytes long. The cause is due to the way base64
9773 BIOs find the start of base64 encoded data. They do this by trying a
9774 trial decode on each line until they find one that works. When they
9775 do a flag is set and it starts again knowing it can pass all the
9776 data directly through the decoder. Unfortunately it doesn't reset
9777 the context it uses. This means that if EOF is reached an attempt
9778 is made to pass two EOFs through the context and this causes the
9779 resulting error. This can also cause other problems as well. As is
9780 usual with these problems it takes *ages* to find and the fix is
9781 trivial: move one line.
9782 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9784 *) Ugly workaround to get s_client and s_server working under Windows. The
9785 old code wouldn't work because it needed to select() on sockets and the
9786 tty (for keypresses and to see if data could be written). Win32 only
9787 supports select() on sockets so we select() with a 1s timeout on the
9788 sockets and then see if any characters are waiting to be read, if none
9789 are present then we retry, we also assume we can always write data to
9790 the tty. This isn't nice because the code then blocks until we've
9791 received a complete line of data and it is effectively polling the
9792 keyboard at 1s intervals: however it's quite a bit better than not
9793 working at all :-) A dedicated Windows application might handle this
9794 with an event loop for example.
9797 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9798 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9799 will be called when RSA_sign() and RSA_verify() are used. This is useful
9800 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9801 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9802 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9803 This necessitated the support of an extra signature type NID_md5_sha1
9804 for SSL signatures and modifications to the SSL library to use it instead
9805 of calling RSA_public_decrypt() and RSA_private_encrypt().
9808 *) Add new -verify -CAfile and -CApath options to the crl program, these
9809 will lookup a CRL issuers certificate and verify the signature in a
9810 similar way to the verify program. Tidy up the crl program so it
9811 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9812 less strict. It will now permit CRL extensions even if it is not
9813 a V2 CRL: this will allow it to tolerate some broken CRLs.
9816 *) Initialize all non-automatic variables each time one of the openssl
9817 sub-programs is started (this is necessary as they may be started
9818 multiple times from the "OpenSSL>" prompt).
9819 [Lennart Bang, Bodo Moeller]
9821 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9822 removing all other RSA functionality (this is what NO_RSA does). This
9823 is so (for example) those in the US can disable those operations covered
9824 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9828 *) Non-copying interface to BIO pairs.
9829 (still largely untested)
9832 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9833 ASCII string. This was handled independently in various places before.
9836 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9837 UTF8 strings a character at a time.
9840 *) Use client_version from client hello to select the protocol
9841 (s23_srvr.c) and for RSA client key exchange verification
9842 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9845 *) Add various utility functions to handle SPKACs, these were previously
9846 handled by poking round in the structure internals. Added new function
9847 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9848 print, verify and generate SPKACs. Based on an original idea from
9849 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9852 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9855 *) Allow the config file extension section to be overwritten on the
9856 command line. Based on an original idea from Massimiliano Pala
9857 <madwolf@comune.modena.it>. The new option is called -extensions
9858 and can be applied to ca, req and x509. Also -reqexts to override
9859 the request extensions in req and -crlexts to override the crl extensions
9863 *) Add new feature to the SPKAC handling in ca. Now you can include
9864 the same field multiple times by preceding it by "XXXX." for example:
9867 this is the same syntax as used in the req config file.
9870 *) Allow certificate extensions to be added to certificate requests. These
9871 are specified in a 'req_extensions' option of the req section of the
9872 config file. They can be printed out with the -text option to req but
9873 are otherwise ignored at present.
9876 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9877 data read consists of only the final block it would not decrypted because
9878 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9879 A misplaced 'break' also meant the decrypted final block might not be
9880 copied until the next read.
9883 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9884 a few extra parameters to the DH structure: these will be useful if
9885 for example we want the value of 'q' or implement X9.42 DH.
9888 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9889 provides hooks that allow the default DSA functions or functions on a
9890 "per key" basis to be replaced. This allows hardware acceleration and
9891 hardware key storage to be handled without major modification to the
9892 library. Also added low level modexp hooks and CRYPTO_EX structure and
9893 associated functions.
9896 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9897 as "read only": it can't be written to and the buffer it points to will
9898 not be freed. Reading from a read only BIO is much more efficient than
9899 a normal memory BIO. This was added because there are several times when
9900 an area of memory needs to be read from a BIO. The previous method was
9901 to create a memory BIO and write the data to it, this results in two
9902 copies of the data and an O(n^2) reading algorithm. There is a new
9903 function BIO_new_mem_buf() which creates a read only memory BIO from
9904 an area of memory. Also modified the PKCS#7 routines to use read only
9908 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9909 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9910 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9911 but a retry condition occured while trying to read the rest.
9914 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9915 NID_pkcs7_encrypted by default: this was wrong since this should almost
9916 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9917 the encrypted data type: this is a more sensible place to put it and it
9918 allows the PKCS#12 code to be tidied up that duplicated this
9922 *) Changed obj_dat.pl script so it takes its input and output files on
9923 the command line. This should avoid shell escape redirection problems
9927 *) Initial support for certificate extension requests, these are included
9928 in things like Xenroll certificate requests. Included functions to allow
9929 extensions to be obtained and added.
9932 *) -crlf option to s_client and s_server for sending newlines as
9933 CRLF (as required by many protocols).
9936 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9938 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9939 [Ralf S. Engelschall]
9941 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9942 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9944 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9948 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9949 DH parameters/keys (q is lost during that conversion, but the resulting
9950 DH parameters contain its length).
9952 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9953 much faster than DH_generate_parameters (which creates parameters
9954 where p = 2*q + 1), and also the smaller q makes DH computations
9955 much more efficient (160-bit exponentiation instead of 1024-bit
9956 exponentiation); so this provides a convenient way to support DHE
9957 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9958 utter importance to use
9959 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9961 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9962 when such DH parameters are used, because otherwise small subgroup
9963 attacks may become possible!
9966 *) Avoid memory leak in i2d_DHparams.
9969 *) Allow the -k option to be used more than once in the enc program:
9970 this allows the same encrypted message to be read by multiple recipients.
9973 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9974 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9975 it will always use the numerical form of the OID, even if it has a short
9979 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9980 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9981 otherwise bn_mod_exp was called. In the case of hardware keys for example
9982 no private key components need be present and it might store extra data
9983 in the RSA structure, which cannot be accessed from bn_mod_exp.
9984 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9985 private key operations.
9988 *) Added support for SPARC Linux.
9991 *) pem_password_cb function type incompatibly changed from
9992 typedef int pem_password_cb(char *buf, int size, int rwflag);
9994 ....(char *buf, int size, int rwflag, void *userdata);
9995 so that applications can pass data to their callbacks:
9996 The PEM[_ASN1]_{read,write}... functions and macros now take an
9997 additional void * argument, which is just handed through whenever
9998 the password callback is called.
9999 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10001 New function SSL_CTX_set_default_passwd_cb_userdata.
10003 Compatibility note: As many C implementations push function arguments
10004 onto the stack in reverse order, the new library version is likely to
10005 interoperate with programs that have been compiled with the old
10006 pem_password_cb definition (PEM_whatever takes some data that
10007 happens to be on the stack as its last argument, and the callback
10008 just ignores this garbage); but there is no guarantee whatsoever that
10011 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10012 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10013 problems not only on Windows, but also on some Unix platforms.
10014 To avoid problematic command lines, these definitions are now in an
10015 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10016 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10019 *) MIPS III/IV assembler module is reimplemented.
10022 *) More DES library cleanups: remove references to srand/rand and
10023 delete an unused file.
10026 *) Add support for the the free Netwide assembler (NASM) under Win32,
10027 since not many people have MASM (ml) and it can be hard to obtain.
10028 This is currently experimental but it seems to work OK and pass all
10029 the tests. Check out INSTALL.W32 for info.
10032 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10033 without temporary keys kept an extra copy of the server key,
10034 and connections with temporary keys did not free everything in case
10038 *) New function RSA_check_key and new openssl rsa option -check
10039 for verifying the consistency of RSA keys.
10040 [Ulf Moeller, Bodo Moeller]
10042 *) Various changes to make Win32 compile work:
10043 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10044 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10045 comparison" warnings.
10046 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10049 *) Add a debugging option to PKCS#5 v2 key generation function: when
10050 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10051 derived keys are printed to stderr.
10054 *) Copy the flags in ASN1_STRING_dup().
10055 [Roman E. Pavlov <pre@mo.msk.ru>]
10057 *) The x509 application mishandled signing requests containing DSA
10058 keys when the signing key was also DSA and the parameters didn't match.
10060 It was supposed to omit the parameters when they matched the signing key:
10061 the verifying software was then supposed to automatically use the CA's
10062 parameters if they were absent from the end user certificate.
10064 Omitting parameters is no longer recommended. The test was also
10065 the wrong way round! This was probably due to unusual behaviour in
10066 EVP_cmp_parameters() which returns 1 if the parameters match.
10067 This meant that parameters were omitted when they *didn't* match and
10068 the certificate was useless. Certificates signed with 'ca' didn't have
10070 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10072 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10073 The interface is as follows:
10074 Applications can use
10075 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10076 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10077 "off" is now the default.
10078 The library internally uses
10079 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10080 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10081 to disable memory-checking temporarily.
10083 Some inconsistent states that previously were possible (and were
10084 even the default) are now avoided.
10086 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10087 with each memory chunk allocated; this is occasionally more helpful
10088 than just having a counter.
10090 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10092 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10096 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10097 which largely parallels "options", but is for changing API behaviour,
10098 whereas "options" are about protocol behaviour.
10099 Initial "mode" flags are:
10101 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10102 a single record has been written.
10103 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10104 retries use the same buffer location.
10105 (But all of the contents must be
10109 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10112 *) Fix problems with no-hmac etc.
10113 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10115 *) New functions RSA_get_default_method(), RSA_set_method() and
10116 RSA_get_method(). These allows replacement of RSA_METHODs without having
10117 to mess around with the internals of an RSA structure.
10120 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10121 Also really enable memory leak checks in openssl.c and in some
10123 [Chad C. Mulligan, Bodo Moeller]
10125 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10126 up the length of negative integers. This has now been simplified to just
10127 store the length when it is first determined and use it later, rather
10128 than trying to keep track of where data is copied and updating it to
10130 [Steve Henson, reported by Brien Wheeler
10131 <bwheeler@authentica-security.com>]
10133 *) Add a new function PKCS7_signatureVerify. This allows the verification
10134 of a PKCS#7 signature but with the signing certificate passed to the
10135 function itself. This contrasts with PKCS7_dataVerify which assumes the
10136 certificate is present in the PKCS#7 structure. This isn't always the
10137 case: certificates can be omitted from a PKCS#7 structure and be
10138 distributed by "out of band" means (such as a certificate database).
10141 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10142 function prototypes in pem.h, also change util/mkdef.pl to add the
10143 necessary function names.
10146 *) mk1mf.pl (used by Windows builds) did not properly read the
10147 options set by Configure in the top level Makefile, and Configure
10148 was not even able to write more than one option correctly.
10149 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10152 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10153 file to be loaded from a BIO or FILE pointer. The BIO version will
10154 for example allow memory BIOs to contain config info.
10157 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10158 Whoever hopes to achieve shared-library compatibility across versions
10159 must use this, not the compile-time macro.
10160 (Exercise 0.9.4: Which is the minimum library version required by
10162 Note: All this applies only to multi-threaded programs, others don't
10166 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10167 through a BIO pair triggered the default case, i.e.
10168 SSLerr(...,SSL_R_UNKNOWN_STATE).
10171 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10172 can use the SSL library even if none of the specific BIOs is
10176 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10177 for the encoded length.
10178 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10180 *) Add initial documentation of the X509V3 functions.
10183 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10184 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10185 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10186 secure PKCS#8 private key format with a high iteration count.
10189 *) Fix determination of Perl interpreter: A perl or perl5
10190 _directory_ in $PATH was also accepted as the interpreter.
10191 [Ralf S. Engelschall]
10193 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10194 wrong with it but it was very old and did things like calling
10195 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10196 unusual formatting.
10199 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10200 to use the new extension code.
10203 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10204 with macros. This should make it easier to change their form, add extra
10205 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10209 *) Add to configuration table a new entry that can specify an alternative
10210 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10211 according to Mark Crispin <MRC@Panda.COM>.
10215 *) DES CBC did not update the IV. Weird.
10218 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10219 Changing the behaviour of the former might break existing programs --
10220 where IV updating is needed, des_ncbc_encrypt can be used.
10223 *) When bntest is run from "make test" it drives bc to check its
10224 calculations, as well as internally checking them. If an internal check
10225 fails, it needs to cause bc to give a non-zero result or make test carries
10226 on without noticing the failure. Fixed.
10229 *) DES library cleanups.
10232 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10233 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10234 ciphers. NOTE: although the key derivation function has been verified
10235 against some published test vectors it has not been extensively tested
10236 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10240 *) Instead of "mkdir -p", which is not fully portable, use new
10241 Perl script "util/mkdir-p.pl".
10244 *) Rewrite the way password based encryption (PBE) is handled. It used to
10245 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10246 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10247 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10248 the 'parameter' field of the AlgorithmIdentifier is passed to the
10249 underlying key generation function so it must do its own ASN1 parsing.
10250 This has also changed the EVP_PBE_CipherInit() function which now has a
10251 'parameter' argument instead of literal salt and iteration count values
10252 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10255 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10256 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10257 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10258 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10259 value was just used as a "magic string" and not used directly its
10260 value doesn't matter.
10263 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10267 *) "linux-sparc64" configuration (ultrapenguin).
10268 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10269 "linux-sparc" configuration.
10270 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10272 *) config now generates no-xxx options for missing ciphers.
10275 *) Support the EBCDIC character set (work in progress).
10276 File ebcdic.c not yet included because it has a different license.
10277 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10279 *) Support BS2000/OSD-POSIX.
10280 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10282 *) Make callbacks for key generation use void * instead of char *.
10285 *) Make S/MIME samples compile (not yet tested).
10288 *) Additional typesafe stacks.
10291 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10295 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10297 *) New configuration variant "sco5-gcc".
10299 *) Updated some demos.
10300 [Sean O Riordain, Wade Scholine]
10302 *) Add missing BIO_free at exit of pkcs12 application.
10305 *) Fix memory leak in conf.c.
10308 *) Updates for Win32 to assembler version of MD5.
10311 *) Set #! path to perl in apps/der_chop to where we found it
10312 instead of using a fixed path.
10315 *) SHA library changes for irix64-mips4-cc.
10318 *) Improvements for VMS support.
10322 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10324 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10325 This also avoids the problems with SC4.2 and unpatched SC5.
10326 [Andy Polyakov <appro@fy.chalmers.se>]
10328 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10329 These are required because of the typesafe stack would otherwise break
10330 existing code. If old code used a structure member which used to be STACK
10331 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10332 sk_num or sk_value it would produce an error because the num, data members
10333 are not present in STACK_OF. Now it just produces a warning. sk_set
10334 replaces the old method of assigning a value to sk_value
10335 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10336 that does this will no longer work (and should use sk_set instead) but
10337 this could be regarded as a "questionable" behaviour anyway.
10340 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10341 correctly handle encrypted S/MIME data.
10344 *) Change type of various DES function arguments from des_cblock
10345 (which means, in function argument declarations, pointer to char)
10346 to des_cblock * (meaning pointer to array with 8 char elements),
10347 which allows the compiler to do more typechecking; it was like
10348 that back in SSLeay, but with lots of ugly casts.
10350 Introduce new type const_des_cblock.
10353 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10354 problems: find RecipientInfo structure that matches recipient certificate
10355 and initialise the ASN1 structures properly based on passed cipher.
10358 *) Belatedly make the BN tests actually check the results.
10361 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10362 to and from BNs: it was completely broken. New compilation option
10363 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10364 key elements as negative integers.
10367 *) Reorganize and speed up MD5.
10368 [Andy Polyakov <appro@fy.chalmers.se>]
10371 [Richard Levitte <richard@levitte.org>]
10373 *) New option -out to asn1parse to allow the parsed structure to be
10374 output to a file. This is most useful when combined with the -strparse
10375 option to examine the output of things like OCTET STRINGS.
10378 *) Make SSL library a little more fool-proof by not requiring any longer
10379 that SSL_set_{accept,connect}_state be called before
10380 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10381 in many applications because usually everything *appeared* to work as
10382 intended anyway -- now it really works as intended).
10385 *) Move openssl.cnf out of lib/.
10388 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10389 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10390 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10391 [Ralf S. Engelschall]
10393 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10394 handle PKCS#7 enveloped data properly.
10395 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10397 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10398 copying pointers. The cert_st handling is changed by this in
10399 various ways (and thus what used to be known as ctx->default_cert
10400 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10401 any longer when s->cert does not give us what we need).
10402 ssl_cert_instantiate becomes obsolete by this change.
10403 As soon as we've got the new code right (possibly it already is?),
10404 we have solved a couple of bugs of the earlier code where s->cert
10405 was used as if it could not have been shared with other SSL structures.
10407 Note that using the SSL API in certain dirty ways now will result
10408 in different behaviour than observed with earlier library versions:
10409 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10410 does not influence s as it used to.
10412 In order to clean up things more thoroughly, inside SSL_SESSION
10413 we don't use CERT any longer, but a new structure SESS_CERT
10414 that holds per-session data (if available); currently, this is
10415 the peer's certificate chain and, for clients, the server's certificate
10416 and temporary key. CERT holds only those values that can have
10417 meaningful defaults in an SSL_CTX.
10420 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10421 from the internal representation. Various PKCS#7 fixes: remove some
10422 evil casts and set the enc_dig_alg field properly based on the signing
10426 *) Allow PKCS#12 password to be set from the command line or the
10427 environment. Let 'ca' get its config file name from the environment
10428 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10432 *) Allow certificate policies extension to use an IA5STRING for the
10433 organization field. This is contrary to the PKIX definition but
10434 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10438 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10439 without disallowing inline assembler and the like for non-pedantic builds.
10442 *) Support Borland C++ builder.
10443 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10445 *) Support Mingw32.
10448 *) SHA-1 cleanups and performance enhancements.
10449 [Andy Polyakov <appro@fy.chalmers.se>]
10451 *) Sparc v8plus assembler for the bignum library.
10452 [Andy Polyakov <appro@fy.chalmers.se>]
10454 *) Accept any -xxx and +xxx compiler options in Configure.
10457 *) Update HPUX configuration.
10460 *) Add missing sk_<type>_unshift() function to safestack.h
10461 [Ralf S. Engelschall]
10463 *) New function SSL_CTX_use_certificate_chain_file that sets the
10464 "extra_cert"s in addition to the certificate. (This makes sense
10465 only for "PEM" format files, as chains as a whole are not
10469 *) Support verify_depth from the SSL API.
10470 x509_vfy.c had what can be considered an off-by-one-error:
10471 Its depth (which was not part of the external interface)
10472 was actually counting the number of certificates in a chain;
10473 now it really counts the depth.
10476 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10477 instead of X509err, which often resulted in confusing error
10478 messages since the error codes are not globally unique
10479 (e.g. an alleged error in ssl3_accept when a certificate
10480 didn't match the private key).
10482 *) New function SSL_CTX_set_session_id_context that allows to set a default
10483 value (so that you don't need SSL_set_session_id_context for each
10484 connection using the SSL_CTX).
10487 *) OAEP decoding bug fix.
10490 *) Support INSTALL_PREFIX for package builders, as proposed by
10494 *) New Configure options "threads" and "no-threads". For systems
10495 where the proper compiler options are known (currently Solaris
10496 and Linux), "threads" is the default.
10499 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10502 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10503 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10504 such as /usr/local/bin.
10507 *) "make linux-shared" to build shared libraries.
10508 [Niels Poppe <niels@netbox.org>]
10510 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10513 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10514 extension adding in x509 utility.
10517 *) Remove NOPROTO sections and error code comments.
10520 *) Partial rewrite of the DEF file generator to now parse the ANSI
10524 *) New Configure options --prefix=DIR and --openssldir=DIR.
10527 *) Complete rewrite of the error code script(s). It is all now handled
10528 by one script at the top level which handles error code gathering,
10529 header rewriting and C source file generation. It should be much better
10530 than the old method: it now uses a modified version of Ulf's parser to
10531 read the ANSI prototypes in all header files (thus the old K&R definitions
10532 aren't needed for error creation any more) and do a better job of
10533 translating function codes into names. The old 'ASN1 error code imbedded
10534 in a comment' is no longer necessary and it doesn't use .err files which
10535 have now been deleted. Also the error code call doesn't have to appear all
10536 on one line (which resulted in some large lines...).
10539 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10542 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10543 0 (which usually indicates a closed connection), but continue reading.
10546 *) Fix some race conditions.
10549 *) Add support for CRL distribution points extension. Add Certificate
10550 Policies and CRL distribution points documentation.
10553 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10556 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10557 8 of keying material. Merlin has also confirmed interop with this fix
10558 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10559 [Merlin Hughes <merlin@baltimore.ie>]
10561 *) Fix lots of warnings.
10562 [Richard Levitte <levitte@stacken.kth.se>]
10564 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10565 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10566 [Richard Levitte <levitte@stacken.kth.se>]
10568 *) Fix problems with sizeof(long) == 8.
10569 [Andy Polyakov <appro@fy.chalmers.se>]
10571 *) Change functions to ANSI C.
10574 *) Fix typos in error codes.
10575 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10577 *) Remove defunct assembler files from Configure.
10580 *) SPARC v8 assembler BIGNUM implementation.
10581 [Andy Polyakov <appro@fy.chalmers.se>]
10583 *) Support for Certificate Policies extension: both print and set.
10584 Various additions to support the r2i method this uses.
10587 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10588 return a const string when you are expecting an allocated buffer.
10591 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10592 types DirectoryString and DisplayText.
10595 *) Add code to allow r2i extensions to access the configuration database,
10596 add an LHASH database driver and add several ctx helper functions.
10599 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10600 fail when they extended the size of a BIGNUM.
10603 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10604 support typesafe stack.
10607 *) Fix typo in SSL_[gs]et_options().
10608 [Nils Frostberg <nils@medcom.se>]
10610 *) Delete various functions and files that belonged to the (now obsolete)
10611 old X509V3 handling code.
10614 *) New Configure option "rsaref".
10617 *) Don't auto-generate pem.h.
10620 *) Introduce type-safe ASN.1 SETs.
10623 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10624 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10626 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10627 that links with OpenSSL (well at least cause lots of warnings), but fear
10628 not: the conversion is trivial, and it eliminates loads of evil casts. A
10629 few STACKed things have been converted already. Feel free to convert more.
10630 In the fullness of time, I'll do away with the STACK type altogether.
10633 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10634 specified in <certfile> by updating the entry in the index.txt file.
10635 This way one no longer has to edit the index.txt file manually for
10636 revoking a certificate. The -revoke option does the gory details now.
10637 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10639 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10640 `-text' option at all and this way the `-noout -text' combination was
10641 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10642 [Ralf S. Engelschall]
10644 *) Make sure a corresponding plain text error message exists for the
10645 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10646 verify callback function determined that a certificate was revoked.
10647 [Ralf S. Engelschall]
10649 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10650 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10651 all available cipers including rc5, which was forgotten until now.
10652 In order to let the testing shell script know which algorithms
10653 are available, a new (up to now undocumented) command
10654 "openssl list-cipher-commands" is used.
10657 *) Bugfix: s_client occasionally would sleep in select() when
10658 it should have checked SSL_pending() first.
10661 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10662 the raw DSA values prior to ASN.1 encoding.
10665 *) Tweaks to Configure
10666 [Niels Poppe <niels@netbox.org>]
10668 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10672 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10675 *) New config option to avoid instructions that are illegal on the 80386.
10676 The default code is faster, but requires at least a 486.
10679 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10680 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10681 same as SSL2_VERSION anyway.
10684 *) New "-showcerts" option for s_client.
10687 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10688 application. Various cleanups and fixes.
10691 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10692 modify error routines to work internally. Add error codes and PBE init
10693 to library startup routines.
10696 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10697 packing functions to asn1 and evp. Changed function names and error
10698 codes along the way.
10701 *) PKCS12 integration: and so it begins... First of several patches to
10702 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10703 objects to objects.h
10706 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10707 and display support for Thawte strong extranet extension.
10710 *) Add LinuxPPC support.
10711 [Jeff Dubrule <igor@pobox.org>]
10713 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10714 bn_div_words in alpha.s.
10715 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10717 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10718 OAEP isn't supported when OpenSSL is built with RSAref.
10719 [Ulf Moeller <ulf@fitug.de>]
10721 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10722 so they no longer are missing under -DNOPROTO.
10723 [Soren S. Jorvang <soren@t.dk>]
10726 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10728 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10729 doesn't work when the session is reused. Coming soon!
10732 *) Fix a security hole, that allows sessions to be reused in the wrong
10733 context thus bypassing client cert protection! All software that uses
10734 client certs and session caches in multiple contexts NEEDS PATCHING to
10735 allow session reuse! A fuller solution is in the works.
10736 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10738 *) Some more source tree cleanups (removed obsolete files
10739 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10740 permission on "config" script to be executable) and a fix for the INSTALL
10742 [Ulf Moeller <ulf@fitug.de>]
10744 *) Remove some legacy and erroneous uses of malloc, free instead of
10746 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10748 *) Make rsa_oaep_test return non-zero on error.
10749 [Ulf Moeller <ulf@fitug.de>]
10751 *) Add support for native Solaris shared libraries. Configure
10752 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10753 if someone would make that last step automatic.
10754 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10756 *) ctx_size was not built with the right compiler during "make links". Fixed.
10759 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10760 except NULL ciphers". This means the default cipher list will no longer
10761 enable NULL ciphers. They need to be specifically enabled e.g. with
10762 the string "DEFAULT:eNULL".
10765 *) Fix to RSA private encryption routines: if p < q then it would
10766 occasionally produce an invalid result. This will only happen with
10767 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10770 *) Be less restrictive and allow also `perl util/perlpath.pl
10771 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10772 because this way one can also use an interpreter named `perl5' (which is
10773 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10774 installed as `perl').
10775 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10777 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10778 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10780 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10781 advapi32.lib to Win32 build and change the pem test comparision
10782 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10783 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10784 and crypto/des/ede_cbcm_enc.c.
10787 *) DES quad checksum was broken on big-endian architectures. Fixed.
10790 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10791 Win32 test batch file so it (might) work again. The Win32 test batch file
10792 is horrible: I feel ill....
10795 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10796 in e_os.h. Audit of header files to check ANSI and non ANSI
10797 sections: 10 functions were absent from non ANSI section and not exported
10798 from Windows DLLs. Fixed up libeay.num for new functions.
10801 *) Make `openssl version' output lines consistent.
10802 [Ralf S. Engelschall]
10804 *) Fix Win32 symbol export lists for BIO functions: Added
10805 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10806 to ms/libeay{16,32}.def.
10807 [Ralf S. Engelschall]
10809 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10810 fine under Unix and passes some trivial tests I've now added. But the
10811 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10812 added to make sure no one expects that this stuff really works in the
10813 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10814 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10816 [Ralf S. Engelschall]
10818 *) Fix the generation of two part addresses in perl.
10819 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10821 *) Add config entry for Linux on MIPS.
10822 [John Tobey <jtobey@channel1.com>]
10824 *) Make links whenever Configure is run, unless we are on Windoze.
10827 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10828 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10832 *) Add a useful kludge to allow package maintainers to specify compiler and
10833 other platforms details on the command line without having to patch the
10834 Configure script everytime: One now can use ``perl Configure
10835 <id>:<details>'', i.e. platform ids are allowed to have details appended
10836 to them (seperated by colons). This is treated as there would be a static
10837 pre-configured entry in Configure's %table under key <id> with value
10838 <details> and ``perl Configure <id>'' is called. So, when you want to
10839 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10840 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10841 now, which overrides the FreeBSD-elf entry on-the-fly.
10842 [Ralf S. Engelschall]
10844 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10847 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10848 on the `perl Configure ...' command line. This way one can compile
10849 OpenSSL libraries with Position Independent Code (PIC) which is needed
10850 for linking it into DSOs.
10851 [Ralf S. Engelschall]
10853 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10857 *) Cleaned up the LICENSE document: The official contact for any license
10858 questions now is the OpenSSL core team under openssl-core@openssl.org.
10859 And add a paragraph about the dual-license situation to make sure people
10860 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10861 to the OpenSSL toolkit.
10862 [Ralf S. Engelschall]
10864 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10865 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10866 Additonally cleaned up the `make links' target: Remove unnecessary
10867 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10868 to speed processing and no longer clutter the display with confusing
10869 stuff. Instead only the actually done links are displayed.
10870 [Ralf S. Engelschall]
10872 *) Permit null encryption ciphersuites, used for authentication only. It used
10873 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10874 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10878 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10879 signed attributes when verifying signatures (this would break them),
10880 the detached data encoding was wrong and public keys obtained using
10881 X509_get_pubkey() weren't freed.
10884 *) Add text documentation for the BUFFER functions. Also added a work around
10885 to a Win95 console bug. This was triggered by the password read stuff: the
10886 last character typed gets carried over to the next fread(). If you were
10887 generating a new cert request using 'req' for example then the last
10888 character of the passphrase would be CR which would then enter the first
10892 *) Added the new `Includes OpenSSL Cryptography Software' button as
10893 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10894 button and can be used by applications based on OpenSSL to show the
10895 relationship to the OpenSSL project.
10896 [Ralf S. Engelschall]
10898 *) Remove confusing variables in function signatures in files
10899 ssl/ssl_lib.c and ssl/ssl.h.
10900 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10902 *) Don't install bss_file.c under PREFIX/include/
10903 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10905 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10906 functions that return function pointers and has support for NT specific
10907 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10908 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10909 unsigned to signed types: this was killing the Win32 compile.
10912 *) Add new certificate file to stack functions,
10913 SSL_add_dir_cert_subjects_to_stack() and
10914 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10915 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10916 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10917 This means that Apache-SSL and similar packages don't have to mess around
10918 to add as many CAs as they want to the preferred list.
10921 *) Experiment with doxygen documentation. Currently only partially applied to
10923 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10924 openssl.doxy as the configuration file.
10927 *) Get rid of remaining C++-style comments which strict C compilers hate.
10928 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10930 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10931 compiled in by default: it has problems with large keys.
10934 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10935 DH private keys and/or callback functions which directly correspond to
10936 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10937 is needed for applications which have to configure certificates on a
10938 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10940 For the RSA certificate situation is makes no difference, but
10941 for the DSA certificate situation this fixes the "no shared cipher"
10942 problem where the OpenSSL cipher selection procedure failed because the
10943 temporary keys were not overtaken from the context and the API provided
10944 no way to reconfigure them.
10945 The new functions now let applications reconfigure the stuff and they
10946 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10947 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10948 non-public-API function ssl_cert_instantiate() is used as a helper
10949 function and also to reduce code redundancy inside ssl_rsa.c.
10950 [Ralf S. Engelschall]
10952 *) Move s_server -dcert and -dkey options out of the undocumented feature
10953 area because they are useful for the DSA situation and should be
10954 recognized by the users.
10955 [Ralf S. Engelschall]
10957 *) Fix the cipher decision scheme for export ciphers: the export bits are
10958 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10959 SSL_EXP_MASK. So, the original variable has to be used instead of the
10960 already masked variable.
10961 [Richard Levitte <levitte@stacken.kth.se>]
10963 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10964 [Richard Levitte <levitte@stacken.kth.se>]
10966 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10967 from `int' to `unsigned int' because it's a length and initialized by
10968 EVP_DigestFinal() which expects an `unsigned int *'.
10969 [Richard Levitte <levitte@stacken.kth.se>]
10971 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10972 script. Instead use the usual Shell->Perl transition trick.
10973 [Ralf S. Engelschall]
10975 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10976 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10977 -noout -modulus' as it's already the case for `openssl rsa -noout
10978 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10979 currently the public key is printed (a decision which was already done by
10980 `openssl dsa -modulus' in the past) which serves a similar purpose.
10981 Additionally the NO_RSA no longer completely removes the whole -modulus
10982 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10984 [Ralf S. Engelschall]
10986 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10987 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10988 [Arne Ansper <arne@ats.cyber.ee>]
10990 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10991 to be added. Now both 'req' and 'ca' can use new objects defined in the
10995 *) Add cool BIO that does syslog (or event log on NT).
10996 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10998 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10999 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11000 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11001 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11004 *) Add preliminary config info for new extension code.
11007 *) Make RSA_NO_PADDING really use no padding.
11008 [Ulf Moeller <ulf@fitug.de>]
11010 *) Generate errors when private/public key check is done.
11013 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11014 for some CRL extensions and new objects added.
11017 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11018 key usage extension and fuller support for authority key id.
11021 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11022 padding method for RSA, which is recommended for new applications in PKCS
11023 #1 v2.0 (RFC 2437, October 1998).
11024 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11025 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11026 against Bleichbacher's attack on RSA.
11027 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11030 *) Updates to the new SSL compression code
11031 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11033 *) Fix so that the version number in the master secret, when passed
11034 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11035 (because the server will not accept higher), that the version number
11036 is 0x03,0x01, not 0x03,0x00
11037 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11039 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11040 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11041 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11044 *) Support for RAW extensions where an arbitrary extension can be
11045 created by including its DER encoding. See apps/openssl.cnf for
11049 *) Make sure latest Perl versions don't interpret some generated C array
11050 code as Perl array code in the crypto/err/err_genc.pl script.
11051 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11053 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11054 not many people have the assembler. Various Win32 compilation fixes and
11055 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11056 build instructions.
11059 *) Modify configure script 'Configure' to automatically create crypto/date.h
11060 file under Win32 and also build pem.h from pem.org. New script
11061 util/mkfiles.pl to create the MINFO file on environments that can't do a
11062 'make files': perl util/mkfiles.pl >MINFO should work.
11065 *) Major rework of DES function declarations, in the pursuit of correctness
11066 and purity. As a result, many evil casts evaporated, and some weirdness,
11067 too. You may find this causes warnings in your code. Zapping your evil
11068 casts will probably fix them. Mostly.
11071 *) Fix for a typo in asn1.h. Bug fix to object creation script
11072 obj_dat.pl. It considered a zero in an object definition to mean
11073 "end of object": none of the objects in objects.h have any zeros
11074 so it wasn't spotted.
11075 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11077 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11078 Masking (CBCM). In the absence of test vectors, the best I have been able
11079 to do is check that the decrypt undoes the encrypt, so far. Send me test
11080 vectors if you have them.
11083 *) Correct calculation of key length for export ciphers (too much space was
11084 allocated for null ciphers). This has not been tested!
11087 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11088 message is now correct (it understands "crypto" and "ssl" on its
11089 command line). There is also now an "update" option. This will update
11090 the util/ssleay.num and util/libeay.num files with any new functions.
11092 perl util/mkdef.pl crypto ssl update
11093 it will update them.
11096 *) Overhauled the Perl interface (perl/*):
11097 - ported BN stuff to OpenSSL's different BN library
11098 - made the perl/ source tree CVS-aware
11099 - renamed the package from SSLeay to OpenSSL (the files still contain
11100 their history because I've copied them in the repository)
11101 - removed obsolete files (the test scripts will be replaced
11102 by better Test::Harness variants in the future)
11103 [Ralf S. Engelschall]
11105 *) First cut for a very conservative source tree cleanup:
11106 1. merge various obsolete readme texts into doc/ssleay.txt
11107 where we collect the old documents and readme texts.
11108 2. remove the first part of files where I'm already sure that we no
11109 longer need them because of three reasons: either they are just temporary
11110 files which were left by Eric or they are preserved original files where
11111 I've verified that the diff is also available in the CVS via "cvs diff
11112 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11113 the crypto/md/ stuff).
11114 [Ralf S. Engelschall]
11116 *) More extension code. Incomplete support for subject and issuer alt
11117 name, issuer and authority key id. Change the i2v function parameters
11118 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11119 what that's for :-) Fix to ASN1 macro which messed up
11120 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11123 *) Preliminary support for ENUMERATED type. This is largely copied from the
11127 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11128 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11130 *) Make sure `make rehash' target really finds the `openssl' program.
11131 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11133 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11134 like to hear about it if this slows down other processors.
11137 *) Add CygWin32 platform information to Configure script.
11138 [Alan Batie <batie@aahz.jf.intel.com>]
11140 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11141 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11143 *) New program nseq to manipulate netscape certificate sequences
11146 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11150 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11151 but the BN code had some problems that would cause failures when
11152 doing certificate verification and some other functions.
11153 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11155 *) Add ASN1 and PEM code to support netscape certificate sequences.
11158 *) Add ASN1 and PEM code to support netscape certificate sequences.
11161 *) Add several PKIX and private extended key usage OIDs.
11164 *) Modify the 'ca' program to handle the new extension code. Modify
11165 openssl.cnf for new extension format, add comments.
11168 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11169 and add a sample to openssl.cnf so req -x509 now adds appropriate
11173 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11174 error code, add initial support to X509_print() and x509 application.
11177 *) Takes a deep breath and start addding X509 V3 extension support code. Add
11178 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11179 stuff is currently isolated and isn't even compiled yet.
11182 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11183 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11184 Removed the versions check from X509 routines when loading extensions:
11185 this allows certain broken certificates that don't set the version
11186 properly to be processed.
11189 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11190 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11191 can still be regenerated with "make depend".
11194 *) Spelling mistake in C version of CAST-128.
11195 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11197 *) Changes to the error generation code. The perl script err-code.pl
11198 now reads in the old error codes and retains the old numbers, only
11199 adding new ones if necessary. It also only changes the .err files if new
11200 codes are added. The makefiles have been modified to only insert errors
11201 when needed (to avoid needlessly modifying header files). This is done
11202 by only inserting errors if the .err file is newer than the auto generated
11203 C file. To rebuild all the error codes from scratch (the old behaviour)
11204 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11205 or delete all the .err files.
11208 *) CAST-128 was incorrectly implemented for short keys. The C version has
11209 been fixed, but is untested. The assembler versions are also fixed, but
11210 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11211 to regenerate it if needed.
11212 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11213 Hagino <itojun@kame.net>]
11215 *) File was opened incorrectly in randfile.c.
11216 [Ulf Möller <ulf@fitug.de>]
11218 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11219 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11220 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11221 al: it's just almost always a UTCTime. Note this patch adds new error
11222 codes so do a "make errors" if there are problems.
11225 *) Correct Linux 1 recognition in config.
11226 [Ulf Möller <ulf@fitug.de>]
11228 *) Remove pointless MD5 hash when using DSA keys in ca.
11229 [Anonymous <nobody@replay.com>]
11231 *) Generate an error if given an empty string as a cert directory. Also
11232 generate an error if handed NULL (previously returned 0 to indicate an
11233 error, but didn't set one).
11234 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11236 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11239 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11240 parameters. This was causing a warning which killed off the Win32 compile.
11243 *) Remove C++ style comments from crypto/bn/bn_local.h.
11244 [Neil Costigan <neil.costigan@celocom.com>]
11246 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11247 based on a text string, looking up short and long names and finally
11248 "dot" format. The "dot" format stuff didn't work. Added new function
11249 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11250 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11251 OID is not part of the table.
11254 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11255 X509_LOOKUP_by_alias().
11258 *) Sort openssl functions by name.
11261 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11262 encryption from sample DSA keys (in case anyone is interested the password
11266 *) Make _all_ *_free functions accept a NULL pointer.
11267 [Frans Heymans <fheymans@isaserver.be>]
11269 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11271 [Anonymous <nobody@replay.com>]
11273 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11274 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11276 *) Don't blow it for numeric -newkey arguments to apps/req.
11277 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11279 *) Temp key "for export" tests were wrong in s3_srvr.c.
11280 [Anonymous <nobody@replay.com>]
11282 *) Add prototype for temp key callback functions
11283 SSL_CTX_set_tmp_{rsa,dh}_callback().
11286 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11287 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11290 *) X509_name_add_entry() freed the wrong thing after an error.
11291 [Arne Ansper <arne@ats.cyber.ee>]
11293 *) rsa_eay.c would attempt to free a NULL context.
11294 [Arne Ansper <arne@ats.cyber.ee>]
11296 *) BIO_s_socket() had a broken should_retry() on Windoze.
11297 [Arne Ansper <arne@ats.cyber.ee>]
11299 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11300 [Arne Ansper <arne@ats.cyber.ee>]
11302 *) Make sure the already existing X509_STORE->depth variable is initialized
11303 in X509_STORE_new(), but document the fact that this variable is still
11304 unused in the certificate verification process.
11305 [Ralf S. Engelschall]
11307 *) Fix the various library and apps files to free up pkeys obtained from
11308 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11311 *) Fix reference counting in X509_PUBKEY_get(). This makes
11312 demos/maurice/example2.c work, amongst others, probably.
11313 [Steve Henson and Ben Laurie]
11315 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11316 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11317 are no longer created. This way we have a single and consistent command
11318 line interface `openssl <command>', similar to `cvs <command>'.
11319 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11321 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11322 BIT STRING wrapper always have zero unused bits.
11325 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11328 *) Make the top-level INSTALL documentation easier to understand.
11331 *) Makefiles updated to exit if an error occurs in a sub-directory
11332 make (including if user presses ^C) [Paul Sutton]
11334 *) Make Montgomery context stuff explicit in RSA data structure.
11337 *) Fix build order of pem and err to allow for generated pem.h.
11340 *) Fix renumbering bug in X509_NAME_delete_entry().
11343 *) Enhanced the err-ins.pl script so it makes the error library number
11344 global and can add a library name. This is needed for external ASN1 and
11345 other error libraries.
11348 *) Fixed sk_insert which never worked properly.
11351 *) Fix ASN1 macros so they can handle indefinite length construted
11352 EXPLICIT tags. Some non standard certificates use these: they can now
11356 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11357 into a single doc/ssleay.txt bundle. This way the information is still
11358 preserved but no longer messes up this directory. Now it's new room for
11359 the new set of documenation files.
11360 [Ralf S. Engelschall]
11362 *) SETs were incorrectly DER encoded. This was a major pain, because they
11363 shared code with SEQUENCEs, which aren't coded the same. This means that
11364 almost everything to do with SETs or SEQUENCEs has either changed name or
11365 number of arguments.
11366 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11368 *) Fix test data to work with the above.
11371 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11372 was already fixed by Eric for 0.9.1 it seems.
11373 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11375 *) Autodetect FreeBSD3.
11378 *) Fix various bugs in Configure. This affects the following platforms:
11382 unixware-2.0-pentium
11386 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11387 before they are needed.
11390 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11394 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11396 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11397 changed SSLeay to OpenSSL in version strings.
11398 [Ralf S. Engelschall]
11400 *) Some fixups to the top-level documents.
11403 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11404 because the symlink to include/ was missing.
11405 [Ralf S. Engelschall]
11407 *) Incorporated the popular no-RSA/DSA-only patches
11408 which allow to compile a RSA-free SSLeay.
11409 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11411 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11412 when "ssleay" is still not found.
11413 [Ralf S. Engelschall]
11415 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11416 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11418 *) Updated the README file.
11419 [Ralf S. Engelschall]
11421 *) Added various .cvsignore files in the CVS repository subdirs
11422 to make a "cvs update" really silent.
11423 [Ralf S. Engelschall]
11425 *) Recompiled the error-definition header files and added
11426 missing symbols to the Win32 linker tables.
11427 [Ralf S. Engelschall]
11429 *) Cleaned up the top-level documents;
11430 o new files: CHANGES and LICENSE
11431 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11432 o merged COPYRIGHT into LICENSE
11433 o removed obsolete TODO file
11434 o renamed MICROSOFT to INSTALL.W32
11435 [Ralf S. Engelschall]
11437 *) Removed dummy files from the 0.9.1b source tree:
11438 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11439 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11440 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11441 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11442 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11443 [Ralf S. Engelschall]
11445 *) Added various platform portability fixes.
11448 *) The Genesis of the OpenSSL rpject:
11449 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11450 Young and Tim J. Hudson created while they were working for C2Net until
11452 [The OpenSSL Project]
11455 Changes between 0.9.0b and 0.9.1b [not released]
11457 *) Updated a few CA certificates under certs/
11460 *) Changed some BIGNUM api stuff.
11463 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11464 DGUX x86, Linux Alpha, etc.
11467 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11468 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11472 *) Add -strparse option to asn1pars program which parses nested
11474 [Dr Stephen Henson <shenson@bigfoot.com>]
11476 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11479 *) DSA fix for "ca" program.
11482 *) Added "-genkey" option to "dsaparam" program.
11485 *) Added RIPE MD160 (rmd160) message digest.
11488 *) Added -a (all) option to "ssleay version" command.
11491 *) Added PLATFORM define which is the id given to Configure.
11494 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11497 *) Extended the ASN.1 parser routines.
11500 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11503 *) Added a BN_CTX to the BN library.
11506 *) Fixed the weak key values in DES library
11509 *) Changed API in EVP library for cipher aliases.
11512 *) Added support for RC2/64bit cipher.
11515 *) Converted the lhash library to the crypto/mem.c functions.
11518 *) Added more recognized ASN.1 object ids.
11521 *) Added more RSA padding checks for SSL/TLS.
11524 *) Added BIO proxy/filter functionality.
11527 *) Added extra_certs to SSL_CTX which can be used
11528 send extra CA certificates to the client in the CA cert chain sending
11529 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11532 *) Now Fortezza is denied in the authentication phase because
11533 this is key exchange mechanism is not supported by SSLeay at all.
11536 *) Additional PKCS1 checks.
11539 *) Support the string "TLSv1" for all TLS v1 ciphers.
11542 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11543 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11546 *) Fixed a few memory leaks.
11549 *) Fixed various code and comment typos.
11552 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11553 bytes sent in the client random.
11554 [Edward Bishop <ebishop@spyglass.com>]